What is the evidence for the effectiveness, appropriateness and feasibility of group clinics for patients with chronic conditions? A systematic review

Purpose of review

The purpose of this systematic review is to examine the available evidence for use of group clinics with patients who have chronic health conditions.

Review question

The review question is as follows:

• What is the current evidence for the feasibility, appropriateness, meaningfulness, effectiveness (FAME) and cost-effectiveness of group clinics/GMVs for patients with chronic conditions?

Specifically:

• What different models of group clinic exist (in the UK and internationally)?

• What evidence exists about the outcomes and cost-effectiveness of these clinics?

• What evidence exists about patient experience of these clinics?

• What are the possible explanatory mechanisms for any reported improvements in outcomes?

Objectives

The primary objective of this review is:

• To identify evidence of effectiveness, or likely effectiveness, of group clinics and, where this is identified, to review evidence of impact, in particular cost-effectiveness of group clinics. This might include measures of efficiencies and clinic/staff time, use of services (hospitalisation rates), patient outcome (and surrogate clinical measures), behaviour, self-efficacy, quality of life and other patient and staff satisfaction indices.

Additional objectives were:

• to understand how group clinics have been conceptualised and to identify different models of use from a review of academic and grey literature

• to relate emerging findings on what works to current practice

• to identify research gaps for funding bodies and researchers.

Scope

This review covers all group clinics which include a component of clinical advice and management, as well as peer learning and support, for chronic health conditions. Terms (largely US) include ‘group medical visits’, ‘cluster visits’, ‘shared medical appointments’ and ‘co-operative health-care clinics’. The focus is on specialist-led services (i.e. replacing hospital outpatient appointments). Patient education and support groups (including expert patient groups) focused on self-management with no clinical advice or input are not the main focus of this review, although there may be some overlap in activity (see Chapter 2 for inclusion and exclusion criteria).

In seeking to inform the review from as holistic a perspective as possible, the team decided to examine the available evidence against the FAME framework. FAME is a mnemonic for the aspects of feasibility, appropriateness, meaningfulness and effectiveness and was devised at the Joanna Briggs Institute. 11 Appendix 1 sets out the FAME framework as used to guide the review process. This framework allows us to:

1. define the scope of the search strategy

2. define inclusion and exclusion criteria to specify types of studies to be included in the final report

3. construct summary tables of all included studies to present key information and findings

4. synthesise the evidence from the included studies.

It should be noted, however, that the FAME framework was selected principally to facilitate the synthesis process. In the interests of brevity we have subsumed considerations of feasibility, appropriateness and meaningfulness elsewhere under ‘appropriateness’ as an umbrella term, as in the report title.

The co-operative health-care clinic model

The specialty co-operative health-care clinic model

The drop-in-group medical appointment model

The shared medical appointment model

A conceptual model of group medical appointments

In order to initiate thinking around the elements of group clinics the team accessed a conceptual framework from the Cochrane GMA protocol21 (Table 2). This identified key structural elements for consideration within any group clinic-based intervention. This conceptual framework helped to identify key differences with regard to the intensity of the intervention (number of GMAs × frequency interval × duration) plus the qualitative consideration of the number of patients per GMA (and by implication the staff-to-patient ratio). The issue of continuity helped to distinguish between drop-in type appointments, those with a cohort of members progressing together and those with more fluid membership. Linked to this is the issue of heterogeneity as explored in issues relating to age, gender, ethnicity and experience of the condition. As our review addresses only chronic conditions, the chronic versus non-chronic was not pertinent except in considering why chronic diseases might be more amenable to a group clinic approach. The children/adults/elderly distinction served as a reminder that, typically for children and adolescents and occasionally for adults and older people, the perspective of family members (e.g. parents or carers) may be an additional factor in assessing the acceptability of group clinics. Finally, the team considerations from the Cochrane GMA conceptual framework highlight the requirement for group facilitation and team training as a resource issue.

High-level theory relating to social support

In order to bridge the often-reported dislocation of empirical intervention studies from their underlying or implicit theory, we conducted a brief literature survey to identify the prevalence of high-level theory in relation to group clinics. Particularly influential high-level theories reflected in the published accounts included social cognitive theory, social comparison theory, social learning theory24 and social support theory (Table 3). In addition, from the perspective of staff delivering the intervention, SMAs may access theories in relation to shared learning and interprofessional working. 27 When introducing group clinics, therefore, attention should be directed to the impact of the programme on staff interaction and interprofessional learning.

Theory relating to group interventions

Hoddinott et al. offer a useful generic framework against which to examine group interventions. 26 Interventions delivered to patient groups are addressed by their framework, which includes:

• the place, setting and context of the intervention

• the design of the intervention, the theory underlying the choice of intervention, the target population and choosing the relevant behavioural outcome to measure

• membership of the group

• how the group will influence people

• intended health outcomes and target populations

• what happens within the group.

Theory relating to chronic disease self-management

Theories relating to the core components of chronic disease self-management, namely the tasks of medical, role and emotional management,28 are particularly pertinent to the operation of group clinic approaches. These are highlighted in the rapid review, A rapid synthesis of the evidence on interventions supporting self management for people with long-term conditions, commissioned by the Health Services and Delivery Research (HSDR) programme. 29 In their review the authors29 highlight the key role of self-efficacy in relation to self-management behaviours. 30 This resonates with Knowles’ theory of andragogy, cited in the trial by Yehle et al. ,31 which proposes that adults are self-directed and that they expect to take responsibility for decisions.

Lorig and Holman32 identify five core self-management skills which can be seen to be accessed within a group clinic approach: (1) problem-solving; (2) decision-making; (3) appropriate resource utilisation; (4) forming a partnership with a healthcare provider; and (5) taking necessary actions. The standard group clinic format may be seen as an opportunity for the clinician and patient to harness all of these skills as targeted by individual components of the intervention.

Theory relating to monitoring

Finally, as highlighted by Taylor et al.,29 we can better understand the role of regular group clinic meetings by examining the complementary and evolving roles of periodic professional reviews and ongoing patient self-monitoring. 33 Group clinics could be conceived as a forum for juxtaposing, bringing these two roles in a potentially helpful synergy.

A symbolic role for group clinics?

Group clinic approaches may also fulfil a symbolic or emblematic role by instilling in the patients a hope and belief in the treatments being offered. 34 Social interchange in a group setting may emphasise the universality of the condition along with recognition that one is not alone in suffering or healing. Instrumentally, the group clinic setting offers an opportunity both to impart information through instruction or dialogue and to clarify any distorted or misleading information. 34 A sense of community may develop over time, with individuals beginning to display altruism and to derive a sense of usefulness from contributing to the group.

The group may provide patients with potential role models in the form of other group members who are better able to manage their condition and, thus, to function more effectively. 34 This may in turn stimulate imitative behaviour. Socialisation may offer potential catharsis through sharing and the destigmatising of chronic medical conditions as well as fulfilling a more pragmatic role as a source of direct advice and the sharing of coping strategies. 34 As a forum for interpersonal learning the group may encourage the sharing of experiences with others and problem-solving as a group. These resources may be more plentiful and more creative than may be offered by an individual clinician with no direct experience of living with a chronic condition. Peer pressure, in its positive sense and as an antidote to the unequal clinician–patient relationship, may encourage patients to become more empowered and, thus, more involved in their care. 34

Stage 1: search for reviews on group clinics

Our initial approach was to scope the literature around group clinics by searching for recent relevant reviews. All studies included in reviews were then scrutinised for inclusion in the review. Relevant terms for the search were found during the scoping exercise. Systematic reviews were identified from the following sources: PubMed Clinical Queries, Epistemonikos (Santiago, Chile; www.epistemonikos.org/), The Cochrane Library, Database of Abstracts of Reviews of Effects (DARE) and Google Scholar (Google^TM, Mountain View, CA, USA; https://scholar.google.co.uk), combining ‘systematic review’ with terms relating to group clinics, SMAs, etc. (see Appendix 2).

Stage 2: search of health and medical databases

The search strategy used a combination of free text and medical subject headings (MeSHs) and can be found in Appendix 2.

We searched MEDLINE and EMBASE via OVID SP, Cochrane Library via Wiley Interscience, Web of Science via Web of Knowledge and Cumulative Index to Nursing and Allied Health Literature (CINAHL) via EBSCOhost. MEDLINE, EMBASE, CINAHL and The Cochrane Library are commonly considered the core databases for identifying evidence relating to clinical topics.

The search strategy was limited to 1999–2014. Bibliometric analysis identified the sudden appearance of group visit studies at around 2000. Evidence was included if it was published between 1999 and 2014 and written in English.

The search results were downloaded into Reference Manager version 12 (Thomson ResearchSoft, San Francisco, CA, USA) where duplicates were removed before sifting for inclusion in the review was undertaken.

Stage 3: search for qualitative studies

A three-part search strategy was used to identify papers reporting qualitative research, as follows:

• Stage 1 – during screening and data extraction, any papers that were relevant and included qualitative data were tagged accordingly in Reference Manager.

• Stage 2 – a search of our Reference Manager database for relevant studies was undertaken using the keywords ‘qualitative’, ‘interview*’ or ‘findings’ in the title and abstract of the records. These terms have been found to have acceptable sensitivity for retrieval of qualitative research. 36,37

• Stage 3 – cited records for all included trials were searched on Google Scholar using the keywords ‘qualitative’, ‘interview*’ or ‘findings’ using the ‘search within citing articles’ check-box function. This would enable the retrieval of ‘sibling’ studies associated with the trials as well as more distant ‘kinship’ studies citing those trials for reasons of topical relevance. 38

Stage 4: search for costs studies

Three separate methods were used to identify studies for the assessment of costs and feasibility, as follows:

• Stage 1 – during screening and data extraction, any papers that were relevant and included costs data were tagged accordingly in Reference Manager.

• Stage 2 – a search of our Reference Manager database for the study was undertaken using the keywords cost*, economic*, charg*, expens*, reimburse* in the title and abstract of the records.

• Stage 3 – a targeted search of MEDLINE and EMBASE was undertaken, with no date or language restrictions, using the following search strategy: ((((shared or group) adj medical adj (visit or appointment or clinic or care)) OR (group adj (visit or appointment or clinic or care))) AND (economic* or cost* or charg* or expens* or reimburs*)).ti,ab.

Stage 5: search for UK studies

Studies conducted in the UK were identified in two ways:

1. Geographical terms for ‘united kingdom’, ‘uk’, ‘britain’, ‘England’, ‘Scotland’, Wales, Ireland were used to retrieve records from the Reference Manager database.

