Improving care for older people with long-term conditions and social care needs in Salford: the CLASSIC mixed-methods study, including RCT

Implementation

1. How do key stakeholders (commissioners, strategic partners) view the SICP, what do they expect from it and how is it aligned with their objectives and incentives?

2. What is the process of implementation of two key aspects of the SICP [the MDGs and the integrated contact centre (ICC)]?

Outcomes

1. What is the impact of the MDGs on the outcomes and costs of people with long-term conditions?

2. What is the impact of health coaching in the ICC on the outcomes and costs of people with long-term conditions?

CLASSIC cohort

Our researchers presented the cohort to a local group. In response to their feedback, we made changes to the survey, including increasing font size and type; improving questionnaire layout; providing an indication of time to complete; providing an explanation of why we are asking the questions; adding a statement regarding confidentiality, especially around data sharing; including examples of question completion; providing name and telephone details of a contact to assist with completion; and adding free-text boxes to enable people to add their own comments. The group also provided a ‘critical friend’ approach to letters and participant information sheets being developed to send to potential participants.

In March 2015, we presented to the PRIMER group, and its members provided advice on encouraging people to stay in the cohort and around how we fed back the results from CLASSIC. It was agreed that providing ongoing incentives would help retain participants.

Health coaching

At the meeting in March 2015, PRIMER members discussed the health coaching model, providing guidance on participant recruitment and retention with the telephone-based intervention.

Members of local groups were consulted in 2015 about recruitment methods and gave feedback that many older people were unlikely to answer their telephone to an ‘unknown’ caller. It was therefore agreed that a letter would be sent to potential participants, which included the telephone number that they would be called from, helping increase uptake.

Study methods: multidisciplinary groups

Multidisciplinary group fieldwork took place from March to December 2015, with fieldwork largely based on non-participatory attendance at neighbourhood MDG meetings. Data collection included 72 hours observing MDG meetings, with sequential fortnightly observations in one neighbourhood MDG for each of the three waves of roll-out (Table 5). Additional observations were undertaken at other meetings supporting the MDG process (including the working group meetings, joint chairpersons’ meetings and administrator meetings) as well as engagement events. Further observations were conducted with MDG nurse and social care co-chairpersons to outline how the work of referring and prioritising patients for discussion and pre-MDG meetings was enacted.

By agreement, field notes made during the MDG meetings did not contain any identifiable patient data. Initials, sex and the general practice were recorded, permitting further questioning around individual cases with the relevant general practitioners (GPs) and to identify potential patients to be invited to participate in qualitative interviews.

Thirty-two face-to-face interviews were carried out with professionals participating in the MDG meetings or those whose work was associated with them. Maximum variation sampling was used to ensure that representatives from all staff groups participating in MDGs were interviewed.

We used routine data (workload and throughput, patient characteristics, links with other services) to contextualise our data. Operational documents were collected from the MDG processes and meetings around them and used to provide information about the implementation.

Study methods: integrated contact centre

Fieldwork took place between October 2015 and July 2016, during which time the single integrated referral point (SIRP) was based within a Salford City Council facility. Colocation with the council corporate team unfortunately meant that permission to carry out observational work within the SIRP was declined. Data collection was therefore based mainly on interviews with 11 ICC staff during which in-depth descriptions of their work were provided in lieu of observations (Table 6).

We explored the various services provided by the centre through individual interviews with participating staff and managers to assess the development of the service over time, how fidelity to the model was achieved and the potential for unintended consequences. We used routine data reported by respondents (workload, patient characteristics, referrals) to contextualise the data.

We described the characteristics of the centre, its staffing and technology, and how the existence and function of the centre is communicated to patients. At the level of the patient, we described the interaction between the staff and patients, through individual interviews with six patients/carers who had direct experience. Observations included 11 hours of non-participant observations of meetings directly related to the centre, including a short visit to the SIRP, observing the locality base, a care homes meeting and initial engagement events promoting the wider SICP. In addition, documents providing evidence of the implementation process and allowing a comparison with the initial plans for the ICC at the start of the SICP were collated. Health coaching data are presented in Appendix 2.

Qualitative analysis methods

Qualitative data from both the ICC and MDG observations and interview transcripts were organised using NVivo 10. Techniques from grounded theory were used for the thematic analysis. 76 Analytical memos were written and discussed to develop a collective understanding of the issues represented in the data. Members of the qualitative team met monthly to discuss emerging themes and to agree subthemes.

Normalisation process theory was used as a starting point to inform the original topic guide used in the qualitative interviews. We considered how data mapped onto the framework, and although there were some connections between concepts, these were limited. We therefore adopted a more responsive approach, using iterative sampling and analysis of data until no new information emerged. This prevented the background framework from constraining the interviews and allowed us to learn from, and develop, the topic guide as the interviews were conducted.

Qualitative data from both the ICC and MDGs observations and interview transcripts were organised using NVivo 10. We conducted a thematic analysis drawing on some techniques from a grounded theory approach, including open coding and the creation of analytical memos as a basis for iterative analysis and sampling as outlined previously. Members of the qualitative team met monthly to discuss emerging themes and subthemes, any unusual cases and to agree the final stage of ‘selective’ coding. These processes of coding and iterative analysis enabled core themes to emerge inductively from the data consistent with a grounded theory approach. 76

Training and supervision

The intervention was delivered by a health advisor (Agenda for Change band 4) who was already delivering health coaching to patients with diabetes mellitus or pre-diabetes mellitus. Training focused on the additional skills needed when dealing with a wider range of long-term conditions and dealing with low mood. A session on long-term conditions was run by a GP from the CLASSIC team, to help prepare for potential queries. Advisors were encouraged to refer people to the NHS Choices website.

Advisors were trained in detecting and working with participants with low mood over a 2-day training session with further updates over time. They were given the opportunity to role play low mood assessments and delivering interventions.They also received a comprehensive manual to aid them in delivery of the low mood component.

The advisors received clinical supervision (initially fortnightly, then monthly from applicant KL) by group for 1 hour. They had an opportunity to discuss challenging calls and risk issues. They were supported in making decisions about how to progress the intervention for each patient who identified as having low mood and in delivering structured low mood intervention such as behavioural activation. The advisors had contact details for the supervisors for immediate concerns.

Recruitment and assignment of interventions

Participants were enrolled through the CLASSIC cohort. In a standard trial, patients receive information and then provide informed consent to participate. At that point, they are randomised. A significant drawback is that patients are told about different treatments in the different arms (including any new treatment), but only half the patients are randomised to that new treatment. This can cause dissatisfaction.

In the cmRCT, patients eligible for the trial are identified from the cohort and randomly selected. Patients who are randomly selected for usual care continue to be followed up in the cohort and are not informed about the trial or the randomisation. Patients who are randomised to the new treatment are then contacted and offered the treatment. They still provide consent to the new treatment and can decide whether or not they wish to receive it. If patients agree to the new treatment, they are provided with the new treatment and continue to be followed up in the cohort. If patients decide that they do not wish to receive the new treatment, they continue to receive usual care and continue to be followed up in the cohort.

We piloted these procedures in 50 patients to test the rate of uptake of the new treatment.

After assessment of eligibility, we selected patients randomly for health coaching or usual care using appropriate central randomisation through a clinical trials unit to ensure concealment of allocation. In this pragmatic evaluation, there was no blinding of patients or providers. All outcomes were either self-reported or routine data.

Sample size and analysis

At the time of study development there were no bespoke methods for powering cmRCTs, and, following existing cmRCTs, we used conventional methods. We powered the study to have 80% power (α = 5%) to detect a standardised effect size of 0.25 on any continuous outcome measure. Allowing for 25% attrition among participants – and assuming that outcome measures at baseline correlate 0.5 with their follow-ups – 504 patients were needed (252 per arm).

The initial uptake rate was lower than anticipated; hence, we selected a further 252 patients to be offered the intervention. However, within the cmRCT framework all 504 patients offered treatment remained in the treatment group in analysis, including those who declined. In consequence, the effect size between arms detectable at 80% power was 0.39 among those consenting to treatment.

Analysis followed intention-to-treat principles and a prespecified analysis plan (see Appendix 1). In summary, we report the trial and analysis in accordance with the updated Consolidated Standards of Reporting Trials (CONSORT) standards and utilising the extension for pragmatic trials. The main test of the intervention was that the overall main effect of the intervention is zero. Condition group was used as a binary variable. Continuous outcomes were assessed using linear regression, controlling, where appropriate, for baseline values of the respective outcome. Outcomes measured using ordinal scales were treated as continuous variables. Results for non-normal variables (skew or kurtosis > 1.0) were confirmed using bootstrap analysis. Baseline values of outcomes and design factors were included in all analyses. Some additional covariates were prespecified.

Owing to implementation delays, no patient was offered treatment up to 6 months after the baseline assessment and for some the offer was not made until month 12 or later. This caused variations in the duration of time before start of the treatment, ranging from 259 to 513 days. Length of follow-up from the end of treatment to 20-month follow-up was similarly variable. Thus, the trial is considered to have run for > 20 months, with patients receiving treatment at any time within that period.

The cmRCT design provides an estimate of the mean effect in people offered treatment. Compared with a pragmatic trial, which provides an estimate of the mean effect in people agreeing to treatment, the effect is ‘diluted’ by the proportion of patients in the treatment arm who do not consent to treatment. An estimate of the effect size in patients consenting to treatment was obtained through application of a complier-average causal effect (CACE) analysis. 105,106 CACE does not increase the power to detect an effect.

Economic analysis

The economic analysis aimed to assess the incremental cost-effectiveness of health coaching compared with usual care.

The primary outcome measure for the economic evaluation was health-related quality of life measured by the EQ-5D-5L,107 a new version developed as a result of concerns over the lack of sensitivity to change of the original. Published English general population preference weightings70 were used to convert responses to a single utility index for each time point.

This was combined with in-hospital mortality information taken from the secondary care utilisation data, applying a utility value of zero to all patients on death. Quality-adjusted life-years (QALYs) were calculated using the ‘area under the curve’ method, assuming linear extrapolation of utility between time points. QALYs experienced in the second year of the trial were discounted at an annual rate of 3.5%, as specified by the National Institute for Health and Care Excellence (NICE) in its reference case. 108

Resource utilisation and costs

Resource utilisation and costs were calculated from the perspective of the UK NHS. Patient-level utilisation data were collected from two sources. Information on GP contacts in the previous 6 months was collected from cohort data at 6, 12 and 18 months. Hospital utilisation data were extracted from linked administrative patient records provided by the NHS, divided into emergency admissions (short stays, ≤ 5 days; long stays, > 5 days), elective admissions, elective day cases, outpatients and A&E attendances.

Utilisation data were combined with relevant unit cost data for the price year 2014/15 to calculate total costs. Unit costs not available for this price year were inflated to 2014/15 prices using the Consumer Price Index. 109 Costs occurring in the second year were discounted at a rate of 3.5%. 108

Unit cost figures were sourced from the Personal Social Services Research Unit’s (PSSRU’s) unit costs of health and social care (2015)110 and national NHS reference costs. 111

Health coaching costs

Costs were estimated combining the cost of training and supervising staff, materials and delivery of the health coaching sessions. The intervention was offered to all participants randomly selected, although only 189 received at least one call and were used to estimate costs.

Missing data

Data required for QALY and cost calculation were missing in a small number of cases (n = 2), and were imputed. Missing information on age and sex was sourced from administrative data (sex, n = 6; age, n = 35) or imputed (missing age n = 30), to ensure independence from allocation. 112

For missing EQ-5D-5L and resource use data, we used multiple imputation by chained equations to generate 50 imputed data sets, assuming that the data were missing at random. The independent variables specified in the imputation models were age, sex, treatment arm and baseline EQ-5D-5L. To account for non-normality, predictive mean matching was used to ensure values observed in the original data set. Multiple imputation was conducted using Stata’s ICE package and analysis using Stata’s MI package (version 14.2; StataCorp LP, College Station, TX, USA).

Cost–utility analysis

The economic analysis estimates the incremental cost-effectiveness of the offer of health coaching compared with usual care at standard UK willingness-to-pay thresholds.

The primary analysis was based on a comparison on the full sample with multiple imputation. A sensitivity analysis was performed using only the complete-case sample (health coaching n = 206, usual care n = 378). Analysis used Stata version 14.

The incremental cost-effectiveness ratio (ICER) was calculated, adjusting for age, sex and baseline EQ-5D-5L index score. 113 To assess uncertainty surrounding the estimates and to account for the typically skewed nature of cost data, incremental costs and QALYs were bootstrapped using pairwise bootstrapping with replacement using 10,000 replications. Cost-effectiveness planes plot these 10,000 bootstrap replications of the ICER estimates to illustrate the uncertainty around the point estimate of the ICER in probabilistic terms. Finally, cost-effectiveness acceptability curves (CEACs) were plotted to represent graphically the probability of the intervention being cost-effective across a range of cost-effectiveness thresholds.

Community asset use

Individuals were asked ‘Have you attended or used any of the following community groups, activities and services in the last 6 months?’ and provided with the following options: (1) trade unions, (2) group for the elderly or older people (e.g. lunch club), (3) environmental groups, (4) youth groups (e.g. Scouts, Guides), (5) parent–teacher association or school association, (6) Women’s Institute, Townswomen’s Guild or women’s groups, (7) residents’ group or neighbourhood watch, (8) social club (including working men’s clubs, Rotary Clubs), (9) education, arts, music or singing groups, (10) sports club, gym, exercise or dance groups, (11) religious group or church organisation, (12) other group or organisation and (13) charity, voluntary or community group. This represented only a part of the SICP community assets programme.

An individual was classified as using community assets if they ticked ‘yes’ to one or more. The list of assets was less comprehensive in waves 2 and 3, as it included only (2) and (8)–(12). However, results were robust to these changes.

We obtained the costs of outpatient attendances, ambulance use and A&E visits from NHS reference costs,111 and cost of a GP visits from a PSSRU costs publication. 110 In 2014/15 prices were as follows: GP appointment £65.00, outpatient attendance £134.22, ambulance callout £96.35 and A&E visit £131.92.

Analytic methods

How do people in the CLASSIC cohort use community assets and how does that change over time?

We present descriptive data from the cohort on community asset use over time.

1. Is community asset use associated with quality of life and health-care utilisation?

We used variation in use of community assets to explore its causal impact. In the absence of randomisation, determining causal associations is complex, as cross-sectional associations may reflect reverse pathways (health influencing asset use rather than vice versa) or unmeasured confounders. We conducted the following analyses:

1. examined cross-sectional associations between asset use and outcomes (quality of life, health-care utilisation), controlling for measured confounders

2. used an instrumental variable approach (using distance to community assets as an instrument) to test whether or not the association was maintained, to provide a more rigorous test of a causal relationship

3. used longitudinal data to assess whether or not change in asset use was associated with outcomes.

Associations between asset use and outcomes

We compared quality of life, health-care utilisation and costs between users and non-users of community assets. We estimated three multivariate models for quality of life and utilisation, including community asset use and:

1. sex, age and socioeconomic characteristics only

2. sex, age, socioeconomic characteristics and the presence of 23 long-term conditions

3. sex, age, socioeconomic characteristics and the presence of 23 long-term conditions limiting activity.

We combined the effects on health-related quality of life and costs to produce estimates of the societal value of 1 year’s participation in community assets using the net benefit framework. 124 This net benefit combines both the benefit and the cost into one single metric. For example, it would be better equipped to deal with a scenario in which an individual’s health improved at the same time as they used more NHS resources (i.e. having higher costs). This involves multiplying EQ-5D scores by a value for a QALY and subtracting the annual cost of health-care utilisation. We use a range of threshold values (£20,000, £30,000 and £12,936)125 and multiplied 6-month costs to obtain annual costs. Net benefit was then used in three further regression models, which included sex, age, socioeconomic characteristics and 23 long-term conditions limiting activities.

Instrumental variable analyses

We received council data containing postcodes of community assets (May 2014–15). These assets were defined by the council as:

. . . buildings or amenities that play a vital role in local life . . . community centres, libraries, swimming pools, village shops, markets or pubs.

The Stationery Office. 126 Contains public sector information licensed under the Open Government Licence v3.0

We included a range of asset types (‘health and well-being’, ‘skills and training’, ‘food’, ‘outdoor space’, ‘physical exercise’, ‘religious’, ‘shopping’, ‘social’, ‘volunteering’ and ‘other’), but excluded those coded as ‘older teenage’.

Data from patient postcodes enabled the calculation of distance between households and assets, by translating postcodes to co-ordinates and calculating distances using the Stata^® (StataCorp, College Station, TX, USA) ‘vincenty’ command via the law of cosines. We calculated (1) the minimum distance to the nearest asset (in miles) and (2) the number of assets within a given radius (1, 2, 5, 10, 15 and 20 miles).

As noted earlier, regression of community asset use on outcomes can produce bias via reverse causality. To overcome this bias, we used an instrumental variables approach, defined as ‘an analytical technique, traditionally used in non-randomised research studies, which uses a variable associated with the factor under study but not directly associated with the outcome variable or any potential confounders’. 127 Instruments are variables that are related to asset use, but not directly to outcomes. We used the distance to nearest asset and the number of assets in a certain radius as instruments. We argue that the way in which assets impact on health is through use only. That is, the distance to an individual’s nearest asset can affect their health only if that individual then uses that asset. Just because an individual lives close to an asset is not sufficient for their health to improve, they must make use of that asset.

Instrumental variable models have two stages:

1. the prediction of asset use conditional on minimum distance and number of assets in the radii (and other cofounders)

2. effect of asset use on health (the first stage prediction of asset use is used in the second stage).

As asset use is binary, we used linear regression with endogenous treatment effects:128,129

1. we estimated three separate multivariate models for the EQ-5D score, which included asset use and sex, age and socioeconomic characteristics only

2. sex, age, socioeconomic characteristics and the presence of 23 long-term conditions

3. sex, age, socioeconomic characteristics and the presence of 23 long-term conditions limiting activity.

We then estimated the same models for utilisation (GP, outpatient and total costs). Finally, we combined the effects on quality of life and costs to produce estimates of societal value.

