The impact of the enhanced recovery pathway and other factors on outcomes and costs following hip and knee replacement: routine data study

Work package 1

Work package 1 aims to explore geographical variation in patient outcomes of surgery. Using Hospital and Community Health Service (HCHS), KH03 and Supporting Facilities data sets linked to the National Joint Registry (NJR), Hospital Episode Statistics (HES) and patient-reported outcome measures (PROMs), we explore how hospital organisational factors (e.g. staff, beds, operating theatres) and surgical factors (e.g. minimally invasive technique, surgeon volume, operative time, implant fixation, thromboprophylaxis) explain geographical variation in patient outcomes, adjusting for patient case mix. The results are displayed as maps, highlighting the level of variation in patient outcomes across CCGs.

Work package 2

Work package 2 focuses specifically on the enhanced recovery care pathway.

National Joint Registry, Hospital Episode Statistics and patient-reported outcome measures

Approval for NJR data was received on 8 July 2016 (NJR internal reference RSC2016/11). Confidentiality Advisory Group section 251 approval was received on 20 September 2016 (Confidentiality Advisory Group reference 16/CAG/0111). A Data Access Request Service application was made to NHS Digital for HES and PROMs data to be linked to data from the NJR. This was approved, with the data-sharing agreement signed on 23 January 2016. Data from the NJR were received on 14 March 2017. HES and PROM data were made available to download on 7 April 2017.

Clinical Practice Research Datalink, Hospital Episode Statistics and patient-reported outcome measures

Independent Scientific Advisory Committee approval was received on 5 July 2016 (protocol number 11_050AMnA2R). The CPRD subsequently announced a new linkage to PROMs, so a further Independent Scientific Advisory Committee amendment was submitted and approval was obtained on 10 January 2017 (protocol number 11_050 AMnA2RA2). CPRD data linked to inpatient and outpatient HES were received on 9 May 2017 and linked PROMs data were received on 25 July 2017.

Top outcomes

Mid-outcomes

Other important outcomes and areas of discussion

• Choice of pain relief.

• Patient education on both pain relief and physiotherapy.

• Continuity of care.

• Length of time between discovery of problem and operation. This fed into the need to stratify according to preoperative pain and function when looking at patient outcomes, as the time taken to get to surgery can have an effect.

• Managing expectations. A group member explained that a friend who had recently had a knee operation was disappointed with the results, as she expected to be able to go back to line dancing club two or three times a week without pain or difficulty with her knee.

• The group agreed that patient satisfaction and overall outcomes are dependent on the patient’s overall outlook and positivity. It was also thought that, if there is a long time between diagnosis and surgery, then the patient becomes that much older, but the operation is not going to wind the clock back.

The group felt that the maps describing geographical variation in outcome need to be clear to patients about what hospitals include (e.g. some hospitals do not take more complex patients). Some group members explained that they chose certain hospitals because of reputation and ‘word of mouth’. One group member said that they found it difficult to find any information on outcomes when choosing a hospital, but that they made a final choice of hospital on infection rate. It was explained to the group that, for all outcomes, patient data will be adjusted for patient situation and comorbidities.

Data source and sample size

Hospital organisational factors from the HCHS Workforce Statistics, the Bed Availability and Occupancy data set and the Supporting Facilities data set were linked to the NJR, HES and PROMs data via the hospital provider code. Details of the NJR, HES and PROMs linked data have been previously described in Chapter 2. For this analysis, we use the most recent 3 years of data, from 2014 to 2016.

Main outcome measures

The primary outcomes of interest are informed by the findings of the patient forum (as described in Chapter 3). As a measure of patient-reported pain and function, we used the absolute change in OHSs and OKSs. 30,31 Each question is scored between 0 (meaning worse symptoms) and 4 (denoting least symptoms). Scores from these 12 questions are added, getting a total score spanning from 0 (the worst possible score) to 48 (the best possible score). We calculated the difference between the total scores 6 months after the operation and at baseline to obtain a measure of change associated with the surgery.