2. Similarly, geographical terms for ‘united kingdom’, ‘uk’, ‘britain’, ‘England’, ‘Scotland’, Wales, Ireland were used to retrieve items from Google Scholar in conjunction with the most common terms used for the intervention, that is, ‘shared medical appointments’, ‘group medical clinic’, ‘group medical visit’ and ‘group visit’. This search approach harnessed full-text retrieval and so added value over the title-and-abstract-based approach listed above.

Progressive fractions method

Following the sifting of 4176 titles and abstracts a further 1212 search results were scrutinised using a method of ‘progressive fractions’. Progressive fractions is a method developed in-house by the ScHARR team for undertaking systematic reviews within a time-constrained period. Essentially, it involves conducting a sensitive search strategy in order to populate a project reference management database. This database then becomes the data set that is progressively ‘mined’ for articles for potential inclusion. Essentially, titles and abstracts are reviewed in decreasing relevance order until no further unique relevant references are retrieved. This method also minimises the likelihood of relevant references being missed through being submerged by excessive quantities of irrelevant noise (Table 4).

Instead of the ‘big bang’ approach that typifies systematic review methods and which conflates terms of low specificity alongside terms of high specificity, the progressive fractions method involves using single string strategies (e.g. ‘group medical visit*’) in decreasing likelihood of unique retrieval, with the team evaluating retrieval results at each point. As each progressive fraction is executed attention is focused on the identification of unique results. When an additional relevant reference is retrieved this yields additional search terms. Quantitative results for the new search terms are used to evaluate whether or not it will be time effective to sift new results, taking into account the number of relevant studies already identified by the combined search strategy and the number of additional records to be sifted. Progressive fractions allows a review team to make iterative and informed judgements about the optimal sensitivity for a systematic review search. After precise search terms were used we had scanned 11% of our sensitive database and retrieved 89.7% of our randomised trials. The remaining four trial citations were identified by citation searching and checking existing systematic reviews. The same precise search sets were also scanned for qualitative studies.

Setting of intervention

Interventions are not initially excluded on the basis of the setting for the group intervention, given the potential for very similar interventions to be delivered in the community or primary care setting as well as in hospital/outpatient settings. Although the review team has justifiable concerns about the additional literature likely to be identified if group approaches in primary care are included within the review scope we cannot identify a sound justification for excluding such studies on conceptual grounds particularly given that the setting for interventions and definitions of ‘specialist’ care may cover a wide range of different settings.

Systematic reviews

Systematic reviews were appraised using the guidelines employed by the Centre for Reviews and Dissemination (CRD) when populating their DARE. 43 This method was employed to ensure consistency of approach between our assessments and existing assessments of published reviews.

Randomised controlled trials

Randomised controlled trials were formally assessed for quality using questions from the Critical Appraisal Skills Programme (CASP) checklist 11 questions to help you make sense of a trial44 in order to explore study limitations qualitatively and the Cochrane risk of bias tables in order to identify likely sources of bias. 45 Assessment of the limitations of included studies was also undertaken using the limitations reported by study authors in the included studies.

Qualitative research

Qualitative studies were formally assessed for quality using questions from the CASP 10 questions to help you make sense of qualitative research checklist. 46 Surveys were not formally appraised and, therefore, were used only to validate findings from qualitative research.

Costs

We undertook an assessment of relevance of evidence to the study objectives by answering three questions about each paper: the currency of the data, the quality of the data sources and the relevance to a UK setting.

UK initiatives

Research studies reporting UK initiatives were not formally assessed given the heterogeneity of study types, making comparability problematic. Nevertheless, all reports of initiatives were reviewed for any identifiable or acknowledged limitations.

Systematic reviews

Literature searches retrieved 13 systematic reviews and one review protocol. This evidence was reviewed in Results of the review of reviews.

Randomised controlled trials

We retrieved 32 papers representing 22 different RCTs. These trials are reviewed in Results of the review of effectiveness.

Qualitative studies and surveys

We identified 12 qualitative papers reporting 10 different qualitative studies. In addition, we identified four surveys that were used to triangulate qualitative research findings. These qualitative studies and surveys are explored in Results of the qualitative synthesis.

UK initiatives

We identified 15 papers reporting 12 UK group clinic initiatives. This review of current practice is examined in Results of the review of UK evidence.

Realist synthesis

Data from the 13 systematic reviews, 22 different RCTs, 10 qualitative studies and 12 UK initiatives (a total of 75 papers) were used to inform Chapter 4.

Cost studies

We identified eight cost studies either nested within RCTs or reported as separate cost-effectiveness or cost–utilisation analyses. These cost studies are analysed in Chapter 5.

From Table 7 it can be seen that group clinics is the most frequently mentioned model, in 19 of the 84 papers. Other frequently used labels are SMAs (n = 12) and CHCCs (n = 10). A further nine labels are used in the 84 papers included in this review, with even greater variation in the non-empirical literature.

Review characteristics and review strategy

As a precursor to our own review of group clinics the review team identified 13 reviews that either matched or overlapped the scope of the planned review. Another review, The effectiveness of group visits for patients with heart failure on knowledge, quality of life, self-care, and readmissions: a systematic review,54 is available only on private subscription from the Joanna Briggs Institute Library and so a summary, commissioned on request from the CRD, was used in assessing the evidence. One Cochrane review entitled Group medical appointments for people with physical illness is currently in progress. 21

Characteristics of previous reviews

An initial task was to seek to characterise existing reviews in terms of their congruity, or otherwise, with regard to the population-intervention-study type elements. In Table 9 total congruity with a particular element is indicated by a +++ notation. Close congruity is correspondingly indicated by ++, while a narrow specific focus is assigned a + notation. In this way, the key reviews with the greatest potential to inform our review question are clearly identified.

Only one18 of the 13 reviews was congruous with our review when matched against both population and intervention (see Table 9). We therefore decided to undertake our review as a more comprehensive update of this systematic review by Edelman et al. 18 Three further reviews55,57,83 articulated the intervention of interest to our review (although they did not employ the precise terminology of ‘group clinics’) but in only one specific disease/condition. We therefore decided to prioritise these three reviews as sources of potential studies for inclusion. The remaining reviews would be checked for their coverage of included studies and for suggestions of further studies for inclusion.

Populations

Two of the included reviews48,56 examined all populations, resulting in a focus wider than that determined for this review. A further seven reviews47,49,51,53–55,87 focused on one specific condition [in all bar one instance this condition was diabetes, with the exception being the review by Slyer and Ferrara54 (heart failure)]. Two of the remaining reviews42,50 were broadly coterminous with our own review, focusing on chronic disease in adults (however, we also included children and adolescents). Only two reviews18,52 covered the exact same population as our review, namely patients of any age group with chronic disease/chronic medical conditions.

Interventions

Three reviews demonstrated a specific group education focus. 42,47,51 A further five reviews had a scope for the intervention that was broader than group clinics: for four of these reviews48–50,54 this focus was labelled ‘group visits’ and for the remaining review52 this was ‘group care’. Two reviews18,55 focused on SMAs and two reviews targeted GMVs;53,57 both of these labels were considered coterminous with our own. The Cochrane review by Rolfe et al. covered a heterogeneous mix of interventions for improving patient trust, one intervention of which was a group clinic approach. 56

This important review-mapping phase has established the potential of our review to provide the most comprehensive and most up-to-date coverage of the topic of group clinics for chronic medical conditions to be found in the published literature.

Review quality

In addition to mapping all 13 of the existing reviews against the population, intervention and study type characteristics (see Table 9), we explored reasons for a mismatch between our review question and that in other reviews (Table 10). We then decided to produce a brief summary of the quality of the four key reviews18,83,85,87 in order to assess any uncertainties underpinning their results (Table 11).

Findings from four key reviews

Summary of findings from other reviews

The Canadian Agency for Drugs and Technologies in Health (CADTH) health technology assessment group52 conducted a review of the clinical effectiveness, cost-effectiveness and guidelines of group care across all aspects of chronic-disease management. They identified eight studies meeting the criteria for inclusion in their review: three systematic reviews, two RCTs, two non-randomised studies and one evidence-based guideline. They concluded that there was evidence for improved glycaemic control for diabetes group care (vs. usual care) and an isolated study in favour of better blood pressure control for group care of hypertension. However, they had been unable to find any information on effectiveness of group care for either chronic obstructive pulmonary disease (COPD) or human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS). A significant observation related to the fact that variations in the structure of group care, together with inadequate detail of reporting for the usual care, meant that the group felt unable to draw meaningful conclusions.

Steinsbekk et al. 51 reviewed 21 studies, reported in 26 publications, involving a total of 2833 participants. For the main clinical outcomes, HbA_1c was significantly reduced at 6 months, 12 months and 2 years, and fasting blood glucose levels were also significantly reduced at 12 months but not at 6 months. For the main lifestyle outcomes, diabetes knowledge was improved significantly at 6 months, 12 months and 2 years and self-management skills also improved significantly at 6 months. For the main psychosocial outcomes there were significant improvements for empowerment/self-efficacy after 6 months. For quality of life the authors were unable to draw any conclusion owing to high heterogeneity. For the secondary outcomes there were significant improvements in patient satisfaction and body weight at 12 months for the intervention group. The review team found no differences between the groups in mortality rate, body mass index (BMI) blood pressure and lipid profile.

In a Cochrane review of group based education for diabetes, Deakin et al. 47 identified eight RCTs (n = 1260) and three observational studies (n = 272). Random-effects meta-analyses showed that HbA_1c and fasting glucose concentrations were lower in the intervention group than in the control group {at 4–6 months [1.4%, 95% confidence interval (CI) 0.8% to 1.9%; p < 0.00001], at 12–14 months (0.8%, 95% CI 0.7% to 1.0%; p < 0.00001) and at 2 years (1.0%, 95% CI 0.5% to 1.4%; p < 0.00001)}. Diabetes knowledge scores were greater in the intervention group than in the control group (standardised mean difference 0.95, 95% CI 0.72 to 1.18) (three trials, n = 432), yet not statistically significantly so. More patients in the intervention group than in the control group reduced their use of diabetes medication over 12–14 months (relative benefit increase 825%, 95% CI 202% to 2738%) (five trials, n = 654). One RCT (n = 314) assessing empowerment and psychosocial self-efficacy reported greater total empowerment scores in the intervention group than in the control group throughout follow-up (p < 0.05). This indicates that the group education element of the group clinic intervention may, in itself, be efficacious. A key issue is the added benefit, if any, that is accrued from employing other supplemental non-group education-based features of the intervention within a group clinic framework.