Longitudinal analyses

We estimated the effect of asset use using a matched control group analysis,130 whereby the effect is defined as the mean difference between the observed and the imputed potential outcome for each individual. We use nearest ‘neighbour matching’ to impute potential outcomes, which takes an average of the outcome of similar subjects that are not treated. We further matched on baseline EQ-5D to better capture baseline health. As well as considering the effects of starting asset use, we estimate effects of stopping. We estimated short- (baseline to 6 months) and long-term (baseline to 12 months) effects (Table 12) of starting and stopping use of assets.

In all models, we matched on baseline age, sex, living arrangements and the presence of limiting conditions. We additionally matched on baseline EQ-5D scores in a later specification. For utilisation and net benefits outcomes, we further matched on baseline levels of these outcomes.

Care plans

The analysis of the effects of care plans used analogous methods to the assessment of community assets, using change in use of care plans over time to explore potential impacts on patient outcomes.

The measure of care plans was the question ‘Do you have a written care plan?’ at baseline and 18 months.

Outcomes were:

1. patient perception of integration; we used a single item from the unpublished QIPP measure, ‘Do you think the support and care you receive is joined up and working for you’, scored on a four-point Likert scale

2. patient activation as measured by the PAM (see Chapter 4)

3. EQ-5D.

Participation in community assets

Fifty per cent (1829/3686) of respondents at baseline reported using community assets. Most reported using one, and those using at least one community asset reported use of an average of two assets.

Table 13 shows participation over time among those reporting data at three waves of the cohort.

Associations between community asset use and outcomes

On average, community asset users had an EQ-5D score of 0.690 compared with 0.596 for non-users (difference 0.094; p < 0.001) (Table 14). Table 14 shows that there were few differences between the two groups in other characteristics, but 9% of community asset users had university qualifications, compared with 4% of non-users.

On average, community asset users had visited a GP three times in the previous 6 months, 0.33 visits fewer than non-users (p < 0.001). Community asset users reported an average of 2.2 hospital outpatient appointments in the last 6 months, 0.3 visits fewer than non-users. Users also reported fewer ambulance call-outs and casualty visits. The average health-care costs over a 6-month period in users was £97 (p = 0.003) lower than non-users (£448 vs. £545).

Instrumental variable models

Table 15 shows the average distance, with users of assets closer to assets [difference –0.026, 95% confidence interval (CI) –0.039 to –0.013]. There is little difference in the number of assets within given radii for users and non-users, suggesting that asset availability is not responsible for the higher quality of life in users.

We ran regressions with different combinations of distance and radii variables. All met validity tests:

1. instruments predict asset use (first stage chi-squared statistic of 316.47; p < 0.0001)

2. they are not overidentified, as shown by a Sargan score of 1.425 (p = 0.84).

The best-fitting model using Bayesian information criterion131 was one with five instruments: (1) distance to nearest asset, (2) this distance squared, (3) number of assets within a 1-mile radius, (4) number of assets within a 10-mile radius and (5) number of assets within a 15-mile radius. Table 16 presents results from the regression. Column 1 shows the first-stage results. Longer distances to the nearest asset reduce the probability of use (as expected). If the distance to the nearest asset increases by 1 mile, then individuals are 66% less likely to use assets. The number of assets in a local area is statistically significant only if that area is defined as within a 15-mile radius. In second stage results (the effects of asset use on quality of life), we observe a significant positive effect of use. Controlling for health conditions and their effects on activity reduces the effect, but it remains statistically significant.

In analyses of the effects of asset use on utilisation and costs, asset use does reduce health-care utilisation and total costs, but the reductions are not statistically significant (Table 17).

Using the current NICE threshold values of £20,000–30,000 per QALY, we estimated that the net benefits of community asset use were £3624.20 (95% CI £1473.52 to £5054.83) to £4665.90 (95% CI £2061.96 to £7269.78) per participant per year (see Table 17). Using the threshold value proposed by Claxton et al. 125 gave a net benefit estimate of £2288.30 per participant per year (95% CI £1067.94 to £3508.70).

Longitudinal analyses: short-term effects (0–6 months)

We present a range of models estimating the effects of starting to use assets (Table 18), with the primary analysis being ‘nearest neighbour with five matches’ (see Table 18). Starting using assets increases EQ-5D scores (0.058, 95% CI 0.031to 0.085), but has no significant effect on GP or hospital outpatient visits. Analysis of net benefit showed large and statistically significant gains at all three thresholds.

Table 19 shows effects for stopping asset use, whereby the reduced benefits associated with stopping were smaller than the benefits of starting and the change in net benefit was not statistically significant.

Longitudinal analyses: long-term effects (0–12 months)

Table 20 shows long-term effects of asset use were similar to the short term. Effects on health-care utilisation remained non-significant, but net benefit increases were significant.

The effects of stopping asset use were not statistically significant for EQ-5D and health-care utilisation, and were significant only in terms of net benefit, whereby the reductions were considerably larger in the longer term than the short term (Table 21).

Use of care plans

The numbers of people reporting a care plan increased over time (baseline 6.4%, 12 months 8.9%, 18 months 10.7%). A similar pattern was observed when restricted to people with data at all three time points (baseline 5.4%, 12 months 8.5%, 18 months 10.4%).

We used a quasi-experimental approach and estimated the effect associated with gaining a care plan. As before, we used the nearest neighbour matching method to match older people who gained a care plan to similar individuals who did not get a plan between baseline and 18 months.

In all models the outcome is either perceptions of integration, patient activation or quality of life at 18 months. We tested the following models:

• model 1 – matched on age, sex and living arrangements at baseline, 1 : 1 matching

• model 2 – model 1 plus presence of limiting health conditions at baseline, 1 : 1 matching

• model 3 – model 2 plus baseline value of outcomes, 1 : 1 matching

• model 4 – model 3, 5 : 1 matching

• model 5 – model 3, propensity score matching.

Results

Gaining a care plan was associated with significant increases in the proportion of patients reporting that their care was ‘joined up and working’ (Table 22). There were no associations between gaining a care plan and either patient activation (Table 23) or quality of life (Table 24).

Early development of programme commissioning and governance

Understanding the commissioning and funding of the SICP is not straightforward, as the programme was implemented and developed in a changing policy context. However, there are two core commissioning processes underlying the SICP: (1) national Section 75 funding and the associated Better Care Fund (BCF) and (2) the local Alliance Agreement.

Since 2006, local authorities and NHS commissioners have been able to pool aspects of their budgets to support joint initiatives, under Section 75 of the National Health Service Act 2006. Historically, these initiatives have been small scale (e.g. focused on joint commissioning for children with learning difficulties). Generally known as ‘Section 75 funding’, this continues to be the statutory underpinning for the funding of the SICP and the developing ICS.

In order to promote greater integration, in 2013 the Department of Health and Social Care announced the BCF,139 an ongoing national programme available to local authorities and NHS commissioners to jointly plan and deliver integrated services, with an initial intention that £3.8B of health and social care funding would be pooled from April 2015. The BCF did not introduce any new legislation; Section 75 remained the statutory mechanism. Local health commissioners were required to identify elements of their existing budgets to pool with local authority budgets. It was assumed by policy-makers that the costs of integrated services would be covered by the savings generated by a reduction in hospital admissions.

The BCF was announced after the start of the SICP. The SICP commissioning partners (CCG and council) had already agreed to pool £98M under Section 75. The BCF process (compulsory for all CCGs and local authorities) required them to submit plans to NHS England and the Local Government Association.

In response, the CCG identified £20M of its £98M pooled budget as its BCF contribution. There were problems with the BCF application, locally and nationally. 139 Locally, the situation was complicated by the existence of the wider agreement to pool a larger budget. This meant retrofitting existing plans to fit BCF requirements and submitting further evidence to satisfy the national process. In addition, there were specific problems surrounding the expectations about estimated reductions in hospital use:

We have been through the assurance process for the BCF; we have been approved with support. This is the second best rating. We have received a number of risk issues that we need to respond to by the end of November, several of them relate to each other. ID 6 spoke to the BCF and is working with the [S]ICP steering and finance steering group, the risks have now been reduced to 10. We are pleased that our plans to reduce emergency admissions have been accepted. Nationally they wanted to reduce emergency admissions by 3.5%, we put forward that we should reduce them by 1% locally. This suggestion looked like a sticking point but it has been accepted.

ID 16 senior city council manager

Taking the Section 75 and BCF funding together, the total SICP funding started out as £98M, rising to £112M in the financial year 2015/16. The CCG contributed two-thirds and the city council one-third. Before agreements were made, a process of mapping was carried out between the council and CCG, focusing on current services for the health and social care in older people. They specifically looked at services that were in scope (at least 50% of the service had to contribute to the care of older people); any services that were deemed in scope had to be 100% within the programme to ensure that the governance arrangements were adequate. This Section 75 pooled budget arrangement was different from any previous historical budget sharing, as the operational management of the budget was governed by an Alliance Agreement, which included both commissioners and providers. The service provision included in the budget was:

• acute and community health via the foundation trust (£50M)

• care services provided or subcontracted by the city council (£30M)

• CCG expenditure including continuing health care and hospice (£10M)

• older people’s mental health services from the mental health trust (£8M).

Although the providers were not funding the SICP, contingencies for a partner organisation’s cost improvement programme were included within the contract. As part of the contract setting, providers had their contracts reduced through tariff arrangements, to ensure that they achieved cost reductions.

The traditional approach to commissioning involves a ‘commissioning cycle’: assessing population need, assessing current service availability and suitability, designing appropriate services to meet needs, assessing the availability of relevant providers and procuring the service, managing demand and monitoring performance (see http://commissioning.libraryservices.nhs.uk/commissioning-cycle; accessed 18 May 2018).

The SICP worked differently. It did not matter whether partners were commissioners or providers: both were equally involved in the decision-making:

I think it’s having a bit of faith and constantly when people raise that, just really assuring them that . . . and I can’t remember at the beginning how many times I said this is a partnership approach, it’s a partnership approach. Because people used to say things like oh well, it shouldn’t be Salford Royal dictating how that happens. Well, it isn’t Salford Royal, it’s a partnership and we’re in the room, we get listened to. And it was just constantly reassuring. And it was a different way of working actually, it wasn’t a commissioner led . . . and it was just constantly reminding people this isn’t commissioner led. It’s a partnership and we have to give and take and compromise.

ID 6 senior CCG manager

In the early phases of the SICP, when contractual decisions were being made, providers were present as members of the Alliance Board, and perceived as experts who can offer useful information, even when the service being discussed was to be provided by a different organisation. This dynamic shifted the balance of the commissioning process:

Then we’ll have individual service contracts, so the CCG or the council will then go away and have contracts in place with the relevant provider organisation. Because it might not be Salford Royal or GMW [Greater Manchester West] that’s the provider. So GMW and provider when they’re in the room are from the alliance perspective, are giving their expert opinion in terms of what will work and how we will deliver the model to achieve the outcomes, not as an organisation we can do that and we want the contract to deliver that. Some instances they will have the contract, in other instances they won’t have the contract.

ID 11 senior CCG manager

Alliance Agreement

In addition to Section 75, the SICP was also underpinned by an Alliance Agreement. This outlined how decisions were made and the governance mechanisms which applied. The document articulated how the organisations would work together as a system of both commissioners and providers. Although sometimes referred to as the ‘Alliance Contract’, it was technically not a contract, as the provisions were not specific enough to be legally enforceable:

I mean if anything on reflection I think the Alliance Contract is about publicly saying, we’re going to work together. Contractually in blunt terms it’s probably not worth the paper it’s written on, if you know what I mean? It’s the principle of it.

ID 15 senior mental health trust manager

The document can be seen as a set of guidelines for behaviour. The document outlined the terms of the agreement (initially 3 years 5 months, with the option for a further 3 years). Such an extension would have to be agreed by all parties (except the general practice provider). The agreement included provision for an annual review, which was intended to ensure that the SICP was delivering its objectives. The Alliance Agreement was not a definitive document; many issues remained to be clarified and it simply set out the process by which such decisions will be made in the future. For example, clause 15 outlined that a framework for risk and benefit sharing needed to be developed. Timelines were included, highlighting when tasks needed to be carried out, but these lacked specificity about who should performing those roles. Furthermore, the document lacked specificity in terms of situations in which conflicts or disagreements may have arisen. The Alliance Agreement proposed a risk and benefit sharing framework, which would outline procedures if the pooled budget faced either an over- or under-spend (Alliance Agreement, internal report).

It was acknowledged that it was impossible to prespecify all possible future scenarios, with the Alliance Board seen as the forum in which any problems would be discussed:

. . . and the reality is, when issues do arise, the situation for each one is different and the factors for creating is different. So the Alliance Board use a forum for those conversations.

ID 5 SICP programme manager

Overall, the agreement could be seen as an agreement to agree at some point in the future, rather than an actual contract. The most specific aspect of the agreement was the scheme of delegation for decision-making (see Appendix 4, Table 61), which clearly defined what decisions could be made within the structure of the SICP without each decision having to be taken back to each partner organisation.

Alongside the Alliance Agreement, there was a detailed ‘service and financial plan’, which provided a breakdown of the pooled budget, with additional detail regarding the investment and disinvestment of services, BCF plans and the general models for implementation.

Although not a legal contract, respondents stated that the process of developing the agreement had been as important in supporting the early development of the programme as the details of the agreement itself. Simply knowing that there was a formal process of sign-up, with clear governance processes, allowed the key organisations to feel secure in decision-making. Presenting the details in the form of an Alliance Agreement made their declarations ‘official’ and provided legitimacy. The process of developing the Alliance Agreement allowed the key stakeholder organisations time and resources to think about what they wanted to achieve, outlining risks and benefits for the organisations:

But the benefit of the Alliance Agreement was primarily the process we went through to agree it. It was refining a shared vision. It was having the difficult conversations about, you know, what are our anxieties, what do we want to achieve. It codified the things we were setting out to do and our expectations of each other.

ID 4 senior foundation trust manager

Furthermore, individuals stated that part of the value of the contract was that the act of signing a document makes the stakeholder organisations more likely to work through issues when disagreements arose:

So it’s a big deal, you know, you sort of owe the other stakeholders once you’ve agreed this. Because people will walk away without any of that control, they always have, and will do. So hence there has to be an overbearing focus on governance, it dominates everything.

ID 3 senior CCG manager

During the period in which the Alliance Agreement was in place, no significant conflicts or disagreements arose. Therefore, the Alliance Agreement arrangement was never tested. It was acknowledged that, in part, this was because a financial buffer had been put in place, limiting the financial risk:

Yeah, so there was some slippage created and that balanced it out last year and then a couple of other things happened this year, so they knew that they’d got underspend within the Alliance budget . . .

ID 5 SICP programme manager

Towards an integrated care organisation

Not long after the SICP was established, moves began towards a formal ICO (mooted in CCG documents from July 2014). This would achieve two things:

1. move beyond the ‘partnership’ model underpinning the SICP, moving staff from the council social care team into the foundation trust alongside community services colleagues

2. extend the population covered to all adults.

The logic underlying this was not clear. In interviews, respondents simply characterised this as ‘the obvious next step’, without clearly explaining why:

I mean, obviously, one of the things that we’ve got to do in terms of the programme, I mean, we’ve moved onto the ICO rather than the integrated care programme but it makes sense because it’s the next step really.

ID 2 senior city council manager

This respondent spoke of greater ‘efficiency’, without clarifying what this might mean and how it might be achieved:

The agreement that was reached as the system was that the ICO itself needs to have benefits in and of itself, so it has to be more efficient to have more people together, so having social workers working in the same teams employed by the same body as district nurses, has to be more efficient. You have to be able to reduce some duplication etc., so the ICO has a target for efficiency. It comes as a result of bringing staff closer together.

ID 25 senior foundation trust manager

An ‘outline case’ briefing document was used internally to provide information for those involved in discussions about the ICO (Box 2). This document set out the advantages of integrated care and suggested that a more formal integrated organisation could overcome deficiencies in the partnership model:

However, no evidence was forthcoming about the existence of these problems in Salford. Indeed, interviews with senior leaders in all organisations had emphasised the strength of the partnership model. The focus appeared to be on anticipating and preventing future problems. Thus, at an engagement event to promote the ICO idea, the focus was on the ICO as an obvious step in order to cement the gains made to date:

We are motivated by success and we have a really strong shared vision for the future. We want to harness what we have got and maximise benefits. All the main organisations have agreed to go a step further so that we can deliver services to all adults in an integrated way; integration will be delivered by an ICO or a lead provider organisation. This is a really big step and change but we will be working with the same outlines as we did for older people. Our vision is to deliver care in a different way, centred on individual needs to maximise the benefits.

ID 6 senior CCG manager engagement event

During interviews, individuals were asked what the ICO could offer that the SICP could not, especially as the programmes of work were essentially unchanged, other than expanded age coverage. The ICO was described as a structural mechanism enabling the transformation that the SICP was trying to deliver:

It’s structural [ICO] . . . by structural I mean employment, legal, etc. The integrated care programme is about transformation, so it’s about doing things differently . . . setting up a preventative way of working, supporting people to support themselves, etc., so clear goals and aims and all the rest of it. The ICO is an enabler to that, so by bringing staff together by removing barriers, by having one team Salford, it would be easier then to transform the work what we do, to deliver the goals of the [S]ICP.

ID 25 senior foundation trust manager

The SICP and ICO were not described as separate entities. Instead, the SICP was perceived to outline the strategy for the programme of work, whereas the ICO has been introduced as a mechanism to deliver the strategy:

. . . but the programme I suppose is a strategy and the ICO is actually the operational delivery, I suppose that’s one way, but they kind of knit it together, they’re not two distinct entities, if you like.

ID 17 senior mental health trust manager

The ICO was described as a way of providing a single leadership structure, which provided the co-ordination of services and reduces fragmentation. Although partnership working was perceived to be good locally, the single leadership structure of the ICO was an opportunity to better co-ordinate care. The ICO was identified as a means to examine care pathways more closely to ensure that there was one service for patients:

That’s four different hand-offs along a pathway. You know, it enables Salford Royal to look at it and go, d’you know what, it makes sense to have one service. People’s experiences is one experience, so . . .

ID 2 senior city council manager

One possible explanation for the decision to establish the ICO lay in the broader NHS. In 2014, NHS England published its Five year Forward View. 140 Central was a call for NHS and social care organisations to develop ‘new models of care’.