Further outcomes included the proportion of complications at 6 months after surgery. For a list of the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10),35 codes used to identify complications of surgery, see Report Supplementary Material 1. We defined complications as one or more from the following list: stroke (excluding transient ischaemic attack), respiratory infection, acute myocardial infarction, pulmonary embolism/deep-vein thrombosis, urinary tract infection, wound disruption, surgical site infection, fracture after implant, complication of prosthesis, neurovascular injury, acute renal failure and blood transfusion. For a list of the OPCS-4 codes used to identify blood transfusion complication, see Report Supplementary Material 2. LOS was calculated as the number of days between the hospital admission and discharge date. We estimated the inpatient cost relating to the index episode using NHS Reference Costs from 2015–16. 36 We estimated the mean cost per bed-day based on the health-care resource use [Healthcare Resource Group (HRG)] for each patient and their LOS. For methods to support the estimation of bed-day cost, see Report Supplementary Material 3.

Predictor variables

Exclusion criteria

We included only patients receiving planned surgery (Figure 3) between 2014 and 2016. We excluded patients without information for the 2001 census lower-layer super output area. Patients with missing data for LOS were also excluded. We excluded patients without information on baseline and/or 6-month follow-up OHS/OKS scores for the analysis of change in OHS and OKS.

FIGURE 3.

Flow diagram showing selection of patients for inclusion in this study (navy shows inclusion and light blue shows exclusion). Reproduced from Garriga et al. 37 © 2019 Garriga C et al. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Reproduced from Garriga et al. 37 © 2019 Garriga C et al. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Sample size

To account for clustering within the data (patients nested within CCGs), we need to inflate the required sample size by the design effect [1 + (n – 1)p], where p is the intracluster correlation and n the mean cluster size. From our previous work, the intracluster correlation was 0.0135 for hip and 0.014 for knee replacement. 38 There are 207 CCGs in England and the CCG is the cluster. If we expect to have 100 patients from each CCG group, the design effect is (1 + 99 × 0.014) = 2.4.

For OHS and OKS outcomes, the minimally important difference between groups39,40 has been estimated to be 5, with a standard deviation (SD) of 10. Using a two-sided, two-sample t-test, with 90% power, at a 5% level of significance, to detect a difference in mean OHS and OKS of 5 requires a sample size of 85 in each exposure group. Assuming a 50% response rate to the 6-month follow-up OHS and OKS questionnaires inflates the sample size to 170 per exposure group. The required sample size per exposure group adjusted for clustering is 2.4 × 170 = 408. Several exposure variables will be considered in the model. Including up to 50 degrees of freedom would require a sample size of around 20,000 patients, assuming equal-sized exposure groups. Hence, this is more than adequately powered. We note that this does not account for multiple testing.

For the other binary outcomes, for complications of both THR and TKR, within 6 months of operation, rates of stroke and myocardial infarction were < 0.5%, and of anaemia, urinary tract infection, wound infection and pulmonary embolism/deep-vein thrombosis were < 3%. 41 The NJR annual report shows 90-day mortality of 0.5% and 1-year mortality of 1.5%. Rates of revision are around 5% at 10 years and rates of revision/reoperation are higher, at up to 20%. 42 For the rarest outcomes, to detect a difference in proportions of 0.5% compared with 1%, using a two-sided, two-sample chi-squared test, with 90% power at a 5% level of significance, requires 6650 patients per exposure group. As HES encompasses elective admissions to all English hospitals, we expect no loss to follow-up, as this information would be captured. For the design effect of 2.4 (for which we assume the same intracluster correlation as above), the sample size increases to 15,960 per degree of freedom. Hence, even for these rarest outcomes, with an actual sample size of > 350,000, we can still include over 20 degrees of freedom in the model.

Geographical variables

Within our HES data set, the smallest level of geography available to us is the lower-layer super output area (LLSOA) that a patient lives in. LLSOAs have a population between 1000 and 3000, and between 400 and 1200 households. There are 32,482 LLSOAs in England as of 2001. For the multilevel analysis, we use this as an intermediate level between the patient level and the CCG level. From the ONS Open Geography Portal, we obtained a polygon shapefile for the 2017 CCG areas in England. Maps were produced using these boundaries.