Two reviews fall short of current practice for systematic reviews. In a narrative review without meta-analysis, Jaber et al. 48 concluded that there are sufficient data to support the effectiveness of group visits in improving patient and physician satisfaction, quality of care and quality of life, and in decreasing emergency department and specialist visits. Significantly, Jaber et al. highlighted a need to abandon old nomenclatures and to clearly define the structure, processes of care, content of visits and appropriate outcome measures. 48

Riley and Marshall49 produced a review of existing reviews including three general reviews8,22,89 as well as the previously mentioned specific review by Jaber et al. 48 They observed that, although ‘a variety of successes are evident from the entire group visit approach, results are inconclusive regarding any specific model for group visits and inconsistent regarding improvement of important patient outcomes’. 49 Nevertheless, Riley and Marshall concluded that there was evidence that ‘group visits may reduce costs, physiological outcomes may be improved, and patient and clinician satisfaction may be enhanced’. 49 They cautioned, however, that ‘The group visit model needs further testing to determine the most effective approach, and the most effective health care provider team to facilitate the group visit, along with standardization and application across a variety of situations’. 49

In a review tangentially related to the topic, looking at interventions for building up trust between patients and clinicians, Rolfe et al. 56 identified three studies that had a group visit component. However, one of these studies was excluded from our review because it involved an induction visit as part of joining a health maintenance organisation. The remaining two interventions were included. The focus on trust is, however, important, as this represents one mechanism by which the group clinic interaction is hypothesised to work.

Overall summary of findings from reviews

All of the reviews of group clinic-type approaches exhibit methodological challenges with regard to the inconsistent use of labels and definitions for the intervention and a lack of detail relating to the intervention components. Mechanisms for action are poorly theorised, and variability in outcomes and in subsequent effect sizes makes attribution of effect problematic. With the review team sensitised to the topic via existing reviews, we attempted to examine the evidence base for effectiveness by bringing together previously identified trials with new studies identified via sensitive search strategies.

Study characteristics

A total of 32 papers involving 22 trials were identified for inclusion in the review. The search of MEDLINE, EMBASE, The Cochrane Library, Web of Science and CINAHL databases yielded a total of 12,819 citations. After adjusting for duplicates, 11,000 remained. Of these, 5255 studies were discarded because after reviewing the abstracts it appeared that these papers clearly did not meet the inclusion criteria. The full text of the remaining 133 citations was examined in more detail, from which the 32 papers were selected and included in the systematic review (Tables 12 and 13). No unpublished relevant studies were obtained. No conference abstracts were identified that met our inclusion criteria and contained sufficient information to address the review question.

Setting characteristics

Of the 22 trials, 17 were conducted in the USA. 20,31,58,60,63–66,68,71,72,75–77,79,80,86 Of the remaining RCTs, two73,74 were conducted in the People’s Republic of China, two82,85 were conducted in Italy and one was conducted in Norway. 70 Not a single RCT was conducted in a UK setting.

Intervention characteristics

Included studies comprised a total of nine different interventions. Of these, the CHCC (six studies) and SMA (five studies) models featured most frequently. SMAs were represented by trials that have occurred during the comparatively recent period of 2010–14, while the CHCC studies occurred during the period 2001–4, with the exception of two recent non-US studies reflecting a resurgence of interest. There were no RCTs for two of the models, the specialty CHCC model and DIGMAs (Table 14).

Intervention components

Edelman et al. 18 has characterised the main features of SMA interventions. Almost 90% of such interventions had an educational component and nearly 65% are delivered by multidisciplinary teams. A behavioural intervention is a feature of exactly half of the SMA interventions. A focus on medicine management is evidenced in the fact that 55% of interventions include medication adjustment. Almost 90% of interventions include peer-to-peer support and just over 40% include clinician training. We did not find it possible to distinguish intervention content for studies not included by Edelman et al. from those studies included in their review, implying that findings from their review are generalisable to a wider population of group clinic approaches. 18 As seen in Appendix 4, our review has completed a very detailed data extraction of intervention components from RCTs. However, the facility to synthesise and analyse these data is constrained by the fact that (1) these data capture the completeness of reporting of each report, not the intervention content for that report, and (2) there is considerable variability in these descriptions, implying that similar components may be described differently or, conversely, that similar-looking descriptions may mask important substantive differences in content, delivery or both. Indeed, even different reports of the same study portrayed different depictions of the same intervention. Notwithstanding these reporting limitations, we found that some element of socialisation was included in 15 of the studies57,60–62,65,66,68–70,73,74,76,77,79,81 and group discussion (i.e. many-to-many interaction) was reported in 14 studies. 63,64,66,68,70,71,73–77,82,85,86 Eleven studies explicitly reported health education/information presentation(s) by individual clinicians,57,60,65,66–68,70,71,73–75,79 with one for health education/information presentation(s) by multiple clinicians63 and two for health education/information via booklet, leaflet or video. 70,80 Seven studies reported medication review and four describe completion of prescriptions. 60,63,66,71,75,80,81 Six studies reported individual consultation within the group session60,65,76,77,84,86, with five describing individual consultation immediately following the group session for all patients58,59,68,69,82 and three for individual consultation immediately following the group session for selected patients. 73,74,83 Six studies reported routine medical checks being performed by multiple clinicians57,65,66,74,79,86 and six reported these checks being made by individual clinicians,60,63,67,73,77,82 while only two studies reported routine medical checks being conducted by the patient. 69,70 Only one study reported telephone follow-up. 69

Narrative summary of study quality

The review of RCTs included 32 papers reporting 22 trials (Table 16). The quality of included RCTs was assessed using questions relevant for RCTs and from these responses a Cochrane risk of bias was determined for each study (Table 17). Of the 22 trials, nine studies (11 papers)60–63,65–70,73 were categorised as having a low risk of bias, 10 studies (11 papers) were categorised as having a high risk of bias20,57–59,64,71,74,75,79–87 and three studies (four papers) were categorised as unclear. 72,76–78 The large number of studies with a high risk of bias means that any conclusions based on these trials should be treated with caution. The discussion on the quality of the RCTs will begin by discussing general problems with the studies then consider the groups of studies with low, high and unclear risk of bias.

A key problem for all of these studies is the possibility of selection bias having impacted on the results. All studies included patients who chose to participate in group clinics. Patients who wish to participate in group clinics are likely to give more positive results on self-reported outcomes. Additionally, a patient’s choice to be involved may indicate greater concern about improving their condition and, as such, they may be more motivated to implement suggested changes to their lifestyle, thereby improving their clinical outcomes.

Another significant problem with these studies was that it was not possible to blind patients or health-care personnel to treatment intervention group, which could lead to bias. Two studies69,83 did state that they have researchers blinded to patient’s treatment groups to measures outcomes. This bias could potentially be more significant with certain outcome measures. Clinical outcomes measures, for example, blood pressure and blood glucose, would be less likely to be affected by this bias. However, outcome measures around patient satisfaction, self-efficacy, self-reported outcomes or outcomes reported by the team delivering the group clinics could be open to bias. Some studies had doctors treating patients in both the intervention and the control group, giving the possibility of a ’halo effect’. 72

The majority of the studies had only a short follow-up, generally 6 months to 1 year, making it impossible to assess the longer-term impact of the interventions. Two of the studies70,79 did have a 24-month follow-up. The 11 studies assigned a low risk of bias60,61,63,65–70,73,78 were generally large, well-conducted trials.

A total of 13 studies were assigned a high risk of bias. 20,31,57,59,64,71,74,75,79,80,82,85,86 One of the studies with a high risk of bias was a pilot study57,58 with small sample sizes, no blinding, patient selection bias and short-term follow-up. Five of the studies with a high risk of bias48,71,80,81,83 had patients with different baseline characteristics.

Three studies were given an unclear risk of bias,72,76,77 owing to insufficient details of the trials methodology being provided. One of these was a pilot study. 72

Study analyses: condition-specific clinical outcomes

Fifteen of the 22 trials included a population with diabetes. By far the majority of the trials (11 out of 22 trials) studied a population with type 2 diabetes. 57,60,63,64,72,74–77,80,83 A further four trials studied a mixed type 1 and type 2 diabetes population,20 a type 1 diabetes-only population70,85 and a population with type 2 diabetes and hypertension. 66,67,69

A further group of studies examined the effects of group clinics in populations with a variety of cardiac problems (heart disease/hypertension71,73 and hypertension/heart failure31). Coleman et al. 65 studied a population with one or more self-reported chronic conditions (e.g. asthma, COPD, congestive heart failure, diabetes and heart disease). Scott et al. 79 also studied a population with a range of chronic conditions (arthritis, hypertension, difficulty hearing, heart disease, liver disease and bladder/kidney disease).

Recent years have seen group clinics extended to a wider variety of conditions. Dorsey et al. 68 studied a population with Parkinson’s disease and Seesing et al. 90 have completed a 6-month follow-up of a population with chronic neuromuscular disorders.

Study analyses: health service utilisation measures

In addition to the biomedical outcomes, several health service utilisation measures have been measured in isolated studies. These are not suitable for meta-analysis but these are reviewed together with an assessment of the consistency around results.

Summary of main findings from randomised controlled trials

In summary, findings from a total of 33 RCTs, of which almost half are considered to possess a low to moderate risk of bias, indicate that biomedical outcomes (e.g. blood pressure and HbA_1c, specifically within the disease context of hypertension and diabetes) are most likely to be significantly affected by group clinic-type interventions,. However, this is by no means the case for all such outcomes. One of our clinical advisers suggests that factors affecting modification of blood pressure and HbA_1c are multifactorial and are, therefore, correspondingly more likely to respond to a complex, multifactorial intervention such as a group clinic. In contrast, measurements such as cholesterol are affected by less complex health choices for which a group intervention may be less appropriate. The reasons for this difference in results across biomedical outcomes require further investigation.

Where such effects to be demonstrated conclusively, these would be of important clinical significance. As Housden et al. state:

Small decreases have . . . substantial clinical impacts: a 1.0% reduction in HbA _1c may be associated with a 37% decrease in microvascular complications, up to a 14% reduction in the incidence of myocardial infarction and a 21% decrease in the risk of death from diabetes. 53

In moving away from easily monitorable and measurable outcome measures it becomes increasingly more challenging to demonstrate a causal effect. For example, disease-specific health-related quality of life demonstrates a significant effect (albeit from only three RCTs), whereas generic health-related quality of life (measured in two RCTs) at a further level of abstraction is only marginally significant. The most recent systematic review and meta-analysis, including only SMAs in a diabetes context,55 concludes that published examples were so heterogeneous as to yield genuine uncertainty about which elements of the intervention make a SMA intervention successful. Furthermore, issues concerning acceptability and sustainability have been raised from the trial evidence and require further exploration. These issues are explored in the following sections examining qualitative, UK-centric and theoretical aspects of the group clinic type of intervention respectively.