Five new models were proposed:

1. primary and acute care systems (PACS), bringing together acute hospitals with primary, community and social care providers

2. multispecialty community providers (MCPs), in which all types of community providers work together with their social care counterparts

3. extended care in care homes, in which local primary and community care providers work together with private- and council-funded care homes to improve care for patients

4. urgent and emergency care, in which different combinations of providers work together across a geographical area to rationalise and improve urgent and emergency care

5. acute care systems, in which providers of acute care work together to rationalise and improve provision of more specialised services.

Volunteer groups of providers (designated as ‘vanguards’) were invited to come forward to test these ‘new care models’, and would be provided with additional funding and support.

This provided an opportunity for the SICP, as national policy supported existing local plans. There had already been discussions about how the SICP could be extended and the model embodied in the SICP (close collaboration between a foundation trust, CCG, general practice providers, and the local authority commissioners and providers of social care) already met many PACS requirements. As the vanguard was aligned with existing local plans, linkage would offer additional financial support, as well as access to a network of providers undertaking a similar journey.

A second motivation lay in the more general environment for large providers. A 2014 review of provider activity and development for the Department of Health and Social Care141 highlighted both the need for innovation in the provision of care and the need for a facilitative regulative structure at national-level supporting providers looking to innovate. The report advocated allowing innovative providers to move ahead rapidly, with light touch regulation and support. The development of ICOs was one of the innovations advocated. The opportunity to be seen as at the leading edge of developments was attractive.

Thus, 2015 saw the introduction of a PACS vanguard with an intention to integrate primary and acute care systems, joining up general practice, hospital, community, social care and mental health services for adults. The ICO was to be established as the ‘prime provider’ of services across Salford, bringing the adult services provider arm of the city council into the foundation trust and subcontracting with other providers (such as non-specialist services at the main mental health trust) for a range of services.

The application set out what the vanguard hoped to achieve, building on the ‘successful’ elements of the SICP, aligned to a federated model of general practice. As discussed above, the ICO/vanguard included the transfer of adult social care from the city council to foundation trust, including the contracts for adult social care homes and domiciliary care. Furthermore, the contract for mental health services for adults and older people was also transferred to the foundation trust, which could be subcontracted to the mental health trust.

The practical issues associated with ICO set up were considerable. The aim was to bring together local adult health, social, mental health and acute care into one place, underpinned by a single health and social care contract for the ICO. For the foundation trust to take on this new role, a large number of staff moved across from council to trust under Transfer of Undertakings in Public Employment (TUPE) legislation:

We’ve been working on what we call an ‘on-boarding process’, so to welcome 400 people into the new organisation, they have to learn new systems for time recording, sequence recording – all sorts of things – e-learning. But that’s just the practical side of things . . . The other part of my role is around the business strategy side. I’ve been involved in the development of various legal agreements, contracts, and risk share agreements, which is hell of a lot of work, just to make sure that organisationally we understand the relationships for the future and there’s no barriers that could be created because we didn’t clearly articulate what the arrangement was, so . . . I think that’s quite a strength for Salford and has been for some time, that we’ve got a group of senior managers across the organisations that get together and think those agreements through in a very methodical way.

ID 6 senior CCG manager

Integrated care organisation implementation was rhetorically designated as a ‘safe landing’. It was agreed by all stakeholders that, for the future success of the ICO, the initial phase was to implement the ICO, transfer staff across and take on the responsibility without initially making any changes to how people work and how services are delivered. The stakeholders wanted to continue to work in partnership and not produce additional risk:

. . . the phrase we use is, safe landing, so nothing should change. The people who are receiving services . . . nothing should change because there should be a smooth transition, so that we’re not creating any risk in the system, but having said that, it is obviously different as well, so you’ve got to get it right for staff and so on. It is a period of change and we’ve been really clear and all partners have agreed that, yeah it will take a few months for that to bed down, but everybody should sort of move in to their new roles if you like, but also we’re all still here, we’re all still a partnership.

ID 25 senior foundation trust manager

Ongoing commissioning of integration: the ‘integrated care system’

As we have seen, the local and national policy context was constantly shifting, requiring those involved with integration services to adapt. The most recent change was the devolution of responsibility for health and social care to Greater Manchester in the devolution settlement. 136 This required all boroughs to produce a locality plan, setting out how health and social care services will work together and providing an opportunity for leaders to develop an overarching narrative, which pulled all the various initiatives together into a coherent whole.

Leaders now talked about an ICS, which aimed to make sense of the changes that have occurred and generate organisational and public ‘buy-in’. The partnership underlying integrated care was rebadged, with the overall aim of improving the health of the population through integrated care. The vision was expressed in the concepts of ‘start well’, ‘live well’ (addressed by the ICO, focusing on the adult population) and ‘age well’ (addressed by the SICP, people > 65 years of age).

The ultimate aim of the ICS was to link general practice more closely with the other stakeholder organisations to deliver services in a neighbourhood model. This would involve general practices working more closely together to provide services across a geographical footprint covering 30,000–50,000 people. To facilitate this, a general practice provider organisation would bring practices together and form a single body with which the ICO could contract and interact. How this would work in practice, and what the relationship would be between general practices and the ICS, remained unclear.

To support these developments, new governance and contractual mechanisms have been required.

In our initial interviews, it was suggested that the Alliance Agreement would form the basis for the ongoing developments. However, once it had been decided to move towards an ICO, a new governance structure was necessary. The ICO formally came into being in July 2016. The new ICS was focused on the integration agenda for the whole of the adult population, which included the original SICP. The new pooled budget covered £246M of expenditure.

The areas of work for the adult population were said to be based on learning from the SICP and will continue to include the programmes of work (community assets, MDGs and ICCs). The focus of the adult population will be across several key priority areas including:

• vulnerable adults

• long-term conditions

• mental health (drug and alcohol dependency in particular)

• dementia

• mental and physical needs.

The aim of the ICS was to bridge a projected £6M financial gap over a 5-year period through service redesign. A new service and financial plan was devised for the whole adult population based on a pooled budget between CCG and council from 2016 to 2021. The budget included all adult social care, learning disabilities, community equipment, foundation trust hospital care, foundation trust community services, adult (non-specialist) mental care services (hospital and community), A&E, continuing health care and NHS funded care, and some CCG grants/contracts for non-NHS providers.

The plan for the ICS was developed in phases:

• Phase 1 – in March 2016, an initial review of the commissioning intentions within the older people’s financial plan was conducted.

• Phase 2 – in May/June 2016, a stocktake of the workstreams within the SICP was carried out, to set a clear work plan for 2016/17; pre-planned city council and CCG workstreams and service reviews were documented alongside the vanguard proposals.

• Phase 3 – the amalgamation of the workstreams above, outlining the outcomes of the plan, shared programme risks and the process by which the plan will be delivered and monitored.

For the ICS to operate, commissioning intentions across Salford have been consolidated over a number of workstreams. The consolidation included:

• older people’s service and financial plan 2014/15–2017/18

• existing or preplanned city council or CCG workstreams and reviews

• aspirations within Salford PACS (the vanguard application).

To oversee the ICS a new governance structure was developed to replace the Alliance Agreement and Alliance Board from June 2016 (see Appendix 4, Figure 13). There are now two new committees: (1) the Advisory Board for Integrated Care, which has a similar membership to the Alliance Board, bringing together commissioners and providers, and (2) an Integrated Adult Health and Care Commissioning Joint Committee (ICJC; see Appendix 4, Figure 14), which involved only commissioners. This was necessary to allow commissioners to make joint decisions about the pooled commissioning budget:

So it was felt to be more appropriate for us to move from an alliance of the two providers and two commissioners to the two commissioners making all the decisions about the pooled budget, the commissioning budget by agreeing a service and financial plan and keeping that as a commissioner only decision-making process . . . Now in reality there’s no way because of our history in Salford we would just go in a dark room and write that as commissioners and present it to the providers, it would still be an iterative process. So we’ve established two groups, there’s the decision-making board, which is a joint committee between the council and the CCG, and then there’s an Advisory Board which is quite similar to the old Alliance Board . . . But they do not have any decision-making. But if they felt really, really strongly about something it is going to have an influence on it, they might not be the decision-makers but it will have an influence. And if we reach a stage where we’ve got the Advisory Board saying one thing and the decision-makers say another then we’re in a bad place really so we have to sort that out.

ID 6 senior CCG manager

The ICJC would oversee the pooled budget across the council and CCG, whereas the Advisory Board has responsibility for the additional investment that has been received from the devolved Greater Manchester transformation fund to support the vanguard.

Devolution added another layer of complexity. The first year of being a nationally designated vanguard saw additional funding to support the work. This was non-recurrent, intended to support ‘transformational’ activity by, for example, allowing investment in infrastructure or to support ‘double running’. In the second year, vanguards were required to submit a further value proposition, setting out what additional funding they required. However, the Greater Manchester Devolution Partnership had negotiated a transformation fund with the Treasury, and it was decided that Greater Manchester’s allocation of vanguard funding was included in this total. The rebidding process caused delays and complications, but, under the transformation fund, Salford accessed £18.2M over a 3-year period:

But because GM [Greater Manchester] has a 5-year funding pot, we were able to secure commitment for more than a single year. So we put forward an application, as I say, over 3 years, so we’ve got a 3-year investment plan, which matches with the 5-year disinvestment plan. And our ICO and the arrangements that we’re establishing, and this investment and disinvestment, makes a contribution to closing Salford locality plans projected funding gap.

ID 4 senior foundation trust manager

However, ongoing funding depended on demonstrating impact:

We always have healthy debates here. We need to demonstrate that we are doing what we said or else GM will terminate the funding.

ID 4 senior foundation trust manager (September Advisory Board)

The contractual mechanism was a so-called ‘prime provider’ contract between commissioners and providers. This form was relatively new and untested in the NHS. A single large organisation is designated as the ‘prime provider’ for a range of services. This provider is then responsible for providing those services, either themselves, or by subcontracting with other providers. The foundation trust was the ‘prime provider’, with the scope of services covered including care for the entire adult population. Thus, the foundation trust will be contracted to provide acute and community care, subcontracting for mental health services. Under this contract, commissioners set the overall strategy and the prime provider was then responsible for service delivery within the specified budget, as well as the quality of subcontracted services. The contract will usually include a range of incentives and risk-sharing provisions. The original intention was that the foundation trust would also subcontract for some services provided by general practices, via a primary care provider organisation. However, this proved difficult.

General practice was an important strand in the initial vanguard plan:

The vision is that general practice, at a neighbourhood level, will collaborate and embrace all community-based health and social care services for a defined population. This includes other primary care based services; pharmacists, dentists and opticians, as well as community and voluntary organisations. It is expected that all care that doesn’t require hospital facilities will be based and managed in the neighbourhood and for pathways of care crossing into hospital to be better for patients. Personalised care planning will benefit from a multiprofessional approach, drawing in specialist expertise and resources as needed.

Vanguard value proposition, internal primary care document

However, interviewees acknowledged that, in developing the ICO, less attention had been paid to engaging with primary care in the development of this vision. General practices, although engaged in the SICP via their participation in MDGs, remained largely detached from the broader ICO development:

So I think . . . well, some obviously GP practices no matter how often you try and talk to them they don’t really have time to listen, do they, really? They weren’t necessarily listening and then there were lots of misunderstanding about what we were doing and then jump to [conclusions] . . . and everything felt like two steps forward, one back and everything.

ID 6 senior CCG manager

An intensive programme of engagement work was launched in spring 2015. Conducted by the CCG, the intention was to encourage member practices to become more involved with the integration agenda. Later, this included the appointment of an external consultant (January 2016), who engaged with local GPs. This engagement highlighted the lack of connection between the broader agenda of the ICS and the wider NHS, and the day-to-day reality of GPs’ working lives:

What was immediately clear is actually GPs were largely disengaged from most of the process that had been going on. So by that what I mean is actually most people were unable to articulate what a vanguard was, nearly everyone was unable to articulate what the ICO was. Very few people had heard of Five Year Forward Views and the like, and actually when you started to talk about the context of integration in general practice, most of the stuff that the GPs wanted to know was ‘is this going to relieve my workload, is this going to help me with my workforce’, and occasionally ‘is this going to help me with my premises, which is also under pressure.’. So, once again, it wasn’t that GPs were pulling against the idea of an ICO, it’s just it wasn’t relevant to their day jobs.

ID 26 independent GP

Summary: commissioning integrated care

As we have highlighted, the development of the SICP and the ICO fundamentally changed the relationship between commissioners and providers. Rather than commissioners identifying needs, specifying services to meet those needs and then procuring relevant services, commissioners and providers in Salford are working in partnership, co-designing services. This was clearly seen during the fieldwork; when observing the Alliance Board meetings, an observer would not be able to tell which participants are providers and which commissioners. The commissioners, any more than any other partner, do not set the terms of the discussion, and there is no ‘challenge’ from commissioners. Rather, the focus is on a partnership approach. This raises questions about the longer-term development of the commissioner role, as it seems likely that future service developments will be provider led. In addition, as a prime provider, the foundation trust will be commissioning and contracting for mental health services on behalf of the CCG, leading to potential longer-term complications around conflicts of interest:

There is the potential the SRFT [Salford Royal Foundation Trust] will change as they commission services as they will have different agendas, will the CCG pass on that infrastructure / resource? We need to be clear what the purpose of the CCG is. The CCG may choose to commission PC [primary care] and they still have an acute agenda, but for MH [mental health] this will be delegated to the ICO. There will be potential conflicts of interest with the roles of commissioning and providing. They need to be clear what the commissioning agenda is and what is the provider agenda is. This is difficult as there is an invisible wall where commissioning ends and providing starts. For example if in the future in the MDGs, SRFT feel that GMWFT [Greater Manchester West Foundation Trust] aren’t providing mental health to the standard they expect, they may challenge as a commissioner but then we could argue that there aren’t enough district nurses which is a problem for them as a provider.

ID 17 senior mental health trust manager

Furthermore, the development of this strong partnership gave providers in the room some power over other (non-partner) providers locally. Contractual arrangements with these other providers were discussed at partnership meetings, and this could give foundation trusts an advantage. Although this can be viewed as potentially detrimental with regard to competition, it could also be argued that having the high-level people sitting around a table offers a forum in which commissioners and providers can have discussions beyond the SICP to bring about change, and that this may be more effective than competition as a means of improving services. For example, at one Alliance Board there was a need for the CCG to commission more beds at the trust:

We’ve invested in SR [Salford Royal] for 18-month contract for beds, etc. whilst the review is worked through. For community equipment we have supported the business case in principal, we need some more information. The good news is that this is major investment area for the programme.

ID 11 senior CCG manager

When the ICO was in development there seemed to be a need to re-establish a distinction between the commissioner and provider roles. During interviews, key stakeholders were asked why the roles were being separated. It was explained that it was a request of the commissioners and was required if they would be working in a prime provider model.

I think there’s a lot going to one place, so I think commissioners needed a way of holding that to account, so I think that’s one thing. I think the second thing is that the city council’s obviously changing its role . . . moving from being a direct deliverer of services to being a commissioner of services, so I think the view was that was part of the process of allowing the city council to move into its new role of being a strategic commissioner.

ID 25 senior foundation trust manager

-

The change in the relationship was described as a mechanism to clarify roles and responsibilities. However, the new arrangements do not override all of the partnership working that has been established:

So that’s slightly delineates responsibilities more, you know, it’s much clearer now. Salford Royal is responsible for making integrated care happen through a supply chain. And the commissioners have the responsibility for commissioning from Salford Royal as the prime provider, the range of integrated care services and outcomes that they want to achieve. That doesn’t mean that we’ve reverted to just a contractual relationship, you know, but that world exists in parallel – and I mean that in a good way – to one where we are sitting down as partners and debating what is it that we want to do.

ID 4 senior foundation trust manager

During an observation of the Advisory Board there was an acknowledgement that there were a number of committees that were commissioner or provider only. However, the majority were still perceived to be following a partnership model of working:

There are other sources of finances separate from here. In terms of the governance, I think it is important that we are still all working in partnership other than a couple of committees.

ID6 senior CCG manager

The consequences of the new working relationships are unclear. However, observations of the workings of the Advisory Board suggested that the separation between the commissioners and providers has led to additional complexities being introduced into the structure of the partnership and the financing of specific projects. The Advisory Board has the decision-making responsibility for the non-recurrent investment funds allocated to the ICO from the transformation fund, whereas the ICJC will oversee the pooled recurrent budget across the council and CCG for the adult population. Although they are clearly separate ways of funding programmes of work, it was identified that some specific programmes of work will obtain funding from both budgets. This will impact on how and where decisions are made, underlining the need to tightly manage communication.

These complexities will arise in other vanguard sites as they establish new models of care. There is work being undertaken by the national team responsible for vanguard development to support them as they navigate the new relationships required. The test will most likely come if a prime provider fails to deliver the outcomes for which they are contracted.

At the onset of this research it was assumed that the commissioning of the programme would be stable, and that it would be possible to explore the relationship between SICP progress and success and the mechanisms used to commission it. In reality, partly owing to the ambitions of those leading the programme, and partly in response to conditions in the wider NHS, the commissioning and governance structures associated with integrated care were in constant flux. It is therefore very difficult to make links between programme commissioning and governance, the extent to which implementation has been successful and the achieved outcomes. Although the structures and governance of the SICP were changing in the fundamental ways, the work of the SICP has continued more or less unchanged. A ‘stocktake’ determined that there was as yet no evidence that the programme will meet the initial agreed targets (e.g. reduce hospital admissions), but no specific changes were suggested other than to continue with the ICO implementation.

In the next section we explore the factors that seem to have supported or inhibited setting up the SICP.

Rule 1: blend designated with distributed leadership

The SICP was strongly led from the top, including strong representation from senior managers on the Alliance Board. Furthermore, there had been stability among these managers. The foundation trust initially took an overall lead given the disestablishment of PCTs and creation of CCGs; as well as initially driving the agenda, it was also prepared to commit more resources.

Furthermore, senior leaders were prepared to commit significant amounts of time. Weekly meetings were held between key people involved with the development of the programme:

And I think one of the things I learnt from the programme is that momentum of meeting weekly, being really disciplined in doing the work and meeting weekly really started to pay off quite quickly, in that it really got the momentum going in that project . . . So it sounded really over the top to me, weekly, it really did. But I think as long as you don’t set yourself massive tasks during the intervals, or give yourself dates to do certain things.