We start by looking at the structure of the geographical data (CCG areas) for a hospital that a patient is treated in. To do this, we use the ArcGIS 10.5 software (ESRI, Redlands, CA, USA) to do a point-in-polygon overlay, providing a spatial join of the hospital points with the CCG areas. This shows us that there are CCG areas that do not have any hospitals within them. For knee replacement, there are 173 CCGs that have a hospital contained within their boundary (175 CCGs for hips). Hence, out of the 207 CCG areas, there are 34 that do not have a hospital in their boundary for hips and 32 for knees. For this analysis, we want a ‘nested data set’; currently, if our data structure is to have patients within hospitals within CCGs, it does not work as there are CCG areas with no hospitals, which, therefore, would appear blank on a CCG map of England when displaying outcomes of surgery.

To have a nested data structure, we instead will base this on the CCG area for a LLSOA that a patient lives in. By doing a spatial join of the LLSOA points with the 2017 CCG polygon file, this provides us with a nested data set of patients (level 1) within LLSOA (level 2) within CCG (level 3) (Table 5).

However, with this structure to the data, we now need to allocate a hospital organisational characteristic according to the LLSOA areas. This is done by using a spatial join based on the distance between two points (hospital and LLSOA). For each LLSOA, we can then identify the nearest hospital and attribute the hospital’s characteristics to that LLSOA area.

Statistical analysis

The hierarchical structure of the data consists of patients (level 1), nested within LLSOA (level 2), within CCGs (level 3). Multilevel regression models are used to describe the association of patient, hospital organisation and surgical factors and patient outcomes of surgery. This controls for evidence of clustering in the data, by allowing outcomes to vary across LLSOAs and CCGs. Failure to control for evidence of clustering can lead to estimates of standard errors that are spuriously precise and a potential source of bias. Analyses are conducted separately for THR and TKR.

The general form of the multilevel model is given as:

where Y_ijk is the outcome variable in patient i, in LLSOA j, in CCG k. CONS is the constant term, the vector of explanatory variables, U_j is the LLSOA residual error term distributed N(0,σj2), V_k is the CCG residual error term distributed N(0,σk2), and e_ijk is the individual-level residual error distributed N(0,σe2).

It is assumed that the set of random effects explains the clustering in the data, such that different observations in the same cluster are independent. We examine continuous outcomes (e.g. LOS, OHS, OKS) using linear regression, and binary outcomes (complications at 6 months) using logistic regression. Normal distributional assumptions are assumed for both linear and logistic multilevel regression models. We use normal probability plots that plot the ranked residuals against corresponding points on a normal distribution curve, to assess our assumption that level 2 and 3 residuals are normally distributed.

In order to generate the predicted outcomes across the 207 CCGs, we first estimate the linear predictor, which is the sum of the betas for each patient i in the data set. The overall predicted outcome in each CCG is then just the mean of the linear predictor for each CCG:

To incorporate evidence of clustering, the multilevel regression model has an estimate of the residual CCG variation. This is additional CCG-level variation in outcomes that is not explained by the variables in the regression model. So, for each of the 207 CCGs, k, we add the mean of the linear predictor to the CCG-level residual:

We fit the following models: (1) null model of the actual observed outcomes; (2) model that adjusts for patient case-mix variables; (3) model further adjusting for surgical variable; and (4) model adjusting for hospital organisational variables. Analyses were conducted using Stata version 15.1 statistical software (StataCorp LP, College Station, TX, USA). Multilevel analyses were imported to Stata from MLwiN version 3.00 (MLwiN Centre for Multilevel Modelling, Bristol, UK), using the Stata command runmlwin.