Characteristics of qualitative studies

The review identified 12 qualitative papers27,97–107 reporting 10 studies (Table 19). Seven of the 10 studies were conducted in the USA,27,98–100,102,104,105 with one each from the UK,97 the Netherlands103 and Canada101,106,107 (three papers). Four studies explored SMAs27,99,102,103 and one examined DIGMAs. 98 The remainder comprised GMVs (two studies, four papers)101,104,106,107 and group clinics (three papers). 97,100,105

Characteristics of surveys

In addition, the review identified four surveys96,108–110 to be used to corroborate findings from qualitative evidence. Three of the surveys were conducted in the USA,108–110 with the remaining survey from the Netherlands96 (Table 20). Two surveys explored GMAs96,110 and one survey examined DIGMAs. 108 Jhagroo et al. 109 reported an adaptation of three models: the DIGMA, the CHCC and the physical SMA. As quality assessment of surveys is problematic, these papers were not critically appraised, and data were used only to triangulate findings and not to generate themes.

Comparison of the distribution of clinic models from the effectiveness literature with that from the qualitative literature reveals that the principal models of group clinic-type approaches are all well represented.

Eight qualitative studies111–118 were excluded from the qualitative synthesis, as they were available only as conference abstracts. However, three abstracts111,117,119 related to UK initiatives and so are examined further in the Results of the review of the UK evidence.

Study populations and settings

We identified a total of 12 qualitative studies of group clinic type interventions. One-third of these (four studies) examined the attitudes of patients only. One study96 explored the views of patients and carers, and four studies investigated both patients and health-care providers. 97,100,101,107 One study investigated the views of providers in isolation102 and one study included views of providers, patients and caregivers. 105 A final study examined the views of students regarding SMAs as an educational experience. 27 The quantitative review had revealed a complete absence of measurement of provider experience in the included studies. The qualitative evidence base clearly has an important part to play in addressing the wider acceptability of the group clinic intervention within a health-care delivery system. 86

Quality of included qualitative studies

The overall quality of the studies included in the review is shown in Table 21, and an assessment of the quality of each included study is given in Table 22.

Population of the conceptual framework

We extracted qualitative data against the elements of the analytical conceptual framework13 (Table 23), deconstituted into fields on a data extraction form (see Appendix 3). The best-fit framework approach provides for inclusion of additional inductive elements once the deductive stage of the synthesis is completed. The qualitative data yielded six principal themes as presented below. However, many of the data have been extracted from one particularly rich qualitative study101 and, therefore, may represent views that are not necessarily typical of the study populations across all of the included qualitative studies. Eight richer studies were particularly influential in populating the conceptual framework and subsequent synthesis. 96–99,101–103,111

Findings 1: feeling supported

A common finding was that the group environment offered individuals an opportunity to derive support from others in a similar or comparable position to them. Such support could be accessed during the initial socialisation sessions or, subsequently, when engaging in group education or interaction with clinicians.

There was some evidence to suggest that this feeling of being supported subsequently led to a sense of security.

You gain; I think you gain a feeling of security, of understanding, of sharing with other people, of compassion, of support . . . so many things that you wouldn’t gain if you were one on one because of the humanity of us as people. You know we try to support one another.

Reproduced with permission from Piper106

In such a climate of trust patients were more likely to share information within the group. This in turn affected the cohesion and a feeling of community within the group, described by one author as an ‘esprit de corps’. 99

The need for feeling supported is illustrated by one extract, which attests to the feeling of isolation a patient may feel if they are not receiving necessary support from either partner (husband) or doctor:

You’ve got a group that can back you up . . . understanding what you’re going through . . . if I tell my husband oh my blood sugar is 2.4 today, he says . . . well you better take some insulin,’ . . . he hasn’t really bothered to even read about it . . . he’ll get irritated with me. Well that’s the last thing you need.

Reproduced with permission from Piper106

Findings 2: learning from each other (reciprocal learning)

A notable finding from the qualitative research studies was that the group setting offered a context in which individual group members could learn from the clinicians, where they could learn from other group members and, significantly, where the clinician could learn from the group members. This last finding was one of a number that signalled at a shift in the power differential from the clinician dominance of the one-to-one consultation. This important consequence of group clinic approaches was explicitly highlighted by several commentators:

Overall, the power dynamic between patient and physician was lessened as the patient now viewed themselves as being able to impart information to the physician. 102

Findings 3: legitimising question answering

A group clinic environment may represent a less intimidating clinical context for patients who are more reticent. Safety, and indeed strength, in numbers may be perceived as an antidote to the power imbalance experienced when a patient encounters a clinician on a one-to-one basis.

Findings 4: structure and content

We were able to map the qualitative findings on the individual components of group clinics to those aspects of self-management (Table 24) identified in the report by Taylor et al. 29

Noticeable from the above mapping process (see Table 24) is that group clinics are able to fulfil many of the extended self-management roles that may be required from any clinician–patient interaction. In particular, the group context is strong in meeting a need for training/rehearsal of communication with health professionals, for ADLs, for practical self-management and for psychological strategies, providing a safe environment in which these activities can be modelled. The group setting is able to fulfil some requirements for social support, especially when these needs are not being met by a patient’s significant others or by their health professional. However, what is missing from the Taylor framework29 (see Table 24) are the functions of ‘groupness’ seen in socialisation, a sense of shared experience, modelling of realistic or ideal behaviours and identity through group cohesiveness. Clearly the group clinic approach cannot be conceived simply in terms of its self-management function, even though this was a major driver in the origins of group clinics.

Findings 5: confidentiality and privacy

One qualitative study in particular107 focused on issues relating to confidentiality and privacy, a frequently expressed concern in the context of group approaches. Certain protections can be easily instituted such as:

1. initiating each session with a discussion of confidentiality

2. setting ground rules with examples

3. gaining permission for disclosure of particular types of information (e.g. laboratory values)

4. emphasising that participation in the group is not dependent on sharing of personal information

5. asking specific consent to share information of a particular patient at times during a session when it is considered potentially valuable as a resource to a group.

This latter approach is described very positively by one participant as a way of allaying initial concerns about attendance:

But he’s [doctor] he’s been very good because he, he makes sure each person gives permission for him to relate any information about them. You have to agree. So do you mind if I talk about your disease or whatever’ and you can say yes or no.

Reproduced with permission from Piper106

Reciprocal learning and circular questioning require similar protections and filters in order to protect the confidentiality of those sharing the learning. An elegant example of how the distinction between the contexts of information sharing and confidentiality is presented in a small-town context:

[O]ne provider explained to the group that if he/she learned something about thyroid disease, then this information could be shared with others. The provider went on to tell the group that what was to remain confidential was who ‘Mrs. Jones, our neighbor’ was the person who has a thyroid condition. 107

Findings 6: the life cycle of the group

It is interesting to observe different views of the group process depending on the stage a person was at in the life cycle of the group. These views can be clustered around the three phases of contemplation, initiation and maintenance of group attendance.

Summary of main findings from qualitative studies

Clearly, socialisation played a large part in the group clinic intervention, with this factor being mentioned consistently across the qualitative studies. Several respondents mentioned the relaxed atmosphere where they were not afraid to share health issues with others. Linked to this is the role of the clinician as facilitator with the group being cast in the expert role – unless misinformation needs correction. 101 Providers benefit from adopting this communicative role101 and also learn more about their patients’ experience of their condition and their medication than they typically might in a one-to-one setting.

There is some evidence of patients benefiting from role models, not necessarily in the sense of modelling ideal behaviours but often in the sense of conveying a realistic expectation for what the patients are going through. 97 Such modelling extended to normalisation of group behaviours, especially with regard to management of their condition.

Information exchange is a key component of the group interaction, with patients sharing technical knowledge of their condition, practical suggestions, detail on available resources and their own personal experience. However, it is interesting to observe that patients do not adopt a particular role within a group setting. To use the analogy of online forums, there are those who are active participants and those who are quite comfortable being ‘lurkers’. Lurkers benefit from information shared within the group. They may also benefit from ‘information surrogacy’, that is someone asking a question of concern to them (either serendipitously or because this question has surfaced during socialisation and is then articulated by a more vocal member of the group). This explains why group interventions can be fulfilling for these quite different personality types:

the interview data did not support this hypothesis [that a more gregarious personality would find the groups situation more acceptable] . . . more private people appeared to be content to read a book or a newspaper or to listen to others rather than to join in . . . 97

Characteristics of UK studies/initiatives

A total of 12 reports97,111,116–118,121–128 reflecting nine initiatives within current UK practice were identified from the literature (Table 25). One further UK initiative, a phenylketonuria group clinic at Great Ormond Street Hospital,119 was identified from web searching. Owing to the limited volume of UK evidence, information from conference abstracts was included, where the initiative met the inclusion criteria.

Quality of included UK studies

Although the remit of this report was to identify all published examples of UK group clinic practice, this approach can be seen to have had a deleterious effect on quality. Of the 13 identified studies, only five97,118,125,127,128 could be considered either research or evaluation and so could be formally assessed for quality (Table 26). Four of these were audits, service evaluations or patient questionnaires, leaving just one observational study125 (moderate risk of bias) and one good-quality qualitative study. 97

Contact with UK advisers

Given the absence of rigorous UK evaluations the review team decided to approach (1) clinicians involved in delivering group clinic interventions and (2) clinicians delivering care to patients with diabetes, as this was the group most represented by international evidence (15 of 22 RCTs). The team contacted three clinicians (two replies) delivering diabetes care and two academics (two replies) involved in evaluation of a group acupuncture initiative (see Acknowledgements). Clinicians were sent a four-page summary of review findings to date as of mid-September 2014. Questions explored with clinicians are reproduced in Box 2.

For ease of interpretation observations from these clinical specialists have been integrated as far as possible with relevant findings from the literature (see Study analyses).

Patient and public involvement

The short time frame for the review and the heterogeneity of group care models, coupled with an overall review strategy that already accommodated patient perspectives from the qualitative and UK research literature, meant that it was not considered feasible to elicit unique perspectives from current or past NHS patients. We accept that had there been more examples of current UK initiatives this could have proved a useful source of additional data. We therefore recommend that any future UK-based evaluations seek to engage patients and the public through robust involvement mechanisms.

Study analyses

Summary of main findings from UK evidence

Thirteen papers were identified that described initiatives from the UK. 97,111,117–119,121–128 None of these represented experience from rigorously conducted experiments. Descriptions of several initiatives were available only as abstracts. Acceptability of group clinics is high among a population requiring group acupuncture for knee osteoarthritis. However, the sensitivity of health and lifestyle topics is not a key issue for this particular population. Even within this context there was an expressed demand for single-sex sessions, including in a Muslim population. 97 Patients considered that single-sex sessions represented good practice, regardless of specific religious and cultural considerations. A good-quality qualitative study from the UK highlighted the importance of situational factors such as a physical space and a flexible appointment system. 97 Patients for whom group clinic sessions may not be as appropriate include those with complex conditions, those with extreme pain97 and those with hearing difficulties.