ID 6 senior CCG manager

At the same time, a robust management structure was also developed, with strong commitment from the middle management in partner organisations:

So we’ve got some executive and chief officer leadership, there’s some sort of second line reports beneath that level where there’s some dedicated time, and that’s largely about kind of the governance of the programme through the finance and steering group now. There’s a dedicated project team, and then we’ve got some operational capacity, so people in city council, in Salford Royal mostly, a little bit in GMW [Greater Manchester West] where we’ve released capacity, we’ve said, actually, this is really important to us . . .

ID 4 senior foundation trust manager

This structure included a clear system of delegated decision-making. This was important in supporting the day-to-day running of the programme. Our interviews suggest that, in the early stages, close partnership working between senior leaders from both the commissioners and providers was important in deciding what the interventions should be and in getting the programme running. This was particularly important in the time and effort invested in developing the Alliance Agreement.

Working together to produce an agreement supported the partners in developing a trusting relationship, which in turn supported the development of the programme as a whole. Once implementation was under way, however, the focus of these senior leaders on the Alliance Board shifted towards the development of the ICO and the vanguard, leaving the middle-level managers to get on with the day-to-day tasks required to run the programme. The scheme of delegation allowed the middle tier of managers to operate with some freedom within the confines of their delegated authority, in keeping with the need to allow those below the most senior level of management to provide local leadership. 44 The leadership for the integration agenda as a whole has remained the same throughout the development and implementation of the SICP, ICO and ICS. The same people representing the four key stakeholder organisations have been involved with the partnership, which has enabled trust and consistency to be developed. This has allowed individuals to carry out the work required of them, rather than working on establishing relationships and trust across the organisations.

Alongside this, the programme was set up with designated managerial time and support, further facilitating ‘distributed’ leadership. Thus, for example, initial work on the programme was carried out with support from an independent innovation and improvement centre.

Rule 2: establish feedback loops

‘Feedback loops’ about performance are an important determinant of success in large-scale change. However, although helpful, other tools to support behaviour change – such as financial incentives – may be important. 138 The SICP was set up with clear performance goals, centring on reduction in admissions and the improvement in patient experience. The Alliance Board received regular reports about performance against these metrics. However, delays in establishing elements of the programme allowed those involved to develop a narrative which emphasised the longer-term nature of expected improvements:

. . . and what people have fed back to me is that it has taken longer than was anticipated to implement, . . . the point at which people say yes, this is up and running and people . . . it’s embedded, it’s become the way of doing things, you would then start to see something . . . some impact, so that’s . . . and I think from what I understand there are starting to be some initial . . . [impacts].

ID 5 SICP programme manager

There were also concerns about the data that was being used to monitor the programme indicators:

Another measure, not a local measure but certainly it gets a lot of national attention and is one of the Better Care Fund measures is our delayed transfers of care. That has significantly reduced since we’ve been doing this programme, however, I am dubious about the data quality of that particular measure and this goes back many, many years in that it is open to a little bit of interpretation when that’s a delayed discharge.

ID 6 senior CCG manager

The governance structures of the Alliance Agreement were such that the Alliance Board – with the senior leaders – was the primary reporting forum for such data. Although the boards of the individual organisations each received progress reports from the Alliance Board, ownership of the performance metrics sat with the Alliance Board. At the same time as these somewhat disappointing metrics were being reported, the Alliance Board was undertaking the complex work of establishing the ICO. This brought about considerable changes to the working lives of those affected, in particular city council staff. As we have seen, motivating such staff and making the considerable changes needed required enthusiasm and momentum to be generated, which in turn required a positive narrative of success. Thus, in engagement events, a positive message about the achievements of the SICP was required. It is possible that this inhibited more critical reflection on progress.

This suggests a further mechanism, which may be important in the operation of feedback loops. The presence of the most senior leaders from the partnership organisations on the Alliance Board was very helpful in establishing the programme. However, as implementation proceeded, it is possible that the fact that ownership of performance metrics sat at Alliance Board level acted to limit independent or critical scrutiny by the boards of the constituent organisations. Thus, our study suggests that, although a strong governance structure is important in establishing a programme such as this, which requires joint working across organisational boundaries, it is also probably important that the individual boards retain ownership of the performance of the collaboration, and that those in receipt of the information are not limited in their ability to respond by the demands of implementation.

The SICP did not establish any specific incentive mechanisms at programme level beyond the overall incentive that if savings were made they could be reinvested in relevant services. As we have seen, the Alliance Agreement made reference to the establishment of risk and benefit sharing mechanisms, but these were not formally elucidated. In the other workstreams, incentives for engagement were indirect, couched in terms of the ability to provide a better service.

Rule 3: attend to history

Our interviewees and those we observed in meetings clearly expressed the opinion that the establishment of the SICP owed much to local history, highlighting longstanding collaborative working and working relationships over a considerable period of time:

So in Salford, we’ve got a very strong record in terms of joint working. So from a commissioner-only perspective between the council and the PCT [primary care trust] that was and the CCG, we’ve got a track record of pooled budgets and the council taking lead for commissioning on services and CCG for the others. So we already had pooled budgets in place for immediate care, for the equipment service, for learning difficulties. And we’ve got joint appointments. So we’ve got a pool of commissioning managers that work across the council and the CCG as well. So we’ve got some very good examples of how we’ve worked collectively as commissioners. So we’ve got that history.

ID 11 senior CCG manager

At the same time, it was acknowledged that the geographical and organisational context (shared boundaries and similar geographical coverage) was important. Moreover, this has been stable for many years:

Salford has quite a reputation for being a very strong partnership-focused district, and if you were to look at the underpinning behind that there are a number of aspects that I think it’s fair to say are drivers for it, so it’s had a stable boundary, so it hasn’t had to go through constant upheaval in terms of boundary changes like Lancashire or other districts have had with bits of it coming in and going out.

ID 10 city council public health

Unlike the earlier study,138 there was no clear evidence that this strong shared historico-geographical context acted to limit the ambition for significant change.

Rule 4: engage physicians

In keeping with the model of major system change,138 we found that engagement must include a much broader range of stakeholders than simply physicians. In the early phases of the SICP there had been some engagement work with community service providers, social care staff and some interested GPs to define the model. Wider engagement was carried out with staff affected by the ICO. Interviews suggested that such outreach work had been fairly successful in generating a shared sense of purpose. Thus, interviews with individuals across the stakeholder organisations, revealed a consistent view about the programme. The development of the SICP was about developing a system of work rather than a series of initiatives:

So I suppose part of this is an all-system approach to seeing that the right people get the right service at the right time. So that is the aim of just enough support, I suppose.

ID 9 city council manager

Individuals in the CCG and the council highlighted the need to focus on prevention:

You’ve got an acute sector that has a very, very unusual leadership style, which basically I think recognised very early on that growing the acute economy was not the future, so a sense that you could not build a sustainable model for Salford Royal on the basis of just hoovering up additional acute activity, but recognising that it needed to think about new areas of business, new models of delivery, get away from simply acute focus care and begin to think about where it could operate full pathways into the communities – only into some of those areas where you might be talking about preventative types of activities.

ID 10 city council public health

However, as we have indicated, engagement of the broader population of local general practices was limited, and this limited the ability of the ICO to deliver its more ambitious plans for primary care. General practices remained independent contractors and, although practices are members of the CCG, it is not clear what this ‘membership’ really means. 142 The local experience suggests that CCG status as a membership organisation does not mean that CCG leaders can speak unequivocally on behalf of the members. The CCG was a key partner, but GPs were more detached:

Salford who had basically done a lot of work on the ICO but were finding that general practices’ role within the ICO was probably actually further away than it had ever been, but with the failure of their local [in inverted commas] GP federation to engage, but also the fact that actually general practice had very different priorities.

ID 26 independent GP

General practice engagement proved problematic in two ways. First, as discussed in implementation 2 (see Chapter 11), in the early stages of MDG development, general practice involvement was not always easy to obtain:

I can talk from personal experience about my practice, my practice is one of the practices involved in the MDGs and our experience has not been positive so far of MDG working, in that we don’t feel we’ve got anything out of it, we don’t feel it’s reduced admissions, readmissions or anything, so I expect it to be quite negative. To such an extent that the partner who’s going stopped going so our practice hasn’t been involved and we’re the only practice in [the neighbourhood] that hasn’t been involved.

ID 14 CCG GP

General practitioners struggled to take time out of practice, and good engagement was obtained only when reimbursement was increased. Second, GP engagement in the ICO was also limited. Investigation by an independent consultant revealed significant pressures:

Then actually that was quite revealing in itself in that there was this perception that GPs were not engaging, well actually it was just the fact that they weren’t informed and it wasn’t that no one was trying to tell them, it was just once again that some of them are actually worried about their very survival beyond the end of this year, so people start talking about stuff that’s going to be happening in the next few years, it sort of becomes abstract for them.

ID 26 independent GP

They also expressed concerns that the foundation trust was becoming too dominant:

I know, and people keeping asking why SRFT [Salford Royal Foundation Trust] are in charge of the neighbourhoods. I’ve said that SRFT are not, this is a partnership. So in terms of primary care, we have suggested that people who join the neighbourhood groups can be represented on the ICO group.

ID 6 senior CCG manager

These findings suggest that the need identified for ‘system-wide authority’ invested in individuals or groups in order to align multiple stakeholders138 may be difficult when general practices need to be engaged in a change programme.

More widely, there were also some issues with engagement with other stakeholders.

For example, some of those working in community assets found that they were isolated from the rest of the programme and experienced confusion around the terminology (often medicalised) used by others in the SICP:

To come in cold and then try and understand what it was about and then try and develop well what does . . . because in the end community assets is a bit, I don’t know, I suppose it sits outside slightly or did sit outside slightly. In the early days people really thought about what that would mean or how that might fit within the programme, people were still perhaps thinking about services, so the council services, about hospital provision, they were thinking much more along those lines. Whereas this project’s grown and developed and community assets have been able to influence a bit more, people are now starting to think about well actually the priority is what can we support to enable people to stay at home and live their lives.

ID 12 voluntary sector

Rule 5: engage with patients and the public

This issue is covered in implementation 2 (see Chapter 11) and in PPI (see Chapter 3).

How do commissioners view the programme, what they expect from it and how it is seen in terms of their objectives?

It was envisaged that the SICP would be seen by the CCG as something they were commissioning, alongside other programmes. In practice, however, CCG leaders were committed to the programme as equal partners. The SICP was seen by the CCG as essential to contain costs and improve quality, and there was a significant commitment to partnership working, whereby the needs of individual organisations were subsumed to the needs of the programme.

However, the development of the ICO and the institution of a ‘prime provider’ contract led to the establishment of formal governance arrangements which reintroduced a separation between commissioners and providers. At the time of data collection, the implications of these changes to the governance and the formal ICO arrangements were still unknown. In particular, it will be interesting to see whether or not the reintroduction of a separation between commissioners and providers leads to a different approach to outcomes and performance management.

How is the programme viewed by strategic partners such as the local authority and how is it sustained under financial pressure?

History played a pivotal role in encouraging the main stakeholder organisations to develop the SICP and the Alliance Agreement. Positive experiences of working together in the past enabled them to develop the SICP and recognised that they all had a vested interest in the programme. All stakeholders discussed the potential difficult decisions that would need to be made. However, they believed that the organisations were willing to face organisational challenges without walking away from the programme. The collective success of SICP was deemed to be more important than individual organisational gains.

The financial pressures that were being faced by all organisations provided an incentive to senior leaders to support their organisations to work differently. Although the foundation trust and CCG were perceived to have a strong financial position, there was recognition that all organisations were working together, and, therefore, the shrinking local authority budget was not considered to inhibit partnership working. However, it remains to be seen in the longer term whether or not developing financial pressures affect the work of the ICO.

The role of Greater Manchester West Foundation Trust (GMWFT) in the SICP was different from the other three stakeholder organisations. GMWFT provided services to a wider geographical footprint, and its interaction with the programme was perceived to be more distant at times. For example, GMWFT did not have as many staff working on the SICP or attending high-level meetings. This different relationship was also identifiable in the move towards the ICO. GMWFT did not have a formal role within the organisation, and its services were to be subcontracted by the prime provider.

The local development of SICP was further supported by national policy and an appetite to introduce new models of care (vanguards), which provided Salford with an opportunity to get additional investment into the local health and social care economy. Moreover, the vanguard status also gave Salford organisations the prospect of recognition as exemplars of new approaches to service delivery. The decision to move to a formal ICO was made collectively by all stakeholder organisations. We were unable to find any evidence that this decision was based upon an appraisal of the impacts of the SICP; the focus appeared to be on the prevention of future problems rather than evidence of SICP success.

Overall, the programme involved a strong partnership between all the organisations involved. All stakeholders that we interviewed expressed their commitment. However, the question as to how the SICP figured in the broader strategic plans of the stakeholder organisations was reduced in relevance by the decision to move towards an ICO. It is possible that the early decision to move towards an ICO was prompted, in part, by financial pressures, but this was not clearly articulated by those involved.

How does the programme impact on the work of the two foundation trusts, in particular how the integrated community and acute provider adapt to reductions in inpatient activity?

This question presupposed that acute hospital activity would be reduced. In practice this has not happened (see Chapter 12).

How does the programme impacts on primary care, in particular general practice?

We have described and explained how the SICP, overall, struggled to engage with general practices, and we have considered some of the underlying causes of this, as well as identified solutions.

How far are the financial incentives (explicit and implicit) in the local health and social care system aligned with the ambitions of the programme?

The SICP did not put in place any specific incentive schemes, other than direct payments to GPs to take part in MDG work and the implicit incentive that any money saved could be reinvested in services. The initial funding for GPs was not deemed to be adequate to cover their costs and time; therefore, additional funding was required by the CCG to ensure GP commitment. The CCG funded local GPs through a local enhanced service for patients with long-term conditions. It was found in implementation 2 (see Chapter 11) that GPs were being funded twice – both by the local enhanced service and through the SICP – for doing similar work. This demonstrates how the objectives of the CCG mirrored that of the SICP, but also highlights difficulties associated with engaging GPs. Over and above the GP-related work, we did not see any evidence of reliance on specific financial incentives to implement SICP, other than the general incentive that it was argued that the programme would improve the sustainability of the local health economy.

What type of integration does the SICP represent?

As discussed in the initial literature review (see Chapter 1), there are many different ways of defining integrated care. Bringing together our observational evidence, interviews and documentary analysis, we have identified two somewhat different ‘programme theories’. Figure 6 sets out our interpretation of the theory underlying the SICP as it was conceived. Figure 7 sets out our interpretation of the programme theory underpinning the latest developments.

FIGURE 6.

Overall programme theory underlying SICP (initial thinking). LA, local authority; SRFT, Salford Royal Foundation Trust.

FIGURE 7.

Overall programme theory underlying Salford ICO/ICS (from July 2016).

In its initial phases, the SICP was underpinned by a strong ideal of partnership working. The three programmes of work were identified as targets for service change and it was assumed that the experience of working together would break down organisational barriers and support wider improvement. At the same time, the programmes of work would support local people in becoming more resilient. This suggests a model of integration in which the focus is on functional and service-level integration, in which staff from different organisations work together across organisational boundaries. From this perspective, the integration of systems [such as information technology (IT) or financial systems] develops out of what is required to make collaborative working possible.

For the ICO (and subsequently the ICS), those involved told us that closer structural integration (with organisations either subsumed into the foundation trust or subcontracted by them) would lead to the embedding of functional integration, in which working together across organisational and professional boundaries would become the norm, leading to better outcomes and patient experience. This view of integration assumes that the integration of systems, financial flows and management systems is a prerequisite for integrated working. In this model, integrated organisations and systems come first, with integrated working flowing from this. The initial vision underpinning the SICP focused on the value that would be derived from the functional integration associated with the three programmes of work. By contrast, the vision underpinning the ICO suggests that, to be successful, such functional integration requires structural integration.

Our study, which has finished just as the development of the ICO is coming to fruition, cannot adjudicate between these two approaches. Indeed, it could be argued that these changing underlying assumptions and beliefs simply reflect the changing context of the NHS. This includes two important macro-level pressures: the national drive towards the formation of new models of care under the vanguard programme, in which elements of structural integration, although not essential, are encouraged; and a worsening financial environment in which all NHS and social care commissioners and providers are struggling to balance budgets. In this environment, the creation of larger, more streamlined organisations which are able to reduce duplication may seem an attractive proposition, regardless of the impact on the delivery of more integrated care.

Supporting and commissioning large-scale service change: comparison with other research

We have used the published framework for understanding large-scale system changes. 44,143 This proved valuable in analysing our findings, and our evidence confirms and amplifies the importance of some elements embodied in the framework.

First, the local experience highlighted the value of the clear scheme of delegation, which allowed middle-level managers to exercise local leadership within delegated resources. This allowed the programme to be implemented while senior leaders were engaged in developing the next phase.

Second, our study reaffirmed the importance of history and geography in providing a context within which change programmes can be implemented. This included good personal relationships, which had developed over many years, and a local geography, which allowed coherence around a core group sharing the same geographical footprint. Although some senior personnel did change towards the end of our fieldwork period, respondents were clear that the trust developed by working together on complex agreements over time was a vital ingredient in its ability to implement the ICO.

Third, our study suggested an interesting nuance in the need to establish feedback loops, particularly in the context of an evolving programme. The establishment of an Alliance Board, which included the most senior leaders from the contributing organisations, may have acted to limit the scrutiny of outcomes. Our study suggests that, although the establishment of a strong governance structure for a collaborative venture is important, it is probably also important that the individual boards of the collaborating organisations retain ownership and scrutiny. In establishing feedback loops, it is thus important that those in receipt of the information are not limited in their ability to respond by the demands of implementation.

Finally, our study re-emphasised the need to engage with stakeholders. Although some emphasise the need to engage with a wide range of local stakeholders,143 the local experience suggested that engagement of GPs had a specific impact. The plans involving general practice included in the ICO vanguard application remain some way from fruition and have required an intensive process of engagement.

Looking at previous research, the most comprehensive analysis is that of the ICPs undertaken by RAND. 51 In Table 26, we compare the findings. The RAND report also highlighted potential ‘enablers’ to integrated care, which are compared in Table 27.

Taken together, it can be seen that the SICP fulfils many of the conditions highlighted by the RAND report as underpinning the successful programmes, and this could be said to be reflected in the success in establishing the three programmes of work making up the SICP (albeit more slowly than intended).