Predictive variables

Variation in outcomes

Length of stay

The observed mean LOS by CCG ranged between 2.5 and 6.2 days for THRs, and between 2.7 and 6.6 days for TKRs. Variability across CCGs was high, with 54 out of 207 CCGs having significantly shorter mean LOS and 58 CCGs having longer mean LOS for THR (Figure 4). Variability between CCGs was even more marked for patients undergoing TKR, with 72 CCGs with significantly shorter mean LOS and 62 CCGs with longer mean LOS (Figure 5). Table 6 shows the five CCGs with the lowest estimates of mean LOS (≈3 days) and the five CCGs with the highest mean estimates (≈6–7 days) for THR and TKR patients, respectively. Maps of England with CCG boundaries show that the London region had longer mean LOS, whereas North England and the East Coast have shorter mean LOS for both THR and TKR (Figures 6 and 7). The mean bed-day costs ranged between £4727 (SD £1026) in the NHS Scarborough and Ryedale CCG (within Yorkshire and the Humber region) and £8800 (SD £1572) in the NHS Hillingdon CCG (within London region) for THR. The mean bed-day costs for TKR oscillated between £4758 (SD £1096) in the NHS Scarborough and Ryedale CCG and £8692 (SD £1507) in the NHS Central London CCG.

FIGURE 4.

Caterpillar plots showing variation in outcomes across CCGs (THR). (a) LOS (54/207 CCGs have significantly shorter LOS than average, whereas 54 CCGs have longer LOS); (b) bed-day costs (56/207 CCGs have significantly lower costs than average, whereas 53 have higher costs); (c) OHS change at 6 months (13/207 CCGs have significantly less change at 6 months, whereas 11 have more change); and (d) complications at 6 months (4/207 CCGs have significantly higher percentages of complications at 6 months than average).

FIGURE 5.

Caterpillar plots showing variation in outcomes across CCGs (TKR). (a) LOS (72/207 CCGs have significantly shorter LOS than average, whereas 62 have longer LOS); (b) bed-day costs (65/207 CCGs have significantly lower costs than average, whereas 56 have higher costs); (c) OKS change at 6 months (25/207 CCGs have significantly less change at 6 months, whereas 26 have more change); and (d) complications at 6 months (7/207 CCGs have significantly lower percentages of complications at 6 months than average, whereas eight have higher percentages).

TABLE 6 - Descriptive statistics of variables included in the hip and knee models (data from 2014 to 2016)

Reproduced from Garriga et al. 37 © 2019 Garriga C et al. This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

THRs				TKRs/UKRs
Five lowest		Five highest		Five lowest		Five highest
CCG	Predicted (SD)	CCG	Predicted (SD)	CCG	Predicted (SD)	CCG	Predicted (SD)
Mean LOS (days)
Scarborough and Ryedale	2.7 (1.0)	Hillingdon	6.1 (1.8)	Scarborough and Ryedale	2.9 (1.0)	West London	6.6 (1.7)
Hastings and Rother	2.9 (1.0)	Harrow	6.1 (2.1)	Northumberland	3.2 (0.7)	Hammersmith and Fulham	6.6 (1.8)
Hambleton, Richmondshire and Whitby	3.1 (0.8)	Camden	6.1 (1.9)	Bassetlaw	3.3 (1.4)	Camden	6.5 (1.8)
Northumberland	3.1 (0.9)	Barnet	5.8 (2.0)	Hastings and Rother	3.4 (1.0)	Central London (Westminster)	6.1 (1.6)
High Weald Lewes Havens	3.3 (1.4)	Central London (Westminster)	5.7 (2.0)	Salford	3.4 (1.0)	Barnet	6.1 (1.6)
Mean bed-day cost (£)
Scarborough and Ryedale	4727 (1026)	Hillingdon	8800 (1572)	Scarborough and Ryedale	4758 (1096)	Central London (Westminster)	8692 (1507)
Hastings and Rother	4880 (1051)	Harrow	8647 (1881)	Northumberland	4930 (720)	City and Hackney	8686 (1247)
Northumberland	4998 (867)	Camden	8250 (1561)	Bassetlaw	4960 (1329)	Lewisham	8676 (1585)
Hambleton, Richmondshire and Whitby	5187 (908)	Barnet	8132 (1788)	Hastings and Rother	5294 (1050)	West London	8639 (1438)
High Weald Lewes Havens	5219 (1305)	Waltham Forest	8060 (1710)	Salford	5344 (1008)	Camden	8600 (1493)
Mean OHS change (points)				Mean OKS change (points)
Mid Essex	24.6 (5.3)	Brent	18.7 (6.2)	Ipswich and East Suffolk	18.8 (4.2)	City and Hackney	13.1 (4.3)
Ipswich and East Suffolk	24.1 (4.9)	Islington	19.1 (6.4)	Eastern Cheshire	18.5 (4.2)	Newham	13.4 (4.3)
Stoke on Trent	24.1 (5.0)	Harrow	19.2 (6.6)	Castle Point and Rochford	18.5 (4.0)	Islington	13.6 (5.4)
Scarborough and Ryedale	23.7 (5.4)	Kingston	19.4 (6.6)	Warrington	18.5 (4.2)	Brighton and Hove	13.7 (4.5)
Nene	23.6 (5.3)	Croydon	19.4 (5.9)	Nene	18.5 (4.2)	Barnet	13.8 (4.7)
Mean complications at 6 months (%)
High Weald Lewes Havens	3.0 (2.4)	Tower Hamlets	5.4 (4.1)	Hull	2.9 (1.4)	Herts Valleys	5.8 (2.7)
Leeds West	3.1 (1.9)	Newham	5.3 (3.2)	Dartford, Gravesham and Swanley	2.9 (1.4)	South Devon and Torbay	5.7 (3.1)
Solihull	3.2 (2.0)	Camden	5.2 (3.4)	Calderdale	2.9 (1.5)	Tower Hamlets	5.5 (2.4)
Herefordshire	3.2 (2.0)	Enfield	5.1 (3.0)	Hambleton, Richmondshire and Whitby	3.1 (1.4)	Camden	5.4 (2.8)
West Leicestershire	3.2 (2.1)	Hillingdon	5.0 (3.3)	Erewash	3.1 (1.6)	Newham	5.4 (2.7)