It should be noted that the absence of empirical studies from a UK context has led to a disproportionate reliance on the reported experience from UK group acupuncture clinics. Two particular considerations are:

1. Group acupuncture clinics differ from other group clinics because patients arrive with the expectation that they will receive treatment. For group clinics in general, treatment and follow-up is more typically contingent on the findings from the monitoring and consultation processes. Potentially, patients attending group acupuncture clinics may have stronger motivations for attendance than those attending for routine monitoring.

2. Acupuncture treatment involves patient downtime (typically at least 20 minutes) as the patient receives treatment. Although the efficiency argument (in terms of number of patients who can be seen by a consultant) is frequently rehearsed in opinion papers the driver for acupuncture clinics may be seen as an example of where a clinical team may be able to ‘work smarter’. Although this driver may be seen to make group acupuncture clinics demonstrably different from other monitoring contexts this could have the potential benefit of showcasing another type of situation that might potentially benefit from group approaches in other disease areas.

Contact with the clinical experts revealed other potentially important issues, in that the acupuncturist was not formally trained in or charged with the task of facilitating the group. As a consequence, group interaction was expected to be more organic and less manufactured. Furthermore, socialisation, as we have termed it elsewhere in the report, could not really be considered a formal part of the ‘programme’. However, potential benefits have been identified where group communication occurs opportunistically, such as (a) the normalisation of symptoms, (b) the sharing of information on resources available and (c) encouragement to adhere (or, more accurately, to continue to attend even though improvement may take a few weeks to become noticeable).

Other contextual UK evidence

The review team also accessed a UK-based discussion on group clinics hosted by the GP-UK (an e-mail discussion list for the UK Education and Research communities focusing on UK general practice topics) discussion list. 129 Several observations from list members are worthy of note. First, one correspondent observed that use of the word ‘clinic’ in ‘group clinics’ might be considered problematic as it might create the impression of an individual session. This might remain the case even though there is provision of explanatory information to the contrary. Two studies,97,103 including one from the UK,97 observed that patients or their carers had different expectations of the group clinic arrangement and this resulted in negative perceptions of the clinic when these expectations were not actually met.

Some patients felt that the explanations supplied by nurses were occasionally incomplete and could be inconsistent, which would result in problems. 98

It (mixed sex clinic) wasn’t something I was expecting, wasn’t something I was told about before I went in . . . you know we’re not all beautiful shapes or whatever, and it’s sort of a bit embarrassing.

Woman in her 50s97

And:

One parent indicated that he/she was not informed properly about the purpose of SMAs, which resulted in incorrect expectations. 103

The GP-UK discussion list also raised concerns about confidentiality as, for example, erectile dysfunction might be discussed during a diabetic clinic and this may be difficult if couples are present. Sudell discusses that women may feel able to discuss such matters with their female friends, but this may result in the partner feeling embarrased. 130

This observation highlights that assumptions must not be made about the content of a group discussion simply on the basis of the condition itself – a sensitive condition might engender sympathetic discussion and yet a general condition may equally yield embarrassment. The critical aspect is the dignity of those who are participating and not the condition per se.

Programme theory 1: ‘feeling supported’

The group clinic model allows for support from two main sources – from clinicians and from other patients – to contribute to what has been described as a supportive environment. 120 In theory this dual support should result in an additive effect over that offered either by a one-to-one consultation – ‘I trust him more when I see that he’s open to learning and figuring out new things that are only happening in group dynamics’101 or from attendance at a peer support group:

You know we try to support one another, it’s kind of human to do that. It’s human to have compassion for other people who have problems and you can show that and you can feel that from other people when you’re in a group, you don’t in isolation. 101

Support may be verbal or may be the effect of perceived solidarity:

I was just a little apprehensive at first, but I saw all the other brave ladies there not flinching or anything, so I thought, ‘Oh well, it can’t be too bad’. 97

However, other implications for this dual source of support are that patients may access support judiciously and appropriately by deciding between the two sources or that they may use the availability of an alternative source of support to compensate, for example, for the perceived inadequacies of support from clinical staff.

The same behaviors are not portrayed as a problem by patients when coming from peers . . .  When such pressures came from providers in a one-on-one CE [clinical encounter], the same behavior was portrayed as abusive or threatening. 101

Of course, the availability of both kinds of support requires careful management when misinformation is being given.

Programme theory 2: ‘building trust’

Following an effectiveness review for the Department of Veterans Affairs in the USA, two experienced researchers, Kirsh and Aron, have undertaken Theory Driven, Context Dependent Studies of Shared Medical Appointments: A Realist Work in Progress. 131 They propose that a key mechanism to the success of SMAs is the build-up of trust in the peer group. We also found evidence of trust being built up in the relationship of the patient with the care provider:

I’ve learned to trust him. I trust him more than I used to and that’s important, that bond of trust has to be there. 101

This establishment of trust with the care provider explains inclusion of group visit interventions in a Cochrane review of interventions to build up trust. 56

Programme theory 3: ‘learning by doing’

Kirsh and Aron also identify the importance of ‘learning in context’. 131 Although this is not a complete match with our concept of ‘learning by doing’, it does share mechanisms by which what is being learnt becomes familiar and therefore no longer carries a connotation of anxiety. We consider that ‘learning in context’ would more appropriately characterise a home-based intervention. In contrast the type of activities that we characterise as ‘learning by doing’ (e.g. taking blood glucose or blood pressure measurements) within a group clinic setting become familiar from experience and support. Although not by any means a home environment per se, the group clinic becomes a ‘safe environment’ where an individual can trial an activity and seek recourse to help before incorporating the activity into his or her independent self-management routine. Trialability is considered an important characteristic that impacts on people’s likelihood to contemplate change.

Programme theory 4: ‘monitoring as ownership’

Fundamental to a group clinic approach, as exemplified by SMAs, is the ‘combination of witnessed and private individualised interactions between patients and their physicians, as well as an educational component’. 132 In the RCT by Edelman et al. ,69 pharmacists and physicians developed individualised plans for alterations in medication and lifestyle management, apparently before meeting with the patients themselves. However, there is sufficient evidence relating to the principles of co-creation to suggest that more effective behaviour modification will result from patients generating their own plans, with a provider simply helping to facilitate. In this context, the involvement of patients in their own monitoring, particularly where this requires hands-on engagement with monitoring equipment, may be both a practical and a symbolic way of getting them to start to engage with their own management.

Programme theory 5: ‘acquiring problem-solving strategies’

The qualitative studies demonstrated a clear role of the group clinics in the context of problem-solving. First, patients were reassured by being placed in a group of people who shared the same problems:

Well it was quite nice being in the group, because you kind of think, well other people have got the same sort of problems, you’re not completely weird!97

Ostensibly, problems do not seem to be conceived as problems if they are issues that other participants also have to face on a day-to-day basis. However, aside from such reassurance, there is a strong line of argument regarding the problem-solving function of the group clinic meetings:

Yeah it’s beneficial in a group from the point of view you’ve got someone to talk to, you’ve got an exchange of ideas or problems or whatever. Whereas if you sit there on your own you’re basically waiting for the clock to tick round to say, ‘Well I’m finished now’. 97

Furthermore, patients gain reassurance from other patients having their own problems resolved:

[Y]ou can follow the other people and you can see what the doctor is doing for their problems . . . we can see where we’re heading and try to stop it before we get there. We know we’re going to get there eventually but we want to slow down getting there. 101

Programme theory 6: ‘gaining information’

The gaining of information is seen by many patients as one of the primary purposes for participating in a group situation. This information may relate to the technical aspects of the condition or how a treatment works, or it may relate to how people cope practically or emotionally with their condition:

There was evidence that participants shared useful information with each other, particularly about managing the arthritis on a daily basis:

Someone says ‘Oh well I find if I lay this way or do that it eases it’ and, of course, it all helps everybody . . . 97

It may relate to facilities or aids that can help to manage the implications of their condition or how to navigate health services or other facilities:

They would exchange ideas, their own experiences, how long that they’d had the condition, how, you know, how much support locally they had, or not (laughs) . . . often they would say ‘Which doctor do you see here? My doctor says this,’ because they might see different consultants in this hospital. 97

They’ll say, ‘This was a nice swimming pool and it was easy to get to’ so it sort of spreads . . . And how the shop-mobility works . . . all sorts of things, really, which you wouldn’t actually have if you were sat on your own in a cubicle. 97

Programme theory 7: ‘legitimising question asking’

In the context of the group clinic patients may feel more empowered to ask questions than they might otherwise be during an individual consultation. The fact that others ask questions during a group session, together with the potential modelling of how questions should (or even should not) be asked, can encourage individuals to feel that asking questions is legitimate. However, there is a corollary because people may be discouraged from asking questions either because the topic is not one they feel comfortable sharing with a group or because they are generally reticent in a group situation. Good facilitation skills are required for the group process so that no question is considered ‘too stupid’ and that individual contributions are valued by the group. The comparative comfort with which patients may ask questions within the group will also depend on the comparator, namely how comfortable they have felt in corresponding one-to-one situations with a health-care provider.

Programme theory 8: ‘information seeking surrogacy’

There is significant evidence to suggest that, in the time-pressured environment of the individual consultation, patients often forget to ask questions that concern them. 133 Furthermore, even if they do remember to ask pertinent questions they often forget the answers that they have been given. 134 Being present when others are asking questions may have several effects:

1. Someone may ask a question that addresses an issue that concerns another patient.

2. A question asked by someone else may prompt a patient to remember a related question that concerns him or her.

3. The asking of any question by someone else legitimises the question-asking process.

4. Observing the question-asking-and-response process may provide a less pressurised environment for the taking-in of information relating to the condition.

The corollary to this is that a patient may be deterred from asking a question in a group setting because of the number or characteristics of the other group members, they may consider their question less important or more trivial than other questions asked by members of the group and they may become more passive in the role that they choose to assume in the question-answering process.

Good facilitation skills are required by the provider in order to elicit questions from patients who may be reticent, to manage the influence of more vocal patients and to correct any misinformation that may arise during the group interaction.

Programme theory 9: ‘looking for warning signs’

One of the functions of the group clinic from a clinician’s viewpoint is that it offers the possibility of unobtrusively observing and monitoring a group of patients and thus of ‘triaging’ those who require specific follow-up. This feature would be particularly important in a model in which individual consultation is not universal but, rather, is reserved for those for whom it is indicated and/or for a selected population with particularly complex or heavy requirements. Although in practice this type of group observation differs little from the observation that might take place during an individual consultation, it is interesting to find that the participating clinician may frame it differently:

It [the GMV] creates an environment that is the trickery in medicine – to think people are having a social gathering and you’re working the crowd and doing the medical work while they are having a good time, I mean that’s optimal. 101

Although the overall impression from this data extract is that the ‘trickery’ is in the patients’ best interests and that it is justified by the fact that the patients are enjoying themselves, this type of comment further illustrates the importance of setting initial expectations about how information gathered through the group component of the process will be used.