Emerging metrics that suggested that targets were not being met were not acted on, and we have already highlighted the failure to fully engage with GPs. Furthermore, as we have highlighted, the goals of the programme continued to develop while it was being implemented, making it difficult to assess fully the impact of these issues. RAND highlighted the impact of increasing scale and complexity, and it remains to be seen how the wider changes enacted under the ICO play out.

Evidence relating to the potential transferability of the SICP

This evaluation is intended to generate learning that could be applied to other areas, and we have tried to do this by using and building on existing frameworks for understanding major system change. However, local contexts and relationship dynamics are perceived by those involved as the key ingredients enabling the development and implementation of the programme:

The programme itself didn’t come from thin air. The programme came from personal relationships built up over a programme of time across organisations, and so there was an awful lot of good [learning disabilities service] . . . So they had something they could see and was tangible within Salford, that they thought, ‘yeah, well if it can work there, it can work on a lot bigger scale, because we’ve seen the benefits for the individuals concerned who receive the service’.

ID 9 city council manager

Thus, the approach taken in the SICP will not necessarily suit all populations. A key early stage in the local journey was to explore the widely praised ‘Torbay model’ of integrated care. 144 Local leaders told us that, although successful in Torbay, on closer review the local population needs were different and therefore tailored solutions were required:

I think that Torbay/Sally Ford storyline is one that illustrates the limitations of thinking that you can do lift and drop. Populations are different. There are aspects of these things which you can mirror, and I’m sure when you talk to the people who are in the detail they will talk about the way in which they’ve looked at some of the work that Torbay did . . . There’s a sense for me about constantly reminding people about the population demographics.

ID 10 city council public health professional

So in Torbay, they are gentrified, relatively affluent, moved to the area, usually short on family connections or they were distant, and quite often are healthy till later in life because of their financial status as much as anything; whereas Sally Ford has quite often lived on the streets she was born in, has a big family network, is considerably impoverished, by comparison to Mrs Smith, and ends up with a long-term condition probably 10 years earlier than her, so the whole pressure in the system is different but also the ways of managing it are very different, because she has a community network of support that we can tap into and use. In Torbay, they didn’t.

ID 10 city council public health professional

Description of the integrated contact centre

Although the SICP focused on those aged ≥ 65 years, the ICC was developed around an existing ‘all-adult’ service and was not restricted to older people, although they formed the bulk of the caseload. The ICC initially comprised three strands: SIRP, health coaching and telehealth. Although health coaching was part of the ICC, the results of the health coaching are presented in Chapter 13. Telehealth (based on text messaging) aimed to affect admission rates by providing monitoring equipment and links to the ICC, but the SICP did not renew the licence.

Intermediate care

Access to intermediate care was via the single entry point (SEP), which had two administrative workers plus a clinician based within the SIRP. In contrast to callers to adult social care and district nursing, callers included health professionals. Intermediate care had around 100 beds based in three units providing rehabilitation and assessment. Although units did not have a permanent medical presence, they were regularly visited by geriatricians and a GP.

Uptake and usage of the integrated contact centre

Calls were tracked to ensure that targets set by the council in terms of response time, satisfaction and call volume were met. Initially the corporate target of answering 70% calls within 30 seconds was used, but the emphasis for SIRP was on the quality of the call. For adult social care staff, call handling was reviewed every 3 months by managers. All SIRP staff had access to data on call numbers and response times. Reports from the date we interviewed SIRP staff showed that 672 calls were received by the three colocated teams (between 07.00 and 18.00).

Colocation

Staff from all three services were colocated, interspersed within two banks of desks. Physical proximity helped them learn informally about colleagues’ work:

. . . you, kind of, want to listen to what their call is because you know you’re going to be dealing with those calls so you want to pick up and see what they do but most of the time you’re on the phone anyway.

ID 10474, ASCCT

Ultimately, staff felt that the integrated approach would improve responsiveness:

. . . eventually when it’s one number that will be even better because sometimes they do ring through and they’ve got a social care query and they might have a district nurse query as well so we will be able to deal with that all at once . . . I’ll just ask the person next to me rather than giving the number to ring somewhere else you’d just deal with that yourself and ask your colleague or leave a message with them.

ID 10474, ASCCT

Staff felt that colocation was essential to support the SIRP administrative team in call management:

I think it works at the moment for social care staff. I don’t think it’s working as well for the district nurse staff. And if we are integrated they’d have to put more professionals with us, and they’d maybe then need to rotate them, I don’t know. But it would need people here so that you’ve got some back up.

ID 10474, ASCCT

Central integration versus fragmenting existing local teams

Although part of the SICP, administrative staff were employed by different employers (adult social care by the council and DNs by the trust). When located separately, the fact that SIRP staff were employed on different pay scales and bands did not cause issues. Since July 2016, the employer of council staff working for the SICP has been the ICO (see Chapter 10), with human resources undertaken by the trust as ‘lead provider’:

. . . the other problem we’ve got at the moment is we’re persistently talking about three groups of people doing the same job, and them three groups of people are all on different pay grades, and how do we move that forward?

ID 10479, manager

Administrative staff from NHS services contrasted their experiences with those of their social care colleagues who had been in place for 5 years and who had managers within the same open-plan office, providing regular opportunities for team and one-to-one meetings. District nursing and intermediate care managers were based off-site. This meant that staff often sought advice from social care managers:

. . . they’ve got their bosses, they’ve got four bosses on the floor . . . We’ve nobody, and we feel a little bit hurt and deserted really. I know they’re contactable by e-mail but it’s not the same . . .

ID 10477, DN administrator

District nursing and intermediate care reported some loss of sense of being part of a team:

I get a lot of hostility at the moment as well because they’re rushed off their feet . . . they’re not fully aware and understand what we do really. We were taken for granted when we were with them [in the locality bases], we were always there; also part of the team. I don’t feel a part of any team now, I don’t feel . . . I hope I make a difference but I don’t feel . . . I’m just a faceless person looking at a screen now . . .

ID 10477, DN administrator

Centralising the SIRP created difficulties for services elsewhere. Removing DN administrative staff meant that new ‘housekeeper’ posts had to be created at each locality base to replace some of the functions that were too costly for band 5 nurses. Having at least one DN within the SIRP supported administrative decision-making and enabled calls to be dealt with in-house, but was viewed as unsustainable in the long term:

. . . there was a district nurse based here with a very specific role and purpose for triage . . . She’d already stopped about three or four calls going through to the team at like half nine, quarter to ten in the morning, which might not sound a lot, but she’s probably done a lot more by then, and that was really successful . . . I think there was the will, but there wasn’t the staff to dedicate a district nurse from here.

ID 10479, manager

Loss of a sense of ‘team’ was also true of intermediate care staff when community rehabilitation, the SEP administrative team, rapid response and the intermediate care units were in separate locations:

. . . even though we’re based with the Salford council, and integrating with them [at SIRP], . . . from our point . . . we’re merging more with Salford council team, but our own team seem to be separated quite a lot.

ID 10476, intermediate care administrator

Similarly, for intermediate care, relocating the clinician away from SIRP meant that administrative staff could not physically pass on messages. As the clincian spent a large amount of time on the telephone, it was difficult for staff to relay messages. Although other SIRP administrative staff relayed messages via e-mail, the consensus was that it was easier to give out direct numbers and suggest that people contact staff themselves, contrasting with the ethos of SIRP. Staff felt that, over time, this might mean that calls that should be routed through the SIRP would bypass them:

. . . we’ve found it a bit harder taking messages, because she likes phone call messages . . . because she’s on the phone so often, she can’t check her e-mails as often, so it’s not great really, with her not being here, for that reason . . . so I’d rather just give [her] number out to that person . . .

ID 10476, administrator

Centralising the contact service aimed to make access easier, but it also created issues when people were referred for multiple assessments: a person might require an occupational therapy and social work assessment and their carer might require a carer’s assessment. This was recorded as three separate referrals. The process was being streamlined to ensure that, to avoid duplication, a single assessment was conducted by the most appropriate person:

I was thinking, ‘hang on a minute, this one person has been contacted three times in 2 days by three different people in the same office.‘. What that person on the end of the phone must have been thinking’s going on here I’ve no idea, because if it was me I’d be like, ‘hello, are you talking to each other?’. So we’ve worked on changing that now and it does work a lot better.

ID 10479, manager

A 3-day test of change was conducted to try to rectify this. It found that colocation of a DN or intermediate care clinician alongside their administrative counterparts from SIRP was helpful in providing clinical advice and led to more calls being dealt with in-house, although this was not sustainable because of staffing shortages. One proposed solution was for clinicians to work on a rotation basis, enabling them to gain experience of SIRP work:

. . . a good mentor is being with nurses . . . I’ve picked up an awful lot of information. I don’t give clinical information out, but I do understand what their procedures are, and what they do now, it’s invaluable that.

ID 10477, DN administrator

Mental health involvement

A lack of representation from mental health within the SIRP created issues. Concerns about mental health were raised with social care via the ICC, either through referrals or through welfare notices from the police or ambulance service:

. . . if something comes through here and it’s mental health we can’t simply say ‘oh, well, we’ll assign that to mental health’ because it doesn’t work, and if we just sent things to them via their generic inbox which may or may not be manned you’ve got no guarantee that’s going to be picked up . . . all we can do here is notify the GP and see if they want to consider referring them to mental health . . . an awful lot of people have said, ‘well, why don’t we just triage mental health here’ . . .

ID 10479, manager

Two main issues seem to have influenced this: (1) the contractual agreements concerning what the mental health trust would provide and (2) the fact that, although the team will assess people, any mental health issues attributable to functional illness are discharged back to GP care without the requirement for the mental health trust to provide any intervention:

I had a meeting with the commissioners and the managers in the old person’s mental health teams and it soon transpired that historical agreements of what GMW were going to provide in terms of social care, with GMW restructuring over the years, agreements no longer were fit for purpose. They no longer fit the service like it is today.

ID 10479, manager

Although the mental health trust was one of the four integrated organisations, its staff were not transferring across to the ICO and it was undergoing its own organisational change.

Complex information technology systems

An issue faced by the SICP in general, but which particularly affects the ICC, is that the NHS and council each invested in different IT systems. Although integration of health and social care services was the primary aim, call handlers were required to work with multiple IT systems.

The trust had developed a system enabling its electronic patient record (EPR), which held hospital data to link with data from primary care via the Salford integrated record interface prior to the SICP. This was then used to develop the shared care record (SCR), which supported the sharing of information at MDG meetings. Although facilitating the MDG meetings, the summary SCR was not available to staff within SIRP. Meetings in spring 2016 showed that links between the SICP on a wider level were being considered, with a suggestion that, if a person known to MDGs telephoned SIRP and consequently had a SCR, their care co-ordinator could be contacted and updated with any relevant details. Subsequent to this, plans were discussed that ICC telephone triage would be used to manage telephone referrals from patients and carers wishing to self-refer to MDGs.

Description of the service

An operational document set out the MDG vision, model and process of service delivery. Core membership included MDG administrators, a social worker (joint chairperson), a DN (joint chairperson), a GP, a mental health lead, a community psychiatric nurse, a consultant psychiatrist, a practice nurse and a geriatrician. The document also suggested wider MDG membership if direct care or support for the individual was required, including a health improvement officer, mental health practitioners (community occupational therapist, social worker, clinical psychologist), a pharmacist, a rapid response team, a housing officer, intermediate care workers and third-sector staff.

Observations at MDG meetings and analysis of the attendance registers collated by the CCG showed that similar membership and attendance existed across all the neighbourhoods (see Appendix 4, Table 63).

The first three staff roles were funded to varying degrees by SICP. One-hundred per cent of the time of administrative and nursing leads was funded, with social care leads funded 50% for MDG work and retaining caseloads for the other 50% of their funding. This meant that nursing leads had more time available for MDG work. This led to tensions when this was seen to be at the expense of direct work with patients:

They see us on a computer and say, well, that’s not a nursing task. When I input it [SCRs] on the iPad [Apple Inc., Cupertino, CA, USA], I used to sit out in the car park because you can still get Wi-Fi there, but nobody can see you. But it’s almost like a dirty little secret doing admin when you’re a nurse, because you should be attending to patients.

ID 59, nursing lead

These frustrations led to the team leader for the MDG nurses arranging a single central workspace for all nurses in one locality base.

The pre work for MDGs was a vital part of the process, and the nursing and social care leads were responsible for ensuring that links between health and social care records were added to the SCR. In some MDGs, nursing and social care staff sat together each using their own system, with the nursing chairperson also inputting data into the SCR. In other neighbourhoods, the nursing chairpersons travelled to the locality bases to work collaboratively, combining this with ‘safety huddles’ to reduce travel. If this was not possible, two nurse co-chairpersons tended to prepare the SCR for their patient caseload then e-mail it to their co-chairperson for their input, before forwarding to others.

The nursing co-chairpersons continue to attend ‘safety huddle’ meetings in the locality bases. At these meetings the local district nursing teams discussed patients they were concerned about, and nurses visiting the same patient shared their expertise. ‘Safety huddles’ enabled the nurses to promote the benefit of MDG working, accepting referrals on behalf of patients who might benefit from discussion within MDGs:

We’ve had some difficulties in describing what the [MDG nursing] role is, so we’ve had some . . . not issues, but the question from district nursing is around, well, what exactly is it you’re doing? Now, at this point, we’re just going to a time where you can see that clinical duties will be part of the role, it was almost as if we had to step out of them initially and to try and find out how we would fit.

ID 62, MDG nursing co-chairperson

In contrast, social care leads remained embedded within their locality teams and retained a small yet complex caseload. Similarly, they were keen to promote the benefits of MDGs among colleagues, and one way was getting colleagues to shadow them at meetings:

We have monthly team meetings . . . I update my team about what we’re doing. I like to keep people informed because they might find themselves as care co-ordinator so it is important for me they understand the work of the MDG and also I am trying to get them involved by getting them one by one to attend the MDG with me, so that will give them a better understanding when I’m asking for information or asking them to do things then they will understand why.

ID 58, MDG social care lead

Weekly micro-coaching meetings at the nurse base were delivered by a facilitator from the quality improvement team. Social care chairpersons were invited to participate, although only half were present at the session observed.

General practitioners were reimbursed for MDG attendance by Salford CCG, with the CCG eventually agreeing to pay for 7 hours, including pre and post work. Practice nurses attending the MDGs were not reimbursed, although in some cases they deputised for the GPs and often shared the pre and post work, particularly when they had more frequent patient contact:

. . . we tend to split the list between us and we each take the patients that we have more information on and know best. There’s quite a lot of work to do after the meetings too and its especially important when we have discussed patients that our colleagues have referred to MDG that we write up the outcomes from the meetings quickly so they know exactly what the outcomes are.

ID 15, practice nurse

Project managers and administrators pointed out that single-handed GPs were keen for support from colleagues. Even where differences existed, this did not affect the functioning of the meetings:

. . . they’re really good groups now. They work quite well together. And the key is regular attendance really from the same people, because you can’t build a forum of trust with new faces all the time, can you? . . . we’ve had the same faces over and over, and we can be really open and honest about the conversations that we have with our colleagues, and then we all agree what’s recorded formally and put into the shared care record.

ID 55, project manager

Practice nurse input was highly regarded by the MDG team, with practice nurses often seeing patients over a prolonged period while providing long-term care.

Multidisciplinary groups chose to involve practice managers in different ways. Initially, they provided a critical function in terms of assisting with allocating ‘Sally’ levels to patients. Some accompanied their GP to meetings or deputised. MDG meetings were seen as a clinical arena for discussing patient care, and it was expected that their input would cease once MDGs became established:

. . . in the newer [MDG] meetings I can forgive it, because a lot of it is about the system and the processes, so what’s risk stratification, what’s the next steps, we need to code them, we need to refer them into choose and book . . . [But] we are experiencing more and more practice managers at MDGs at the moment, and I’m not particularly comfortable with it . . . Then it was fed back to me that actually we could do with practice managers being put on there [SCR access list], because if you want information copied from one clinical system into another, who’s going to sit there and do that?

ID 55, project manager

In many neighbourhoods, the geriatricians were appointed later, and only at pilot sites were they involved at an early stage. Even once identified, staff shortages and increasing demand limited their input:

. . . recruiting has been difficult across each of the different services . . . only three geriatricians qualified last year . . . I know that there’s a geriatrician coming to those [two MDG] areas in August, September time . . . because they are hospital-based staff really, they could, and it’s not the same as having them there in the forum for conversation, but they could liaise with us via the shared care record.

ID 55, project manager

The geriatricians’ role within the MDG was viewed as key, with high expectations about the value of their input. GPs saw the geriatricians as a key reason to engage:

I think once the geriatrician is involved, and we establish good links with him, I think it could actually be much more stimulating, and much more rewarding, in terms of educating us, and improving our standards of care. But, obviously, we’ve spent 6, 7 months without anybody . . .

ID 70, GP

During observations, the geriatricians were able to access hospital patient data, enriching the discussions with information not contained in the SCR. They were involved in many key actions and frequently advised GPs on medication reviews:

. . . earlier we were discussing a case where a lady had a rapid deterioration of a wound and I asked if she might be diabetic. The geriatrician was able to pull up her blood glucose results confirm she hadn’t been tested and sent a message for this to be done. MDG reviews mean more people can contribute to a possible solution that may otherwise have not been considered.

ID 15, practice nurse

Involvement of mental health professionals could be an issue owing to resource limitations:

I think there’s buy-in from mental health . . . you know what they’re able to deliver is fairly limited and the need is massive . . . mental illness, it goes on over a period of time and, sort of, we’re so far off anticipating what the need might be, because we’re worried about putting more need [support] in there than what’s required.

ID 89, geriatrician

Process and content of multidisciplinary group meetings

Patients were stratified into four groups of ‘need’ to allow services to be better matched. Patients receiving three or more visits per week from district nursing and/or social care teams (around 3100 patients) were deemed most likely to benefit from MDGs. Practices had voluntarily participated in an exercise to identify the 2% of their population at greatest risk of admission (national ‘enhanced service’). Although existing algorithms existed [e.g. Patients at Risk of Rehospitalisation (PARR), PARR+] to identify patients at risk, dissatisfaction led to new ones being developed to code patients to ‘Sally’ levels (Table 30 and Appendix 4, Figure 15).

Although there were tensions between the amount of work and pre work provided by various parties, there was mutual respect for the different perspectives provided. During interviews, the concept of teams tended to be reserved for each group’s specialty, and team working seemed more limited at the MDG meeting itself. An exception was where the nursing and social care co-chairpersons worked closely together and formed strong dyads.

An electronic SCR was used to enable information to be shared between the relevant statutory agencies, and these were projected during MDG discussions. The majority of SCRs were initially created by the MDG nurses, with input being provided by the social care chairperson and then being sent onto GPs, practice nurses and mental health staff. An agreed shared care plan within the SCR was then developed based on MDG discussion.

Initially, there was an emphasis on the creation of the SCR, and targets to encourage this could have affected the quality of data they contained. A working group was set up to look at content of the SCR, especially when there was an expectation that a summary SCR (in the form of a plan) would be shared with patients:

. . . we’ve, kind of, learned about the shared care record, about the data and how to look through it quickly and how working alongside the social worker with their system open and shared care record, how together, you know, you can get really good quality information, it’s much quicker to do it that way, so more timely.

ID 62, nurse lead

. . . the patients eventually, they’re going to be getting a copy of this . . . you don’t want to write anything that’s going to upset them. There’s a lot of debate at the moment about what we should be writing, and we’ve had a few little working groups about trying to standardise documentation.

ID 59, nurse lead

By the end of observations, SCRs were still not being shared with patients. Patients often had not been told they had been assigned a care co-ordinator and did not have their contact details.

Involvement of general practitioners

In terms of the factors identified by the ‘five simple rules’, one of the core issues facing the MDGs related to GP involvement. Initial participation was not always enthusiastic; not all practices took part, and identifying a GP to be the MDG lead was difficult for some practices. Incentives were required, although over time the involvement of GPs grew (as did that of practice nurses). The SICP delivered a somewhat unique neighbourhood model (compared with a practice-based one), which meant that GPs were exposed to MDG work around patients in other practices, which might be seen as inefficient, although it could have advantages.

Two MDG managers visited practices to explain the MDG concept and worked with practice managers and a GP lead at each site to encourage engagement. As a result, most practices in Salford agreed to participate, although some delayed joining owing to staffing and/or relocation issues:

I don’t remember being made to feel compelled to do it, I could be wrong . . . for me, we had to try it, so I went in thinking, ‘I’ll give you 6 months, at the very most I’ll give you 6 months.’. I went in with a very negative attitude. Fortunately, I didn’t allow my head to rule my heart and I started being a bit more open about it . . .

ID 85, practice staff

Practices agreeing to participate met together and underwent three training sessions initially. The first meeting introduced MDGs and how they operated. The second planned to carry out a Belbin assessment of all the team members to identify strengths and weaknesses and to facilitate better project working. At this second meeting the process to refer patients and the SCR were demonstrated. At the third training meeting, practices were each asked to identify one patient as a trial run for MDG discussion. In the MDG meetings that followed, practices were asked to refer other patients into the system (initially two or three patients per practice and then five patients per practice).

There were competing initiatives with these groups of patients, such as those run by the CCG:

[Long-term conditions] creates a lot of work. It’s a local initiative and the CCG want us to review patients twice yearly who have a LTC [long-term condition]. Discussion re[garding] symptoms, medication, health promotion advice and agree an action plan with them. If the individual has a respiratory condition or diabetes they receive an action plan for this and a LTC action plan. This is a lot of duplication and the process is time consuming. Reviews have become a tick-box exercise and the discussion about agreed actions is comprised due to time constraints.

ID 15, practice nurse

Measurement

The second simple rule related to measurement. Again, the MDGs had ready indices to assess progress and care delivery (such as numbers of SCRs). The initial risk stratification identified 3100 people potentially requiring discussion at MDGs, creating a tension between creating SCRs (to meet targets) and collecting data of sufficient quality to aid care co-ordination and planning.

One of the main ways in which learning was shared across the neighbourhoods was through the ‘MDG joint chairpersons’ meeting. MDG project managers led this meeting, which managed issues and facilitated learning. The number of level 3 patients with SCRs was fed back in terms of their achieving red, amber or green status. This feedback aimed to encourage a sense of competition between the neighbourhoods, but also highlighted difficulties when individual practices had not met the requirement of creating SCRs for 80% of their identified level 3 patients. In addition to this in-depth monitoring, which was updated on a weekly basis, ‘dashboards’ were created, which allowed interrogation of the data by individual practices.

Patient involvement

The NHS National Collaboration for Integrated Care and Support4 adopted a narrative definition of what integrated care should feel like for the patient [e.g. ‘I can plan my care with people who work together to understand me and my carer(s), allow me control, and bring together services to achieve the outcomes important to me’]. Recent qualitative work has highlighted some of the dimensions of this experience, identifying six themes around person-centeredness (holism, naming, heed, compassion, continuity of care, and agency and empowerment) and exploring the patient experience of being ‘unseen’ by their practitioners. 11

In principle, there are two key processes by which MDGs may help deliver patient-centred care. The first is care planning, which involves an assessment of the needs of the individual patient and the development of an individualised care plan reflecting the patient’s personal circumstances and needs. Second is care co-ordination, which is the process of helping patients to navigate services by enhancing communication and providing advocacy. Nevertheless, there are potential tensions between patient-centred care and MDG working. Core features of patient-centred care include detailed knowledge of the patient (and application to decision-making) and an effective ‘therapeutic relationship’, both of which may be more difficult to achieve in the context of team-based care. A patient seeing a variety of health professionals may feel less able to build up a relational continuity. 145 Patient involvement can be difficult in the context of MDGs and team-based clinical decision-making.

Multidisciplinary group guidance indicated that patients and carers should be contacted and a discussion had with them ahead of MDG discussions. In reality, the pressure to create a SCR and wider service pressures meant this rarely happened. Virtually all MDG staff interviewed felt that consent should be sought ahead of discussion:

I think the idea is that if they’re open to a particular worker within social services that we would have that discussion with them, if they’ve got district nurses going in regularly that they would or that if they are regular visitors to the GP that they would . . . I guess it’s sort of one of them where ideally the care co-ordinator would be in place and it would be that person that would then go and have that discussion with them and give them the feedback from the meeting and everything.

ID 60, social care lead

In December 2015, the CCG launched standards for services across practices, with guidance regarding consulting patients ahead of MDG discussions and feeding back outcomes. However, observations at MDG meetings showed that the discussions were of patients receiving feedback in the future.

We identified three core themes concerning how MDGs delivered patient-centred care:

1. structure of MDGs and person-centred care

2. processes within MDGs supporting person-centred care

3. barriers to the delivery of person-centred care.

The structure of multidisciplinary groups and person-centred care

The SICP used the ‘Sally Ford’ character to highlight the issues faced by older people. The multidisciplinary nature of MDGs provided a platform for a biopsychosocial assessment. By working together, health and social care professionals better appreciated each other’s roles. A second critical structural support was the creation of the SCR, which used data from the existing locally integrated primary, secondary and clinical settings, and into which social care staff could input data. The SCR was displayed during the discussion for each patient and updated with outcomes and actions arising.

Processes within multidisciplinary groups supporting person-centred care

All staff viewed the MDGs as an opportunity to share knowledge and provide a more holistic approach to care provision:

It’s bringing everything together so that we’re truly thinking more holistically about a person and considering not just individual kind of professional service-specific issues but seeing the overview and what might really matter to a person or what might be a priority and then coming up with an action plan that actually meets that.

ID 56, project manager

Although colocated previously, the arrangement whereby MDGs were co-chaired by staff from district nursing and social care provided greater insights into each other’s work and changes to practice:

. . . I’ve seen where there have been suspicions between health and social care, and crucially where people have different priorities . . . the pressure on hospital beds, which means people need to be discharged . . . Now having a better understanding around how the social care, how we do things that will help them to understand that it’s not as straightforward to just discharge somebody. We need to discharge and ensure that they are safe wherever they are going, and I think that wider integration [including psychological, medical and community services] can help address that.

ID 58, social care lead

Multidisciplinary group discussions yielded rapid solutions, particularly when local authority-commissioned social care packages required amendments. Small changes, termed ‘tinkering at the edges of care’,146 enabled patients to attend one-off hospital appointments and also regular exercise classes:

. . . we were saying about this fellow going at night to [a hospital] appointment, and I was saying about the care agency going in later, the social workers can arrange that; and she said, ‘but they finish at 8 o’clock’; so everything I was saying was being answered really negatively . . . But then we realised that maybe . . . the evening district nurses could go and put him in bed that night.

ID 70, nurse

Knowledge of patients’ interests helped MDGs find some solutions to frequent hospital attendance:

. . . one time the [MDG] nurse was in the house, [the patient] went in the other room and rang an ambulance, this gentleman. [Then] it’s somebody who just thought . . . ‘such a person’s not rang, what’s going on?’. [Then] . . . somebody realising that whenever, say, the tennis was on or the Olympics, he never went into A&E. He never rang A&E, he never went in. Then the practice nurse must have had a conversation with the son and he might have just said, he really loves sport. His son went out and bought him a Freeview box and it’s got sports channels on. Since that date, he’s never had an admission to A&E.

ID 63, administrator

A second process concerned MDGs sharing knowledge from health and social care perspectives about local services, encouraging a broader approach to meeting patient needs. Although time-consuming, MDG meetings provided a forum for professionals to learn about services available. GPs found it difficult to keep abreast of service changes, especially those delivered by non-statutory organisations:

. . . some of the other GPs . . . they’ve said, ‘oh well, I’ve learnt that there’s this service or that service.’. But they’ve learnt about maybe a handful of services in coming up to a year of . . . a part of the MDG you’ve got to ask is that actually a good educational use? . . . A lot of those services I knew about before . . . [It’s] hard to look at all your e-mails ’cause you get bombarded with so many, it’s easy to miss a service or you read one you think, ‘oh, that sounds fantastic’ and then you forget about it ’cause you won’t refer to it very much.

ID 79, GP

Multidisciplinary groups could also support person-centred care by involving those best known to the patient. A professional already known to the patient was usually assigned to the role of care co-ordinator, and the details of all those involved in their care were recorded in the SCR, pulling data from many agencies into a single source to aid care co-ordination.

Continuity of care was seen as vital to having effective knowledge of patients and their families, with GPs being seen as key. Despite this in-depth knowledge, the geriatricians were continuing to see patients face to face during assessments, and they questioned whether or not admissions can really be avoided. Geriatricians suggested that more proactive care planning, by GPs in particular, is required to reduce future admission rates and support more appropriate (rather than earlier) discharge for those who are hospitalised:

They [GPs] know them inside out, yet, they’re not necessarily thinking about what might tip them into, you know, non-elective emergency responses . . . but they [GPs] may not see that there’s any other alternative.

ID 89, geriatrician

Barriers to person-centred care

Although consultations during the development of the SICP had included patients and carers, MDG discussions did not directly include them:

. . . why wouldn’t they want to be involved? What is being discussed there that the person and their families wouldn’t find relevant, appropriate or interesting? . . . and is that right and what is the purpose of those [MDGs] if it’s, if it’s not to engage, you know, . . . [Is this] a process for just professionals to speak short hand about, together, about the integrated care management of a person and the contact with the person takes place elsewhere, possibly?

ID 81, third sector staff

Observations showed that although GPs fed back the outcome of MDG meetings to their patients, patients did not always agree to the suggestions made. This was especially true when MDGs recommended memory or social care assessments, which caused anxiety.

Aspects of MDGs limited the degree to which patient-centred care could be delivered. Nurses recognised the value of visiting patients at home, building a relationship and getting quality information to inform care planning, but flagged up difficulties with this when workload pressures limited the information supplied by colleagues:

. . . we need to be taking the patients that we’re seeing for MDG, to go and meet them, . . . to actually have a good picture of how they live and what’s going on, and the dynamics of families, and just get an overall picture. Because you can’t [get that] from looking at a screen.

ID 59, MDG nurse

Some questioned the utility of the SCR:

I can’t tell you how annoying I find that . . . It’s not a plan. There is no care plan. It is a record of information about a person . . . There are no directions on that record that give you a prescription for treatment. So it’s not a care plan . . . And as nurses, we pick up on that, they don’t.

ID 93, nurse

Towards the end of our observations, there was a change in MDG discussions, which coincided with the inclusion of a ‘summary of existing risks and issues’ on the SCR, which helped generate a more proactive care planning approach.

How is the work of the multidisciplinary groups experienced by patients and carers?

Patients and carers knew little about the SICP or the MDGs. Of those who did, many had been directly involved in the initiative through the Older People’s Partnership Board or the Citizens Reference Group.

This lack of awareness, combined with shared care plans not being shared with patients, meant that people were unlikely to know that they were part of a MDG. As patients were not routinely consulted ahead of MDG meetings, this limited the number of interviews that we were able to undertake.

Among those who were interviewed, there was limited evidence of active ‘case management’, and any care co-ordination seemed to focus on integrating records from health and social care via the SCR. Most patients could not identify a ‘care co-ordinator’.

Most GPs (during interviews and at MDG meetings) said that they would tell patients about discussions with colleagues if they saw the patient, but would not make a specific telephone call to discuss the MDG. If patients ended up being referred to health improvement services, then the GP would put the referral on hold while they confirmed that the patient was happy for their details to be shared:

The [GP] did mention it to me once I think, something about she’d been talking and she was going to put my name forward, something to do with this kind of thing [MDG], yeah . . . you can only give it a try and if you think it’s going to do you good well you’ll do it, won’t you?

ID 69, patient

Some patients don’t even know they are being discussed. I will try to contact patients or ask when I see them about discussing them at an MDG, but sometimes if we can’t contact them and there is a pressing issue that needs discussing we will have the meeting and then afterwards I will talk to them about it.

ID 15, practice nurse

There were examples of patient involvement in MDGs. An 80-year-old woman with a history of frequent A&E visits arising from repeated falls had lost confidence and was no longer going out of her flat, citing fears about using her stairlift. She gave consent to be discussed at MDG by her GP, who was concerned about the changes occurring:

I’ve known her for nearly 20 years, and she lives in a sheltered housing complex. She used to go and visit her mates [in another housing scheme], and get involved in all this social stuff . . . I also know she’d lost her confidence after various falls, and she wasn’t going out . . . she’s scared to leave her front door. I have a very good idea about what the patients I know very well want.

ID 70, GP

I had about five or six [falls], from last Christmas, in the flat and everywhere, you know, and I’ve lost my confidence . . . I’d left my [care on call] band in the bathroom . . . You know, if I fall I can ring them right away, . . . So sometimes I was sitting here 2 or 3 . . . about 3 hours before anybody came . . .

ID 69, MDG patient

Prior to the MDG discussion, the GP requested a change of housing. Once relocated to a bungalow in the same complex, she was assessed for, and received, mobility equipment from the rehabilitation team and was able to re-engage with other residents after the health improvement team invited all residents to attend a meeting.

It is possible that limited patient involvement limits the benefits of MDG work and the creation of relevant care plans. The process to refine the shared care plan, however, took far longer than anticipated, and there were many issues around what format the plan should be in so it could be shared with patients:

. . . we were discussing the logistics of how printing off a copy of the shared care plan for the patient and how that’s difficult at the moment. It doesn’t look nice. It’s not very ergonomic. It’s not very user friendly at all. In fact it’s horrible.

ID 55, project manager

. . . the patients eventually, they’re going to be getting a copy of this, . . . you don’t want to write anything that’s going to upset them. There’s a lot of debate at the moment about what we should be writing, and we’ve had a few little working groups about trying to standardise documentation.

ID 59, MDG nurse

Patients and carers were often overwhelmed by the sheer number of services they received. Where carers were aware of the MDGs and knew the appointed care co-ordinator, they were happy to remain involved in making arrangements:

[Mum] had visits from nurses, doctors, hospitals, GPs, physiotherapists, dieticians, intravenous teams, and others . . . care agency, et cetera, so her house is like Waterloo Station. It’s very busy . . . [If you’ve got a hospital visit] I phone [the care agency] usually a couple of days before, and say, my mum’s got a hospital visit on whatever day. Can you leave her in the wheelchair in the morning, and don’t do her lunch. I’ll do her lunch when I bring her back. So . . . and that always works. I don’t want to do that sort of stuff through a [care] co-ordinator. I’m happy to . . . on day-to-day issues like that, that’s fine.

ID 1, carer

Some carers felt too much was sometimes being done to patients in the NHS, which was unnecessary:

The NHS has a concept of ‘doing things to the older person’ . . . It pathologizes. It treats people as patients ‘cause that’s how they present rather than standing back a bit . . . older people tend to be hypochondriacs and they’re costing the NHS a lot of money. So rather than not constantly respond to them as patients, we might be sensible to stand back and actually talk to them a bit.

ID 33, carer

Regression results

Table 32 shows the results comparing all adopters with the four comparator sites. Graphs of trends for all outcomes are also provided (see Appendix 4, Figures 16–23).

We focus on the comparisons using the largest numbers of observations: all adopters compared with the rest of England. We find that the intervention sites showed increased numbers of A&E attendances, by 0.008 per quarter per person. Between 1 April 2015 and 31 March 2016, the intervention sites saw 1063 (95% CI 664 to 1462) more A&E attendances.

We find that the intervention sites showed a similar absolute effect on health and social care referrals to A&E, which also increased by 0.008 per person per quarter. This represents an increase of 1063 (95% CI 797 to 1462) in the number of attendances over the same period. Self-referred A&E attendances had a more modest increase of 0.002 per person per quarter, which represents an increase of 266 (95% CI 0 to 532) A&E attendances.

We find that the intervention was associated with an increase in emergency admissions of 0.003 per person per quarter, which represents an increase of 399 (95% CI 133 to 664) additional emergency admissions between 1 April 2015 and 31 March 2016. Emergency admissions via A&E increased by 0.6%, an increase of 797 (95% CI 399 to 1063) admissions. We found a fall in direct emergency admissions by 0.001 per person per quarter, resulting in a fall of 133 (95% CI 0 to 133) admissions. The effect of the intervention on avoidable emergency admissions is not statistically significant.

We found that the intervention led to a decrease in the number of patients who are discharged to the usual place of residence of 0.005%. The intervention resulted in a larger relative decrease in the number of patients returning to their usual place of residence among early implementers (0.012%) than among late implementers (0.004%).