The same clinician then seeks to explain how such trickery might work:

. . . people are enjoying themselves . . . and yet there’s a team going around getting all the information that needs to be gleaned . . . that’s the secret, so you turn it into a really positive experience for the patients so that’s why they want to keep coming back . . . 101

Programme theory 10: ‘gaining understanding’

Although much of the rhetoric of the group clinic literature relates to efficiency for the provider and improved information and social support for the patient, we were able to identify qualitative benefits to the health-care professional in terms of their improved understanding of the patient’s situation, the constraints of his or her condition and, specifically, issues relating to his or her medication or wider treatment.

I think that it [the GMV] has helped me to be more creative in looking at ways to meet people’s needs. Some of that just comes from the patients themselves because they often have some really neat ideas about how to overcome challenges or difficulties in dealing with the diabetes. 101

Furthermore, the group clinic interaction also served to enhance providers’ skills and awareness:

Through interaction with patients, providers reported having gained a more advanced communication repertoire, and developed greater self- and situational awareness. 101

Programme theory 11: ‘observing a difference’

One of the motivations for attending a group clinic is observing a difference that is perceived to have the potential to make a difference to a participant’s own life. Such a difference may be seen in a reduction in unhelpful behaviours:

A number of the patients mentioned that they had reduced their use of pain killing drugs as a result of participating in the acupuncture clinics:

I took it upon myself to reduce the medication . . . I have reduced it by 50% so, you know, that is a big difference, but ultimately I want to not be taking diclofenac at all. 97

In other cases patients may observe a change in their underlying health condition and be encouraged to persist. Observability is considered another characteristic that impacts on people’s likelihood to contemplate change.

Programme theory 12: ‘modelling positive behaviours’

One of the putative mechanisms for effect in a group context is that other group members may model the desired behaviours required from the patient group and, therefore, participants will adopt the desired behaviour. This is supported at a theoretical level by the various social theories itemised in Chapter 1. We found some empirical data to support this effect. For example:

. . . participants specifically mentioned the usefulness of meeting role models—women who were successful in coping with their disease. 127

We always know we’re not the only one in that boat, when you’re in a lot of pain you think, ‘Oh I don’t know, is it just me, am I exaggerating? Is it mental? ‘ like this. And you see how everyone else suffers and how they cope with it. 97

Programme theory 13: ‘normalising on group behaviour’

Kirsh and Aron identify an important mechanism as ‘motivation to comply with others’. 131 This corresponds quite closely with our conception of ‘normalising on group behaviour’. A group member can establish a benchmark against which they can critically appraise their own behaviour. Although this mechanism is linked with the idea of other group members ‘modelling positive behaviours’, which may then encourage an individual group member to comply with others, there is some evidence that models may exhibit realistic, and hence reassuring, behaviours that might allow a person to aspire to slight but feasible behaviour modification rather than to a more dramatic and thus less attainable change. A potential adverse effect of the concept of the ‘opinion leader’ as such is that the very characteristics that make them stand out as an opinion leader may be the same characteristics that make their example seem unattainable to an ‘average’ person targeted by a group intervention.

A further issue relates to the implied need for homogeneity within the group in order to harness shared norms and values. This issue, which is apparent in web-based articles, has not been explored in depth in the research literature. It is conceivable that minority interests in the group may be overlooked or neglected and that the minority individual may feel marginalised. Interestingly, in the study by Raballo et al. ,135 the concepts most used by patients with type 1 diabetes to define group visits were as follows: ‘comparing,’ ‘knowledge,’ ‘educational’ and ‘friendship.’ In patients with type 2 diabetes, the group visit resonated with ‘friendship,’ ‘I feel good,’ ‘I like this,’ ‘I learn’ and ‘interesting.’

Programme theory 14: ‘signalling importance’

As illustrated in the UK evidence (see Chapter 3, Results of the review of the UK evidence) there are data to suggest that one mechanism for engagement relates to signalling to patients, and indeed to clinicians, that the group clinic, and by implication the activities that take place there, are considered important. These ‘signals’ may be literal (i.e. in the communications sent to the patients) or tacit (e.g. in the premises and activities assigned to the group clinic activity). There is reason to believe that there may be an asymmetrical effect in operation, in that negative perceptions of the premises may have a more powerful effect in deterring attendance than positive perceptions of the premises might do in encouraging attendance. However, this needs further investigation.

Programme theory 15: ‘making a difference’

For clinicians, persuasion that group clinics can make a difference is important if they are to contemplate the not inconsiderable organisational and professional adjustments that may be required for such clinics’ successful implementation. The qualitative studies appeared to indicate that clinicians were monitoring whether or not the group clinic interventions were making a difference and this had a positive effect on their own belief in the intervention. When clinicians witness the achievement of the group clinic approach against an implied inability to engineer change, they are moved to contemplate the advantages of the intervention:

They will self-manage with the group. It isn’t just me sitting telling you what to do. They hear from their peers . . . people will change . . . I could tell them ten times and as soon as somebody beside them with the same condition tells them to do it they listen . . . 101

This statement indicates the influence of observability and relative advantage, both of which are important characteristics when someone is contemplating behaviour change.

Programme theory 16: ‘joining up care’

Group clinics are perceived by some staff as an opportunity to develop shared team approaches to patient care. 100 In a UK context a further mechanism relating to joining up care relates to bringing together the clinical consultation activity and the group education activity that have previously existed separately. Such co-ordination may result in potential efficiencies but may also be seen symbolically in signalling the importance of a co-ordinated approach to chronic disease management. Joining up care is therefore not simply about bringing the two activities together but about emphasising their genuine partnership as activities of complementary importance.

Summary of consideration of theory

Methods

The methods for this section are found in Chapter 2.

Results of the literature search

The results of the three-stage literature search are presented in Table 29. The analysis of costs used eight studies.

What are the key elements in examining the costs of group clinics?

What evidence exists for the costs of group clinics?

Discussion

Group care is more expensive to set up and run, and although not many studies have actually calculated these increased costs, they have reported increased use of physician time, increased educational resources, increased frequency of appointments per patient and the existence of one-to-one appointments for patients on group care, all of which will increase costs when compared with usual care. The lack of information relating to the costs of the intervention in the studies we examined means that it is challenging to draw conclusions about the cost of group clinics. From the data from the RCTs, we can understand more about the key cost elements of group clinics. However, this information would need to be considered in a full economic analysis in order to be meaningful.

From the studies we examined, we can make better judgements on the cost savings as a result of patient participation in group clinic interventions. The majority of studies examined addressed the changes in utilisation and the subsequent changes in costs. There was a mixed pattern of changes in utilisation, with some studies reporting that intervention patients used fewer health services overall while others reported an increase in some areas (primary care, inpatient and outpatient). This mixed pattern was repeated in the assessment of changes in costs; understandably, in studies where utilisation decreased, there was a decrease in costs. With this mixed set of results, it would not be meaningful to cluster studies together in terms of utilisation and cost changes.

It would have been informative to identify whether or not the savings identified are realised over a longer period of time. We found evidence to suggest that the US health-care system reimbursement process means that these interventions will always be delivered in a standard way to ensure that insurance claims are reimbursed, thereby making costs across interventions (although not cost savings) more uniform. It may be possible to hypothesise that as group clinics become more widespread, staff costs will decrease as more staff become trained (training being a major part of establishing a group clinic, as identified in the main review).

Clancy et al. 58 aimed to determine why costs were higher for intervention patients than for control patients. In addition to small sample sizes, they note that participating in an intervention such as a group clinic might ‘activate’ patients who had previously missed care to catch up with the care that they had missed, thereby increasing health service utilisation. In addition, the length of study is important: improved self-care (which is often an outcome of group clinic interventions) may have a time lag, and so for a shorter study 6 months is not sufficient time to demonstrate a decrease in utilisation and, therefore, to observe a decrease in costs.

Summary of included studies

Our assessment of costs and feasibility across a heterogeneous set of studies has showed mixed effects of group clinic interventions on costs and savings. A full economic analysis of group clinics, along with the robust collection of costs data alongside group clinic interventions, is recommended. A full economic analysis could allow for data included in RCTs, such as the type of clinician delivering the intervention and how long each group clinic lasts, to be costed in order to get a more complete picture of the costs of group clinic. Primary research assembling information on the running of group clinics and the costs that are saved specifically within a NHS setting would be essential to inform decisions about group clinic provision in a UK context.

Overview

Feasibility conflates many issues, such as acceptability to patients and providers, practicality in terms of required procedures (whether alongside or as a substitute for existing practice) and affordability, in terms of financial considerations and available equipment and premises. The evidence to be mapped against this domain is drawn from qualitative studies of provider and patient attitudes, implementation studies not otherwise included in this review and an overall picture of likely cost-effectiveness, as has emerged from Chapter 4. Feasibility includes general issues to be considered within any context for implementation of group clinics and specific issues relating to implementation within a NHS context.

What are the key considerations regarding feasibility?

Key to a consideration of feasibility in this context is affordability. Although claims are made of cost savings these are based on either (1) US studies of limited geographical or temporal relevance or (2) a simplistic argument of more patients seen by a clinician per hour. In particular, there is limited evidence of cost implications in a UK study. Indeed, although the insights from group acupuncture clinics are informative in terms of the group interactions and dynamics within a UK context, the actual assessment of costs would be potentially misleading. As will be explained later in this report the achievements of the group acupuncture clinics are located within a ‘work smarter’ treatment delivery model. These otherwise promising achievements have, therefore, limited relevance to the monitoring model that is fundamental to group clinic provision.

A further concern relates to acceptability. Our clinical advisers point out that there is a strong expectation in the NHS of being seen by a specialist clinician within an individual consultation. Even if group clinic provision were to become the default position, a sizeable proportion of the population would still require access to the more traditional model, perhaps because of the complexity or severity of their condition or because they would demand it through exercising patient choice. Such preference may be affirmed on commencement of treatment or, as illustrated by UK group acupuncture clinic qualitative data, may emerge following patients’ experience of the group clinic provision. In particular, the willingness of patients to try a new modality of service provision should not be interpreted as those patients’ commitment to that service modality on a long-term basis.

Practical issues relate to the requirement to be able to access all patient records and results in advance of a single SMA. This may place a burden on diagnostic services but may also prove problematic for the individual specialist, who would have to make time for review of the notes. The latter factor is examined in a US context of uncompensated clinician time. 132

Other feasibility concerns relate to the need for clinician training, particularly in group facilitation, and the need for suitable premises. Within the wider picture of feasibility it would be worth exploring whether or not the individual components considered essential to the group clinic approach could be delivered in an alternative format. For example, in some circumstances the socialisation or the interaction with a group facilitator could be offered virtually, offering the opportunity for the clinical team to identify those needing particular help.

What evidence exists on the feasibility of group clinics?

Little evidence exists on the feasibility of group clinics, even though much of the literature suggests how they might be introduced. Particularly noticeable is a shortage of data from the UK. The wider non-NHS-specific literature informs such aspects as implementation and confidentiality. A feasibility study104 revealed such positive aspects of GMVs as personalised attention (77%), self-care education (69%), access to medication refills and examinations (69%) and advice from peers (62%). Negative aspects included insufficient personal attention (23%), logistical barriers (8%) and loss of confidentiality. 104

Kirsh et al. 27 have explored implementation issues relating to SMAs. They identified such important promoting factors as the formation of a core team committed to quality and improvement with strong support for the clinic leadership from other team members. Notably, tailoring had to take into account such ‘key innovation-hindering factors’ as limited resources (such as space), potential to alter longstanding patient–provider relationships, and organisational silos (disconnected groups) with core team members reporting to different supervisors. The last point emphasises that group clinics should be seen not in isolation but as a potential vehicle towards interprofessional team working, with all of the associated culture changes that this might necessitate.

Concerns relating to confidentiality were raised consistently in the reviewed literature. This issue was examined specifically in a study by Wong et al. 107 This study aptly highlights that GMVs can impact on the clinician–patient relationship as patients are ‘able to draw upon more informational resources and social support from attendees and often feel more empowered to pose questions to their providers than they might otherwise in individual encounters’. 107 However, providers reported that ‘the most common reason for not attending a GMV was patients’ concerns about confidentiality and hence a preference for individual visits’. 107 Nevertheless, one overall finding from the study was that patients who did attend a GMV consciously selected which information they were comfortable sharing in a group situation. 107 Although it could be perceived as a drawback that patients filtered the information they felt able to share, some interventions included a discussion of confidentiality with practical examples as a component of the initial group clinic sessions.

Discussion

The review team has identified specific concerns relating to the interpretation of predominantly US data within a specific UK context. In particular, many of the interventions have been delivered within the context of health-care financing that determines the exact configuration of approved packages of group clinic provision and, for example, requires guaranteed access to an individual consultation if requested. Advice from our clinical advisers suggests that a model in which an increasing amount of the content of the previous individual consultation is assumed in a group context, facilitated perhaps by a member of staff who is not the specialist clinician, may be an alternative form of substitution. This might facilitate shorter individual consultations, although this issue remains to be investigated. Importantly, however, such provision would need to be in a context where group education is seen as more central to the chronic disease management process and not as an optional extra.

Summary of included studies

Although the evidence from the USA and that from a UK group acupuncture clinic context does inform a discussion of feasibility, a specific need remains for further investigation of the monitoring model of group clinics within a UK context. This research requirement sits naturally alongside the suggestion made in the previous chapter for a full UK-centric economic evaluation and the need for qualitative exploration of the attitudes of NHS patients, providers and caregivers towards group clinic provision. In addition, there is a requirement to explore the feasibility of ‘substitution’ of specific functions from the individual consultation with a corresponding group-based provision, along with any training and role development issues this might occasion.

Health outcomes

By far the majority of studies examining clinical outcomes relate to diabetes and focus on basic biomedical outcomes that are relatively easy to monitor routinely. It is, therefore, difficult to extrapolate these effects to other chronic conditions.

Health service outcomes

Perceived and actual benefits and disadvantages of a group consultation when compared with an individual consultation

Although crude analyses compare the number of patients seen within a group session with the number of those seen individually in the same time period, such an approach is inadequate for the purposes of a rigorous evaluation. There is substantial evidence that provision must be made for individual consultations and also that costs may be displaced to other parts of the health-care system. The cost of individual consultations must factor in the provision for such consultations within the group session, for those who are displaced to sessions outside the group clinic and for those for whom group provision is either inappropriate or unacceptable. Although one assumption encountered in the literature is that the reduction of health costs will take place over an extended time period, studies that have been conducted to date have not covered a long enough evaluation period to demonstrate this realisation of cost benefits.

The value of group education

The cost benefit of group clinic approaches depends on whether or not current provision (usual care) includes existing group education provision and, specifically, whether this is delivered by health professionals or by lay peer supporters. Group education has been found to have an effect on some of those biomedical measures addressed by group clinics but not typically to the degree realised by most group clinic studies. The cost issue, therefore, becomes ‘what is the demonstrable cost–benefit to be realised by delivering the specific group clinic intervention compared with the individual consultation plus group education sessions?’. As indicated by some of the foregoing, this question is complicated by the issue of what can be quantified as benefits. In particular, is the evaluation framework to be exclusively that of cost savings – in which case group clinics are unlikely to deliver against this agenda – or is the evaluation to be situated in the context of joined-up improved-quality interprofessional care?

The value of multiprofessional approaches resulting from simultaneous clinical involvement

We found some evidence that involvement in group clinics may have accrued particular advantages in relation to interprofessional team working and mutuality:

[T]he flexibility of the individual team members is manifest during the SMA sessions. A weekly meeting . . . continues to occur to discuss patients and processes to assure that all team members have an open forum to voice concerns and make group changes. 7

The literature around uni-professional, multi-professional and inter-professional working emphasises flexibility of roles and a degree of interchangeability as the means by which interprofessional working might be achieved. 138

Under what circumstances do patients with chronic conditions agree to participate in group clinic approaches?

From the theoretical literature, we have identified four principal components of a group clinic approach, as follows:

1. Monitoring: this is a traditional activity in the individual consultation but there is some evidence that group clinic approaches may turn this into more of a shared activity between the patient and clinician, with the patient becoming involved in some of the monitoring activities.

2. Self-management: the group clinic approach encourages patients to become more active in managing their condition. In contrast to an individual consultation, the group-based approach may offer both role models in those who manage their own condition and tips on techniques and resources acquired from fellow patients.

3. Peer support: this is a completely discrete activity from the individual consultation and one which offers additional sources of support beyond the clinician and the significant others of the patient. Commonly in the UK there is a separation between clinical activities and group education approaches.

4. Education and information: quantitatively, there is the opportunity for the clinician to share information with several patients at the same time, reducing duplication and repetition, and resulting in greater consistency in information provision. Qualitatively, patients may respond better to information shared in a less didactic manner or to information originating from fellow patients. More reticent patients may benefit vicariously from questions asked by more proactive members of the group, in effect becoming ‘lurkers’ within the group.

Typically, patients with chronic conditions appear to make an overall assessment of the benefits of participation before agreeing to participate. There is some evidence that the disadvantages of participation are not adequately explained to participants by clinical staff. A significant proportion of those invited to participate decline, largely because they do not recognise the benefits of group clinic approaches against the perceived advantages of an individual consultation. Expectations of being seen in an individual consultation, whether specified by a health plan as in the USA or through cultural conditioning in the UK, appear to militate against the use of a group clinic approach. Alternative provision will probably be required for this sizeable group of patients and the very availability of such an alternative may have a negative effect on uptake.

Having agreed to attend group clinics, why do some patients with chronic conditions decide not to attend any group clinic sessions?

Constraints related to the logistics of attending the group clinic appointment (e.g. timing, other commitments, etc.) play a major role in determining whether or not patients with chronic conditions will attend. However, these same constraints are also present for those seeking individual appointments. The primary consideration, therefore, appears to be the flexibility of attendance patterns. In particular, attendance may depend on whether group clinics employ a closed cohort-based approach or more of a drop-in model.

Having started to attend group clinics, why do some patients with chronic conditions discontinue a group clinic programme?

There is some evidence to suggest that some patients will attend with a specific goal of receiving sufficient information for self-management of their condition. Once they feel that they have obtained this information, their motivation for attendance wanes. For others, the social aspect of a group is particularly important and this may contribute to their motivation for ongoing attendance, even if the other benefits of attendance degrade over time. Finally, there are others for whom the sense of shared community persists, which is recognised from their transition from being beneficiaries to becoming donors to the overall group process.

Which is the most sustainable model of group clinic delivery?

The identified research literature does not support a detailed analysis of sustainability. Most initiatives were evaluated over only a relatively short time period. For example, Cohen et al. claimed to have demonstrated ‘that the pharmacist-led group intervention program was an efficacious and sustainable collaborative care approach’ and yet only evaluated the initiative over a period of 2 years. 63 In fact, within the context of group clinics such an evaluation period is comparatively long. Housden et al. 53 reported that 15 out of 26 studies were 12 months or shorter in duration, and six studies were up to 2 years in duration. The study with the longest duration followed patients for only a period of 5 years after the intervention. We conclude with Housden et al. that:

. . . the long-term or sustainable outcomes of group medical visits are unclear, and it is difficult to know if the outcomes were maintained for a substantial length of time after the intervention. 53

Qualitatively, there is very little discussion in the published literature about the practicalities of managing different models of group membership. Such contrasting models have considerable implications for facilitation, educational content and the group dynamic. These are now briefly discussed, together with their possible implications.

Feasibility (evidence from qualitative research, cost studies and UK studies and informants)

UK informants highlighted a current separation between the clinical consultation and the provision of group education, as evidenced in diabetes care. Even within existing UK provision the coverage and quality of group education is believed to be extremely variable. Wider issues relating to feasibility concern appropriate premises for delivery97 and training in facilitation skills for participating clinical staff. 101

Appropriateness (evidence from qualitative research, UK studies and UK informants)

The evidence for the appropriateness of a group clinic approach, as perceived by patients, is largely equivocal. Patients report substantively comparable perceptions of improvement across both group and individual interventions, with the groups sharing concerns about appointment availability. There was little evidence of dissatisfaction with care from those actually receiving group clinic approaches. However, other considerations may result in patients’ poor adherence to the group clinic regime. More typically, those who had vocalised reservations regarding group clinic approaches expressed this concern by withholding consent to entry into a group intervention arm. We have made some initial observations based on data available on recruitment and maintenance from included studies. This suggests that any assessment of effectiveness should pay close attention to those who, although eligible, exit the intervention prior to its commencement. In practical terms this population will require alternative health-care provision, which may make a dual model of service delivery particularly problematic.

The perceived advantages of group-based approaches include greater flexibility in duration of appointments and more time with the clinician. 97,128 Improved flexibility is expressed in the fact that a group clinic ‘can be altered to fit various patient populations, specific physician practices/organisations, and a number of health care delivery systems)’. 48 Recent qualitative evidence suggests that the group clinic approach may have a beneficial effect in terms of challenging the previous clinician–patient dynamic, thereby producing a ‘levelled playing field’. 101

Corresponding disadvantages include a perceived lack of privacy, although this was not found to be a significant problem in existing UK studies (albeit in the context of group acupuncture clinics). Of significant concern, however, is the fact that participants attending individual sessions perceive little apparent advantage in switching to a group-based approach and report difficulty in imagining how such group-based approaches might be feasible.