Results using Greater Manchester, West and West (extended) show similar results to using all of England as the comparator group, when focusing on all results from all implementers. The signs of the coefficients remain the same for all estimated effects of the intervention along with similar absolute magnitudes. The small differences in estimate effects are explained by the changes in the sample size.

Robustness checks

Reducing the pre-intervention period by removing the first four and first eight financial quarters of the sample produced results that were consistent with the results obtained using data from the full sample (Table 33). We find that emergency admissions increased by 0.003 and 0.004 per person per quarter for models removing the first four and eight quarters, respectively.

The results from the difference-in-differences estimator were not statistically significant. However, the point estimate of the effect of the intervention using all of England as the comparator group is identical to the main results. Therefore, both difference-in-differences and lagged dependent variable approaches resulted in impacts of similar magnitude, but the lagged dependent variable generates a more precise estimate.

The process of health coaching in patients who consented

The defined ‘dose’ of the PROTECTS intervention was monthly calls of around 20 minutes for a period of 6 months, with the option for additional calls to deal with complex patients or issues of risk.

A total of 189 out of 504 (38%) people who were randomised to the health coaching intervention agreed to take part and were referred to the health coaching team from July 2015 to March 2016. In terms of adherence, 167 out of 189 (88%) participants who consented to the intervention received four or more telephone calls.

Economic results

Complete-case analysis requires data on age and sex at baseline, and EQ-5D-5L responses and resource utilisation at all four time points. Complete data necessary for the economic analysis were available for 45% of the sample (584/1306).

Table 38 presents the baseline characteristics of the full sample compared with the complete-case sample.

Resource utilisation and costs: health coaching

The resources required to deliver health coaching are presented in Table 39. They consisted of training sessions on intervention delivery, staff supervision, information materials for staff and patients, and the cost of coaching sessions.

The main training session was delivered by project staff and lasted 2 full days. A second half-day training session focusing on long-term conditions was delivered by a GP. There were five additional training sessions, again delivered by project staff, each lasting 90 minutes. Thirteen 1-hour supervision sessions were conducted with project staff. Manuals and scripts for health coaches and information sheets and notes pages were printed for participants. Telephone coaching calls lasted between 15 and 25 minutes, with an additional 10–15 minutes of staff preparation before or after the call. On average, each participant had one unanswered or rearranged call, which utilised an additional 10 minutes of staff time. The average cost per individual receiving the full course of health coaching (six calls) was £148.27, of which £44.38 related to training, supervision and paperwork and £103.89 related to delivery of health coaching.

Additional resource utilisation over trial follow-up

In addition to the direct costs of the health coaching, the economic analysis also considered wider NHS resource utilisation. Table 40 reports the average utilisation by resource category for the complete-case sample, separated by treatment arm, and summarised as the mean number of contacts with each service per individual.

Although there was variation in the use of services over time and between treatment arms, some consistent patterns can be seen. The most frequently utilised category across both treatment groups was outpatient appointments, followed by GP appointments.

The average number of emergency admissions (short stays) and A&E attendances was lower for the health coaching arm at all three time points, with the differences between the treatment arms increasing over time. The health coaching arm also had lower emergency admissions (long stays) and day cases at all three time points. Elective admissions and outpatient appointments were higher at all three time points among the health coaching arm, with the difference in outpatient appointments increasing over time.

Resource use was higher at 18 months than it had been at 12 months for all resource use categories in both treatment arms. However, in the health coaching arm, emergency admissions, day cases, outpatient appointments and A&E attendances fell at 12 months compared with 6 months, before rising again. Although emergency admissions (short stays), day cases and outpatient appointments also fell in the usual care arm, no such drop was observed for emergency admissions (long stays) or A&E attendances, with use rising at each consecutive time point in the usual care arm.

Overall, there was a pattern of greater use of emergency care among the usual care arm, whereas health coaching patients utilised more planned services.

To cost both arms, unit costs were applied to individual utilisation. Table 41 presents a list of all unit costs used, together with their source. Table 42 combines the resource utilisation of the complete-case sample with the unit costs to obtain the average costs associated with this resource utilisation over the trial period.

The most costly category of resource use was outpatient appointments, followed by elective admissions and GP appointments. These are all planned care services, the costs of which were higher in the health coaching group. Conversely, the costs of emergency admissions (short and long stays), day cases and A&E attendances were higher in the usual care group. Overall, mean costs were higher in the health coaching group (£4000.88) than in the usual care group (£3424.16).

Intervention costs in the health coaching averaged £79.29. This was lower than the £148.27 estimated for full treatment because not all individuals took up the offer or completed the full course.

Outcomes

Table 43 summarises the EQ-5D-5L utility scores at each time point and the total QALYs gained over the 18-month follow-up period for the complete-case sample. On average, patients who received health coaching reported slightly lower health-related quality of life at baseline than those who received usual care (mean utility 0.696 and 0.708, respectively). This value fell steadily over the follow-up period in the usual care group, reaching 0.664 at the 18-month follow-up, but in the health coaching group remained relatively unchanged at the end of follow-up, at 0.691. The mean unadjusted QALYs for usual care and health coaching over the study period were 1.105 and 1.124, respectively.

Cost-effectiveness analysis: full sample with imputation

Table 44 presents the adjusted estimates of the effect of the offer of health coaching on the incremental costs and QALYs over and above usual care in the full sample with imputed data. This analysis controls for age, sex and baseline utility. Health coaching is associated with a mean incremental total cost increase of £150.583 (95% CI –£470.611 to £711.776) and a mean incremental QALY gain of 0.019 (95% CI –0.006 to 0.043).

Although there are no statistically significant differences in either costs or QALYs, the point estimate of the ICER is £8049.96 per QALY. This would represent a cost-effective intervention with respect to the standard cost per QALY threshold of £20,000–30,000. However, it is important to consider the uncertainty surrounding this estimate.

The cost-effectiveness plane plots the 10,000 bootstrap replications of incremental cost and QALY estimates to illustrate uncertainty in probabilistic terms (Figure 9). The replications are clustered predominantly in the north-east quadrant, reflecting a positive health gain at an increased cost. Health coaching resulted in an incremental QALY gain in 94% of bootstrap replications and costs were higher than usual care in 69% of replications.

FIGURE 9.

Cost-effectiveness plane: full sample with imputed data. Reproduced from Panagioti et al. 147 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Reproduced from Panagioti et al. 147 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/

The CEAC (Figure 10) demonstrates how the probability that health coaching is cost-effective increases with the decision-maker’s willingness to pay. At the lower bound threshold of £20,000 per QALY, there is a 70.2% probability of health coaching being cost-effective. This rises to 79.0% at the upper bound of £30,000. Compared with usual care, health coaching is likely to be cost-effective in ≥ 50% cases if decision-makers are willing to pay ≥ £8180 for 1 QALY.

FIGURE 10.

Cost-effectiveness acceptability curve: full sample with imputed data. Reproduced from Panagioti et al. 147 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Reproduced from Panagioti et al. 147 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/

Cost-effectiveness analysis: complete-case analysis

Table 45 presents the adjusted estimates of the effect of the offer of health coaching on the incremental costs and QALYs over and above usual care in the complete-case sample.

Health coaching is associated with a mean incremental cost of £497.99 (95% CI –£189.19 to £1185.19). Although higher than the incremental cost associated with health coaching in the full sample with imputation, this difference in cost between arms still does not reach statistical significance. The incremental QALY estimate is also higher, at 0.037 (95% CI 0.0037 to 0.070), indicating that there was a significant QALY gain associated with health coaching among the complete-case sample. This results in an ICER of £13,506.27 per QALY, which is again below the lower bound of recommended cost-effectiveness thresholds.

It is again important to examine the uncertainty around these point estimates in probabilistic terms.

Figure 11 shows that the bootstrapped replications are again clustered predominantly in the north-east quadrant. There is notable shift north above the x-axis and to the right of the y-axis compared with the full sample with imputation, as even more of the replications now fall into this quadrant, representing incremental increases in both costs and QALYs above usual care. Health coaching was more costly in 92.5% of replications and led to an incremental QALY gain in 98.6% of replications. The CEAC illustrates that, at the lower bound (£20,000/QALY), there is a 67.2% probability that health coaching is cost-effective, rising to 82.0% at the upper bound of £30,000 per QALY (Figure 12). Compared with usual care, health coaching is likely to be cost-effective in ≥ 50% cases if decision-makers are willing to pay ≥ £13,570 for 1 QALY.

FIGURE 11.

Cost-effectiveness plane: complete-case sample.

FIGURE 12.

Cost-effectiveness acceptability curve: complete-case sample.

Implementation of the mechanisms of integration in the SICP

As noted earlier, the progress on structural integration through the SICP to the ICO was not closely connected to the delivery of integrated care services, and the various management processes put in place by the SICP ensured that progress continued on delivery. The next section will consider how much progress was achieved in delivering the planned mechanisms of integration.

The MDGs were the largest mechanism in terms of scope and resource, and the most rapidly implemented (with some delays around planned pilots and ‘tests of change’). There were some significant challenges in implementation (owing to issues with engaging with GPs), dealt with in time through relevant contractual arrangements. SICP staff involved in the MDG process reported that consistent input from all professional groups was achieved around August 2016, when it might be said that the complete model was in place, placing it towards the end of the CLASSIC study timeline and thus limiting our ability to assess impact of the complete model given the data that were available.

Multidisciplinary group meetings were generally well attended, and staff were broadly positive about the model. They reported some issues with what was perceived to be slow progress, as well as a focus on patients at certain levels of need who may have fewer opportunities to experience proactive care. These concerns echoed the general debate about risk stratification of patients and about where resources are best placed to generate real gains in health and reductions in utilisation. 78

Early implementation in MDGs focused on process measures (such as creation of records). Actions arising from the MDGs were often relatively limited, echoing similar interventions elsewhere in the UK. 153 Although ‘care co-ordinators’ were allocated to patients being discussed by MDGs, interviews with patients and carers showed that MDG discussions were taking place without their involvement. As MDGs were quickly implemented across the locality, we could not conduct a formal experiment. We compared admissions in Salford with those in other areas, which showed little difference from the national trend, with the strongest evidence suggesting an increase in emergency admissions. The SICP is not the only integrated care initiative to report such outcomes. 33 There is a legitimate question as to when the MDGs could reasonably expect to see effects on utilisation, and the timing of the full implementation may have reduced the ability to show benefits of MDGs in the CLASSIC timeline. Further analysis can continue to explore utilisation to see if the initial effects change over time.

The ICC faced some significant challenges in implementation and clarity of vision. Patients’ experiences of the ICC were mixed, which in part reflected the need to adapt to new ways of engaging with services. It was not clear that the centre was functioning to enhance patient experience of an integrated service, although colocation was noted by staff to be a potentially important mechanism of integration in terms of understanding roles and identifying new possibilities for support, which suggests that initial benefits may be focused on staff rather than patients. 153 Such benefits in the process of care are likely to show benefits in patient outcomes over the longer term, if at all.

We conducted a formal trial of a health coaching intervention within the ICC. Patients using the service reported that it was acceptable and useful. Impacts on self-management and self-reported health were small, but the economic analysis suggested a reasonable probability that the intervention was cost-effective at conventional levels of willingness to pay. However, the intervention delivered improved health-related quality of life at an additional cost, and did not reduce utilisation overall (although the pattern of use suggested less use of emergency care and more use of planned appointments).

The SICP sought to increase access to community assets, and we saw a small increase in reported use of assets plausibly attributable to the SICP. Asset use was associated with better quality of life, when accounting for other factors, but impacts on utilisation were not significant. As with coaching, and in line with other quality improvement projects in integrated care, improving outcomes seems easier to deliver than reductions in utilisation. 154,155

Overall, the SICP made some modest improvements in outcomes to some patients, including patient experience and quality of life, although the numbers impacted were relatively small and the magnitude of the benefits was limited. There was much less evidence that the programme was leading to reductions in health-care utilisation.

Interpretation of the findings: overall

As noted in Chapter 1, there is a fairly large body of international literature on integrated care, and this has demonstrated some benefits of this model, although effects are somewhat inconsistent and impacts on economic outcomes are difficult to demonstrate consistently. However, there are issues concerning the interpretation of that very broad set of studies and applying the results to the particular context of the NHS. Empirical studies of integrated care in the NHS have demonstrated more modest impacts. A key motivation for the present study was to explore whether or not a particular model of integration (the SICP) delivered in a receptive context49 could demonstrate improved outcomes. Overall, the results suggested that the receptive local context facilitated moves to structural integration, but it is not clear if benefits translated to the delivery of the specific mechanisms of integration, which faced the challenges and delays that have been reported in other evaluations. 36,37 There were initial indicators of benefits among a modest proportion of patients (increases in reports of community assets and care plans, themselves associated with improved outcomes).

Interpretation of the findings in the context of the wider literature: multidisciplinary groups

Multidisciplinary groups (and linked case management interventions) have a peculiar place as mechanisms of integration in the NHS. They are very common, with a survey of CCG plans finding that around 80% included some form of MDG. 79 Nevertheless, since Evercare,29,30 evidence that this model can reliably reduce hospital admissions is weak. 18,19,33,77,78,156 The recent National Audit Office report stated that:

While popular approaches, such as multi-disciplinary teams focusing on patients with multiple and complex needs, may improve the care experience for a minority of patients, the evidence to date does not suggest that they will achieve the widespread efficiencies and outcomes needed in the current financially constrained times.

Reproduced with permission from the National Audit Office25

A realist model would hypothesise that the mechanism of MDGs might activate only in certain contexts. We identified a number of contextual factors that might be present in the SICP, including the partnership underlying the SICP, high-quality IT, the potential for self-management support via the interface between MDGs and other mechanisms of integration in the SICP, support for MDG development through quality improvement and ‘tests of change’, and the use of a ‘neighbourhood’ model for MDGs. Although many of these contextual facilitators are plausible effect modifiers, there is no strong evidence either way as to their actual impact.

The quantitative analysis was unambiguous in showing that the introduction of the MDGs failed to reduce admissions, with the strongest evidence showing an increase. This core result was robust to a variety of analytical models, comparators and the staging of the MDGs, and reflects similar outcomes found in previous work. 33

There are a number of possibilities that account for this lack of effect:

1. MDGs are not effective. As noted, the lack of effect demonstrated here is not unusual. 19 There are potential competing interventions to reduce admissions, such as enhanced continuity of care,157 although they are more difficult to test in a controlled fashion.

2. MDGs are effective, but are so widespread that showing differences for an intervention in any one area compared with comparators is difficult (as noted, most CCGs include the intervention in some form). The SICP MDGs may be unable to outperform similar interventions elsewhere, rather than being ineffective per se.

3. The contextual facilitators we describe are incorrect or ineffective. We did not find strong evidence that IT was a strong facilitator of the MDGs, linkage to other parts of the SICP (such as community assets) to support self-management was limited, GP engagement was variable and the advantages of the neighbourhood model not fully articulated or observed.

4. MDGs are effective if targeted. Critics suggest that excessive focus on ‘high utilisers’ means that interventions are provided to a patient group that is both too small to show meaningful change in admissions and impervious to change because of the severity of their conditions. 78 A variety of risk algorithms have been used worldwide, but these were not adopted here, which may have attenuated effects, although there is little strong evidence that particular algorithms are effective. Targeting of specific groups and specific contexts, such as hospital-led case management for heart failure, may increase impact but clearly restricts the scope of benefits. 20 It is interesting to note that the health coaching intervention (which is based on an assessment of patient activation, not future risk of admission) actually led to reductions in emergency care use, providing some support to the idea that a focus on prevention among larger groups of lower-risk patients may be an alternative strategy (although the evidence here is also weak).

5. The evaluation of MDGs was too early to show benefits. A significant focus of the early work on MDGs was on forming the groups, identifying patients, developing ways of working and organising required systems. The numbers of patients managed by MDGs and the various performance indicators in place may have led to a focus on certain aspects (such as the creation of records), to the detriment of others (such as referral to other parts of the SICP to support self-management). Analysis of the work of similar teams in north-west London also found improved collaboration, but a lack of actions to take forward,158 as well as suggestions that the benefits might be more apparent for professionals (via shared learning) than patients. 153 There was some evidence that MDG function improved over time. Improved patient outcomes may become apparent only later.

6. Patient involvement in MDGs was insufficient. There were aspirations for patients to be involved (such as being informed that they were subject to a discussions and involvement in care planning). However, the numbers of patients to be managed and pressure of performance targets may have limited scope for patient involvement. There may be aspects of integrated care that can be achieved without significant patient involvement (such as better co-ordination between agencies) and through which patient benefits may accrue, even if integration is ‘behind the scenes’. The empirical evidence that the effects of MDGs are multiplied by effective patient involvement is not strong. However, there is an argument that some of the potential benefits of MDGs are lost without that involvement. Recent research has outlined some of the ways in which they could contribute and the barriers to them doing so. 159

Interpretation of the findings in the context of the wider literature: health coaching

Compared with MDGs, health coaching is focused on patients lower down the ‘risk pyramid’, with a simpler intervention focused on improving self-management. The benefits of such an intervention on high-cost health-care utilisation (such as admissions) is likely to be deferred in time, as patients are not at high risk. However, such prevention models may be more effective in the longer term, as the numbers of patients in this group are so much higher and their potential to benefit from intervention may be greater. 160,161

Of course, realising those benefits requires two features: (1) a significant level of uptake of the intervention among patients with needs who are not necessarily seeking help and (2) enduring impact on health attitudes and behaviour which will translate to longer-term benefits, especially in relation to care utilisation.

We did not demonstrate significant impacts on patient-reported outcomes, despite largely positive assessments from patients in the qualitative research. An important limit on effectiveness was the low uptake among patients. In terms of uptake, the levels found (40%) are reasonable, given that this was not a help-seeking group and the invitation came from outside usual services routes (such as GPs). This suggests a willingness to take part in such ‘preventative’ interventions among an older population with long-term conditions. It is, of course, important to remember that the CLASSIC cohort was itself selected, in that patients had to agree to be part of it.

The cmRCT estimates the effect of being selected for treatment, regardless of whether or not that offer is taken up. This is different from a conventional trial, in which patients who do not give consent are excluded from the trial and the effect being estimated is that of randomisation to treatment. Therefore, demonstrating effectiveness in a cmRCT is even more challenging than in a conventional pragmatic trial, as the cmRCT faces loss of potency through variability in quality of delivery and adherence as well as overall uptake, as opposed to the former alone in a pragmatic trial.

We argue that the treatment effect estimated is an appropriate one for health coaching, as it is a model of integrated care designed to be proactive, using identification of patients in the community according to external criteria rather than their own help-seeking, and designed to achieve population benefit. However, it cannot be assumed that the rates of consent in PROTECTS would generalise to delivery in routine practice. 162

The economic analysis suggested that the intervention was likely to be cost-effective, delivering modest improvements in quality of life at some increased overall cost. The methods of the cost-effectiveness analysis are not based on notions of clinical or statistical significance. Decisions to adopt one intervention over another are based on the expected cost-effectiveness of the interventions and the probability of making the correct decision. 163

Interpretation of the findings in the context of the wider literature: community assets

Although there is a consensus that community assets improve health, the experimental literature around assets and related interventions, such as social prescribing, is limited. 164 We found a small increase in use of assets among cohort participants and found that increases in asset use were related to improvements in quality of life, but not care utilisation. It is possible that benefits in quality of life have a longer-term impact on utilisation, in line with the prevention model underlying some discussions of integrated care. 160

Strengths and weaknesses of the CLASSIC study

There is an increasing recognition that the scale and pace of reform and reorganisation in the NHS calls for new methods of evaluation that balance rigour with speed and responsiveness. 41

Our broad methods were conventional, using routine data and non-randomised comparisons to assess causal relationships, allied to detailed descriptive and qualitative research to provide depth, explore context and assess mechanisms. 29,30,33,51

The use of the cmRCT represented a methodological innovation. The cohort provided an ongoing assessment of patient experience (absent from assessments based on routine data only), eased recruitment of the trial sample and allowed modelling of the impact of mechanisms of integration, such as community assets, by providing a measurement framework into which interventions were introduced, rather than measurement following the intervention. The cohort also provided the opportunity to produce a variety of papers related to the general management of patients with multiple long-term conditions. 114,115 Attempts were made to run additional trials, but funding bids were not successful.

It had been hoped that mechanisms of integration in the SICP could have been allocated in ways that allowed more scope for causal inference (such as ‘stepped’ allocation). In reality, the practical demands of the programme were such that the only ‘staging’ of change involved the early introduction of the MDGs in two neighbourhoods, driven by a desire to pilot the service.

The challenges faced in the delivery of the SICP are not dissimilar to many of these reported in the wider literature, with slower than anticipated implementation a common feature of integrated care evaluations that causes difficulties for the timing of any evaluation. 165 As noted previously, the impact of investment in community assets, delivery of care plans and significant engagement by MDGs with large numbers of patients had yet to materialise, and thus clear evidence of benefit may be deferred. As noted in Chapter 2, the formal ‘end date’ for evaluation of the SICP is 2019–20. The original CLASSIC bid was for 5 years, and extending the evaluation could assess longer-term effects.

We took a population perspective, exploring the effects of the mechanisms of integration on a large cohort (for self-reported outcomes) and the population of Salford (for hospital utilisation), rather than a more restrictive view exploring effects on those patients who were specifically targeted by particular SICP mechanisms. The latter approach may be more likely to show impacts, as there is a direct link between the mechanisms and the outcome. However, the restricted view would ignore wider effects; for example, MDGs might generate benefits by focusing on a sample of patients, with worse outcomes among those who are not targeted and thus receive less attention.

Although some outcomes may take time to become apparent, some previous evaluations of integrated care have reported impacts on issues such as patient experience at an earlier stage. 33 We found only limited evidence of such impacts here, with the most notable change being in reports of written care plans and access to community assets. Although there were initial concerns among the research team that the SICP may have been implemented before the necessary ethical and governance arrangements were in place for CLASSIC, in reality, the introduction of CLASSIC proceeded at a pace that may have been too rapid for the benefits of the SICP to be evident.

The combination of the population perspective and the timing of the SICP and CLASSIC mean that the limited impacts shown at this point are provisional; the SICP has not demonstrated significant population-level impacts at this point in its development. It remains to be seen if those effects will become apparent in time.

The CLASSIC study does represent a single-site evaluation, with consequent limits on external validity. We have provided as detailed a description as possible of the local context (within the limits of the report format) to enhance interpretation. Of course, there is no consensus on the numbers of sites and amount of variation between sites that would allow stronger statements about generalisability in the context of ‘large-scale transformation’.

The study was premised to a degree on the idea that the local context provided a particularly fruitful one for integration. The health and social care system was largely coterminous (with one council, one commissioner, one mental health provider and one principal provider of acute and community health), and included a long history of partnership working and an integrated care record. If the overall outcomes had been more positive, then a debate would have ensued about the ability to replicate it in other settings. Given the more modest impacts, despite the context, the interpretative task is different. In the realist model, mechanisms of integration will show more or less potency in certain contexts. We have outlined the ways in which the chosen site was one in which integration was enabled by history and geography. The fact that there was only limited evidence of improved patient outcomes could be seen as a challenge for the realist model, as mechanisms of integration should arguably be more potent – more modest effects might be expected in areas with a less amenable context, such as multiple providers and more complex interorganisational relationships. This will be most evident in comparative data that uses external locations, as external sites may potentially lack both the ‘mechanism of integration’ and the contextual enabler. Of course, an alternative interpretation is that the effective mechanisms of integration are yet to be identified.

Meaning of the study: possible explanations and implications for clinicians and policy-makers

The SICP has been implemented consistent with the original vision. However, there has been more rapid success in establishing new integrated structures, rather than delivering mechanisms of integration at the level of services which impact on patients.

As with many integrated care transformations, it has proved challenging to deliver transformation in care, which means that the evaluation data reported here may be premature. It is still not clear if greater improvements in outcome will flow from integrated structures.

The failure to evidence improvements in patient outcomes may reflect a number of issues. Although the chosen mechanisms of integration were broadly supported by existing reviews (in the case of health coaching), or consistent with wider service activity (in the case of MDGs), there is still limited evidence about each. The ‘decoupling’ of the system and service-level integration means that the SICP is not wedded to these particular mechanisms. The potential for improved integration in the ICO (such as staff working in a common organisation) could support other mechanisms of integration.

The data would suggest that there is an urgent need to develop and evaluate new mechanisms that could be deployed from integrated organisations. It is also possible that more modest ambitions than the ‘triple aim’ (better outcomes, enhanced experience, lower costs) need to be set for those interventions, given the difficulties in achieving health gain without increasing cost. 24 Other writers have emphasised more radical changes to the approach, such as an additional focus on prevention,79 and the recent plan for health and social care under the devolved arrangements in Greater Manchester suggests a preventative focus, allied to a drive to ‘enhance our primary care services, with local GPs driving new models of care and local care organisations forming to include community, social care, acute, mental health services, the full range of third-sector providers and other local providers such as schools’. 161 The Salford vanguard (see Chapter 10) is one of the new organisations mooted in the plan.

The SICP has introduced new services, but it is not clear that the scale or ‘dose’ of the programme is sufficient to have an impact on wider population health. To talk of a ‘dose’ of integrated care may seem inappropriate, as it suggests that the complexity of the SICP can be reduced to a simple measure. Nevertheless, considerations of ‘dose’ may have some utility. Box 3 outlines some relevant data. These are crude indicators of ‘dose’, and ignore wider changes in the structure and organisation of care that may not be immediately apparent or easily measured. Examples would include MDGs that impact on general ways of working, with the potential to impact on all patients. However, some issues are worth noting. The programme that involved the largest numbers of patients (community assets, of relevance to all patients and most relevant to the larger numbers of patients at lower levels) showed only limited evidence of direct engagement with patients in the study period. This is supported by analyses in Chapter 9, which showed that large numbers of older people reported engagement with assets at baseline (unlikely to have been affected by the SICP), with only small changes over the timeline of the programme. The numbers directly affected by the ICC were more significant, and the impacts were more likely to translate to immediate benefit. Nevertheless, the proportion of patients was still relatively small. Finally, the MDGs delivered a ‘mechanism of integration’ to a relatively large number of patients, especially in proportion to those at levels 2 and 3 who were eligible. Nevertheless, as detailed earlier, actual MDG activity with respect to many patients was limited, with minimal patient involvement and small changes in care delivery.

Although the data related to dose are only indicative and do not cover all the changes made as part of the SICP, they suggest that, although many older people may have received some form of integration, the dose at the level of the population may have been limited.

Unanswered questions and future research

Demonstrating significant health benefits from integrated care interventions remains a significant challenge. In part this represents difficulties in defining integration of care in a way that makes it amenable to evaluation. When integrated care programmes are part of a larger-scale transformation of services, there are additional complexities around the timing of evaluation. The enormous challenges associated with integration at scale do suggest that benefits may be deferred. However, if evaluation is delayed, then services may already be so committed to a model that responding to the results of the evaluation may be difficult. Identification of comparators (and keeping them from adopting new services rapidly) is also made more problematic.

Some of the results reported here are in line with other evaluations of integrated care: slower than anticipated progress on aspects of the programme36,37 and limited impacts of interventions on patient experience and admissions (including increases in some outcomes). 33 In Chapter 2, we hypothesised that context would be a significant enabler in the SICP, given the long history of joint working, but the context seems to have facilitated the speed of higher-level organisational change, rather than enhancing the delivery of mechanisms of integration and consequent impact on patient outcomes. Although there is consensus in the health services research literature as to the importance of context,166,167 the ways in which it interacts with discrete interventions and impacts on patient outcomes are less well understood. A recent exploration of the context–mechanism–outcome relationships in integrated care for diabetes mellitus reported many barriers involving organisational issues (such as deficits in IT and provider self-management support skills), whereas facilitators more often involved social issues (such as involvement in staff in decision-making, leadership and culture). 168

The future research questions around MDGs are complex, as there is now a weight of evidence suggesting limited benefits, despite the popularity of the model. 79 Within the SICP, and integration initiatives more generally,36,37 the usual patient under consideration is an older patient with multimorbidity. One of the barriers faced by integration initiatives is that, despite large amounts of evidence about the challenges faced by these patients,169 the evidence base concerning interventions that demonstrably improve the care of multimorbidity is very poor,98 as highlighted by the recent NICE guidelines,170 with consequent funding calls from NIHR. Indeed, there is a debate about whether the concept of multimorbidity is optimal, or whether a focus on frailty would be better, as the populations captured under these terms are related but not identical. 171 The success or otherwise of integrated care may be dependent on the development of that evidence base.

It is possible that the model needs improved targeting of patients or disease subgroups to demonstrate effectiveness, although there is the challenge of identifying suitable subgroups. Although the targeting of patients at high risk has face validity, effectiveness will be dependent on the ability of services to make significant improvements to their care without increasing the risk of other patients who are not targeted. The best way to undertake risk stratification and the types of patients who should be targeted remains controversial,78 and the field awaits the results of trials of different methods. 172

One of the limitations of risk stratification models is that they only identify risk and do not necessarily drive intervention (which has been called the ‘impact’ of such prognostic tools). 173 There is clearly a role for further detailed qualitative examination of the actions generated by such multidisciplinary working, given the findings of this study and those of others that suggested limited impacts on patient care. 132,153 It is important to explore if these limited impacts reflect the nature of work in these groups, the amount of support that these patients are already receiving and the limitations in available resources to manage problems that have been identified through multidisciplinary meetings. Nevertheless, the weight of evidence would suggest a more limited role for the model in integrated care and the consequent need to identify alternative mechanisms.

Health coaching is in a somewhat similar position as an evidence-based intervention, as early reviews of small studies in particular groups showed promise,85,87 whereas larger, more pragmatic evaluations suggested more modest gains. 90 Overall, the evidence seems stronger than that for MDGs, although the search for a reliable model of coaching that can be implemented at scale remains. The benefits reported here, although cost-effective at conventional levels of willingness to pay, would not translate to savings for commissioners or providers. This means that the model may have a limited role in achieving the sorts of reductions in health-care utilisation that are critical to many integrated care initiatives. Nevertheless, the positive patient experience (as reported in Appendix 2) and the changes in patterns of resource utilisation (see Chapter 13) should not be discounted.

The data suggest a role for community assets in generation of gains in quality of life. Further quantitative exploration is required to assess longer-term benefits, especially longer-term impacts on utilisation, which may be generated by short-term improvements in quality of life. It would also be useful to explore the patient and contextual factors that drive asset use and that influence the level of benefit derived by older patients. For example, many social prescribing schemes have a focus on mental health, in terms of both inclusion criteria and measurement of outcomes. Appropriate qualitative methods might be employed to explore the factors underlying access to assets (beyond the effects of proximity explored in Chapter 9), and the psychological and social mechanisms by which they achieve their effects (such as improvements in self-efficacy or activation, reductions in loneliness or improved social support).

Community asset interventions as described here are linked to social prescribing interventions, which have received recent interest from policy-makers,174 although the evidence base remains weak. 164 Social prescribing schemes are generally based on a GP referral mechanism, which differs from the broader community assets intervention employed in the SICP. The relative impact of the different schemes could be usefully explored. Referral through a GP may well increase access to assets in principle, and provide greater linkage to other aspects of management of long-term conditions through an effective primary care-based care planning process. However, enthusiasm for such schemes may be highly variable among GPs. There is a useful debate to be had about the amenability of such schemes to formal test, compared with the quasi-experimental models employed here, but examples of randomised evaluation do exist. 175

As noted previously, it remains an open question whether or not health impacts from community assets convert to genuine savings in care utilisation. Although the ‘triple aim’ remains part of the vision of integrated care, the pattern of results reported here (achieving improved outcomes at increased cost) is less likely to be sustainable.

The use of the cmRCT design in the CLASSIC programme has identified a number of issues with the design, including the management of low levels of uptake among patients and the logistics of follow-up. As more trials utilise the design, synthesis of the findings will be required to better understand the advantages and disadvantages of the model and how it can best be used to meet the challenges of evaluation in the NHS in years to come. 41

Although many integrated care initiatives share common mechanisms (especially variants of the MDG), there is no reason why a wider variety of interventions might be employed and tested within the context of integrated services. For example, many integration initiatives have an initial focus on prevention36,37 and future work could explore preventative interventions that could be facilitated through integrated services to improve the outcomes of older people.

Main analysis

The main analysis is principally concerned to test for a treatment effect between the two trial arms on each primary and secondary outcome variable at the 20-month follow-up.

Sample selection

The sample size and power calculation described above were undertaken under the assumption that all 252 patients offered the treatment would consent to receive it, whereas only 100 (40%) actually did. To account for this, the treatment was offered to a further ‘top-up’ sample of 252 patients of which 107 consented, producing an overall consent rate of 41% (207/504). Under the cmRCT design all 504 patients selected for treatment remain in the treatment arm for analysis. This includes 51 patients who could not be contacted and hence were never offered the intervention. To exclude these would risk introducing bias.

Definition of baseline and principal follow-up

The baseline for all patients will be time zero and the measures taken at that time point. Although no patient received treatment until after the 6-month follow-up, the measures at that point were post-randomisation for half of the sample and therefore potentially affected by arm allocation. The principal follow-up will be the 20-month time point. This will ensure that all patients across both arms have equal length of time duration between baseline and follow-up. We consider this to be the approach with least chance of selection bias under the null hypothesis.

Timing of treatment

Due to implementation delays, no patient was offered treatment before 6 months after the baseline assessment and for some the offer was not made until month 12 or later. This caused variations in the duration of time before start of the treatment, ranging from 259 to 513 days. Length of follow-up from end of treatment to 20-month follow-up was similarly variable. Thus the trial is considered to have run over the 20-month period, with individual patients receiving treatment at any time within that period.

Treatment effect estimation

The cmRCT design provides an estimate of the mean effect in people offered treatment. Compared to a pragmatic CRT, which provides an estimate of the mean effect in people agreeing to treatment, the effect is ‘diluted’ by the proportion of patients in the treatment arm who do not consent to treatment. An estimate of the effect size in patients consenting to treatment can be obtained through application of a Complier Average Causal Effect (CACE) analysis, and we will undertake that here. We note however, that CACE does not increase the power to detect an effect.

Research hypotheses to be tested

There are no overall significant differences in the primary outcomes between intervention and control groups at 20 months.

• Patient Activation Measure. A standardised spreadsheet in excel is used to score the PAM. Overall score (0–100) on the PAM scale where higher scores indicate high patient activation. A total score will be generated where participants answer at least 10 out of the 13 PAM questions. The continuous scores are categorised into four levels for descriptive purposes using the standardised spreadsheet. Higher scores indicate higher patient activation.

• Quality of life – physical health domain. Physical health domain of the World Health Organization Quality of Life instrument (WHOQOL-100). WHOQOL is a 26-item measure which includes two items on general overall QoL and health, plus 24 which are scored in four QoL domains: physical, psychological, social relationships and environmental QoL. Facets are scored from 1 to 5, and higher scores indicate better QoL. Domain scores are transformed onto a scale from 0 to 100.

Test of the intervention

The overall effect of the intervention will be tested by the main effect of treatment group from the regression analysis.

The primary analysis will be a complete-cases analysis and we will support this with a main sensitivity analysis that uses multiple imputation to include cases with missing baseline or follow-up data.

Analysis method

Multiple regression analysis of complete cases. Baseline data was collected for all the primary outcomes.

Dependent variable: primary outcome at 20 months.

Independent variables: trial arm, baseline value of primary outcome, pre-specified covariates.

Level of significance: 0.05.

Alternative tests if distribution assumptions are violated

Regression with bootstrapping.

Adjustments of significance and confidence levels due to multiplicity of outcomes

Each primary outcome addresses a different hypothesis and therefore adjustment for multiple testing will not be applied.

Sensitivity analysis

See relevant section below.

Research hypotheses to be tested

There are no significant differences in the secondary outcomes between trial arms. Secondary outcomes include a number of outcomes at 20-month outcomes.

Analysis of secondary outcomes

Patient-level multiple regression analysis of complete cases.

Dependent variable: secondary outcome.

Independent variables: trial arms; baseline value of secondary outcome; pre-specified covariates.

Level of significance: 0.05.

Adjustment of significance and confidence levels due to multiplicity of outcomes

The analysis of secondary outcomes is regarded as exploratory and therefore not subjected to adjustment for multiple testing.

Alternative tests if distribution assumptions are violated

Regression with bootstrapping.

Sensitivity analysis

See relevant section below.