Given the outstanding questions about the sustainability of group clinic approaches, the severity of attrition and the lack of long-term outcome measurement (with the longest follow-up being 5 years),84 it is problematic to consider the use of group clinics in the context of lifelong chronic disease management provision.

Group clinics may not be appropriate for certain patients or under certain circumstances. In addition to religious and cultural considerations, as exemplified by the specific reference to Muslim women made by a nurse in one study (see Chapter 3, Results of the review of UK evidence, Study analyses, How do UK health providers feel about group clinics?), group clinic approaches may be less acceptable to patients of an older generation or in instances where there is a perceive threat to patients’ dignity or an increased likelihood of embarrassment (e.g. the revealing of unsightly varicose veins, etc.). Although concerns regarding privacy are underplayed in the included studies, this may reflect the types of conditions being reported and a focus on those who have consented either to randomisation (for RCTs) or to a group intervention (for other research designs). Some concerns may be alleviated by such factors as the design of single-sex group sessions or groups offered to particular ethnic populations, although providers should be sensitive to the fact that such measures may not necessarily result in the addressing of all of the target population’s concerns.

Other considerations regarding appropriateness were implied by the exclusion criteria employed in the included trials. Many trials have purposely excluded patients with dementia or cognitive impairment. Others excluded those with hearing difficulties or other communication-related constraints. The exclusion of patients who do not speak the predominant language was also evident. For qualitative studies it was less obvious whether such exclusions relate to the specific group nature of the intervention or were a function of the methods of investigation. In either case, it is clear that the group clinic approaches are not suited to particular segments of the population. For other patients, concerns about access and attendance, for example in the case of those who do not have their own transport or those who work during clinic hours, are also evident.

Meaningfulness (evidence from UK studies and UK informants)

Individuals using the NHS have, as a marker of good-quality individualised care, a general expectation that they will receive an individual consultation with a clinician. This impression may be strengthened by use of the word ‘clinics’ and by the fact that several patients will have specific expectations of the group clinic, notwithstanding any information provided. Furthermore, group education in the NHS is seen by some as an alternative to the individual consultation but typically presented as a ‘bolt-on extra’ and, as a result, could be regarded as optional or less important by patients and/or health providers.

Effectiveness (evidence from systematic reviews and randomised controlled trials)

We identified 13 systematic reviews, including multiple variations of the GMV. Ten of these studies were analysed in detail, while one was at the protocol stage and another was unavailable and used only in summary form. The majority of these reviews are disease specific, with the primary focus being on diabetes. One Cochrane review included two studies of group visits as interventions designed to increase patient trust in their clinician, one of the putative mechanisms of the group clinic effect.

Taken as a body of evidence, the reviews shared common conclusions:

• evidence of a significant positive effect in terms of HbA_1c and systolic blood pressure

• non-significant effects in relation to LDL, HDL and total cholesterol

• a significant effect in relation to disease-specific quality of life

• a moderately significant effect on generic quality of life.

• equivocal evidence in relation to potential cost savings.

Many of the reviews concluded that the heterogeneity of group clinic-type interventions made it problematic to classify such initiatives, to isolate the effect of specific intervention components and, subsequently, to evaluate their effects.

We identified 22 RCTs (32 papers) published between 1999 and 2014. A total of 17 of the 22 studies were conducted in the USA,20,31,58,60,63–66,68,71,72,75–77,79,80,86 with two conducted in Italy,82,85 two in the People’s Republic of China73,74 and one in Norway. 70 Included studies recruited a total of 5572 patients. Diabetes was the most represented condition, being present in 23 of the 31 papers and representing, in turn, 15 of the 23 RCTs. 20,58,60,63,66,67,69,70,72,74,79,80,82,85,86 One further study was conducted in a pre-diabetes population. 64 Other conditions included asthma,65 cardiovascular disease, heart disease/hypertension (three RCTs),31,71,73 stroke/transient ischaemic attack, and Parkinson’s disease. 68

We found six trial reports64,66,67,73,74,76 (five trials) published over the period 2012–14. Only one previous review55 had included any of these reports (n = 1).

Efficiency (cost-effectiveness) (evidence from cost studies)

The evidence for the cost-effectiveness of group clinic approaches is equivocal. The efficiency of group clinics is determined by the perspective from which the group intervention is being examined, the level of current (comparator) provision and whether or not there is recognition of a need for the provision of such enhancements as training for clinicians (e.g. to act as facilitators) and accommodation for group activities. A full economic evaluation is required in a UK setting with recognition of the factors described above regarding feasibility and the other realities of implementation.

The substitution argument

An initial attraction of group clinic approaches, as encapsulated in our review protocol, was the assumption that such approaches might offer a viable alternative to and substitute for individual consultations. In reality, many models make routine provision for individual consultations, offer follow-up consultation on demand or use the group setting as a mechanism for singling out those requiring specific support. The implications of these three different approaches are similarly varied. In the first instance efficiencies are gained only to the extent that the information giving that would have taken place in an individual setting is provided in a group setting and the corollary that the duplication of genuinely shared concerns may be commensurately reduced. In the second instance, the numbers of on-demand consultations may be difficult both to predict and to provide for, with the consequent dangers of under- or overutilisation of clinical staff and, in the latter case, decreased patient satisfaction. The third variant, whereby those requiring an individual consultation are ‘triaged’ through the group processes, is heavily dependent on the clinician’s capacity to identify genuine need amid a preoccupation with group processes and facilitation. Perversely, those least likely to communicate or engage in a group setting may be the very patients who are most need of supplemental individualised care.

We found no compelling evidence that, within the context of the entire health system, the group clinic approach offers efficiencies over the usual care system. Considerations here are that a large proportion of patients will not take up group clinic provision – either because of initial preferences or following personal experience of the approach – and will require individual consultations. Furthermore, the large majority of group clinic approaches make provision for individual consultations within the model, with additional cost consequences. Investigation of this phenomenon, which ran counter to the original perceived rationale for conducting this review, revealed that this may be primarily an artefact of funding arrangements in the USA, where most evaluations have taken place. For example, Blue Cross/Blue Shield Corporate Reimbursements will not cover group visit (SMAs) if the patient is not able to access an individual consultation with the physician and both individual and group interaction must be documented in the patient’s medical record to receive reimbursement. 140 Detailed evaluation in a UK setting is required in order to assess the proportion of patients who would avail themselves of an individual consultation in addition to the group interaction or who would find a group clinic unacceptable.

The quality of care argument

The achievement of positive biomedical or associated outcomes is variable. Although it is conceivable that ongoing self-monitoring, allied to hands-on experience of aspects of self-care and the positive support of realistic models and peers, may achieve a beneficial effect, it is unclear if group clinics are the optimal method for harnessing such mechanisms.

The acceptability argument

Although concerns over confidentiality and privacy are not as plentiful as might be expected, it must be recognised that the views of those who are not willing to enter into a group clinic trial at all are imperfectly captured by either quantitative or qualitative studies. In addition, individuals may be able to enter a group clinic arrangement on an experimental basis but may subsequently decide that the experience was not positive enough for them to continue such an approach beyond the lifespan of the trial. Indeed, there is little evidence on the sustainability of group clinic approaches.

The enhancement model

Typically, group clinic approaches have been investigated as an alternative to individual consultations. Comparisons between different types of group intervention of differing intensities and with/without clinical input are required to examine the differential benefit of the added group clinic-specific input. Considerations for the feasibility of group clinics may centre on whether or not group clinics are seen as an entirely new intervention or whether or not they represent a means of systematising and joining up existing group education and individual clinician input and, therefore, placing group education provision in a more central role than currently appears to be the case.

Search strategy

1. group visit$.tw.

2. group clinic$.tw.

3. *Group Processes/

4. group appointment$.tw.

5. group care.tw.

6. group meeting$.tw.

7. group medical visit$.tw.

8. group medical clinic$.tw.

9. group medical appointment$.tw.

10. group medical care.tw.

11. group medical meeting$.tw.

12. gmv.tw.

13. gma.tw.

14. shared medical appointment$.tw.

15. shared medical visit$.tw.

16. cluster visit$.tw.

17. (group outpatient$ adj1 (visit$ or clinic$ or appointment$ or meeting$ or care)).tw.

18. or/1-17

19. limit 18 to (english language and yr=“1999 –Current”)

Search strategy

1. group visit$.tw.

2. group clinic$.tw.

3. *group process/

4. group appointment$.tw.

5. group care.tw.

6. group meeting$.tw.

7. group medical visit$.tw.

8. group medical clinic$.tw.

9. group medical appointment$.tw.

10. group medical care.tw.

11. group medical meeting$.tw.

12. gmv.tw.

13. gma.tw.

14. shared medical appointment$.tw.

15. shared medical visit$.tw.

16. cluster visit$.tw.

17. (group outpatient$ adj1 (visit$ or clinic$ or appointment$ or meeting$ or care)).tw.

18. or/1-17

19. limit 18 to (embase and english and yr=“1999 –Current”)

Search strategy

ID Search

#1 “group visit*”:ti,ab,kw (Word variations have been searched)

#2 “group clinic*”:ti,ab,k this term only

#4 “group appointment*”:ti,ab,kw (Word variations have been searched)

#5 “group care”:ti,ab,kw (Word variations have been searched)

#6 “group meeting*”:ti,ab,kw (Word variations have been searched)

#7 “group medical visit*”:ti,ab,kw (Word variations have been searched)

#8 “group medical clinic*”:ti,ab,kw (Word variations have been searched)

#9 “group medical appointment*”:ti,ab,kw (Word variations have been searched)

#10 “group medical care”:ti,ab,kw (Word variations have been searched)

#11 group medical meeting*:ti,ab,kw (Word variations have been searched)

#12 gmv:ti,ab,kw (Word variations have been searched)

#13 gma:ti,ab,kw (Word variations have been searched)

#14 shared medical appointment*:ti,ab,kw (Word variations have been searched)

#15 shared medical visit*:ti,ab,kw (Word variations have been searched)

#16 “cluster visit*”:ti,ab,kw (Word variations have been searched)

#17 “group outpatient visit*”:ti,ab,kw (Word variations have been searched)

#18 “group outpatient clinic*”:ti,ab,kw (Word variations have been searched)

#19 “group outpatient appointment*”:ti,ab,kw (Word variations have been searched)

#20 “group outpatient meeting*”:ti,ab,kw (Word variations have been searched)

#21 “group outpatient care”:ti,ab,kw (Word variations have been searched)

#22 #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 or #17 or #18 or #19 or #20 or #21

w (Word variations have been searched)

#3 MeSH descriptor: [Group Processes]

Search strategy

Search strategy