FIGURE 6.

Map of patient outcomes across 2017 CCGs (THR). (a) LOS (days); (b) bed-day costs (£); (c) complications at 6 months (%); and (d) OHS change at 6 months (points). Ultrageneralised clipped boundaries in England V4. Reproduced with permission. 43 Source: Office for National Statistics licensed under the Open Government Licence v.3.0. Contains OS data © Crown copyright and database right 2019.

Reproduced with permission. 43 Source: Office for National Statistics licensed under the Open Government Licence v.3.0. Contains OS data © Crown copyright and database right 2019.

FIGURE 7.

Map of patient outcomes across 2017 CCGs (TKR). (a) LOS (days); (b) bed-day costs (£); (c) complications at 6 months (%); and (d) OHS change at 6 months (points). Ultrageneralised clipped boundaries in England V4. Reproduced with permission. 43 Source: Office for National Statistics licensed under the Open Government Licence v.3.0. Contains OS data © Crown copyright and database right 2019.

Reproduced with permission. 43 Source: Office for National Statistics licensed under the Open Government Licence v.3.0. Contains OS data © Crown copyright and database right 2019.

Main findings

There is substantial variation in patient outcomes of THR and TKR across CCG areas. This variation remained after adjusting for patient case mix and surgical factors. Hospital organisational factors had little to no influence on explaining this variation.

Strengths/limitations

Strengths of the study include use of the NJR data set, which, to the best of our knowledge, is the largest arthroplasty data set in the world, allowing us to generalise the results to the UK. The NJR has near complete coverage of all operations, particularly in the most recent years of data. Linkage to HES allowed us to examine a wide range of confounding factors and enabled us to link in hospital organisational factors; however, the limitation of this is that analysis was restricted to England only and private operations are not included in the HES data set. The large sample size has allowed us to explore geographical variation in rare complications. The main limitation of the study is missing data, which was particularly prevalent for the hospital organisation factors. To overcome this, we used multiple imputation methods, but only single imputation was possible given the complexity of the multilevel regression models fitted.

What is already known

A large number of studies within the literature have identified factors predictive of patient outcomes of THR and TKR. In respect of patient case-mix variables, it has been shown that baseline levels of pain and functional disease severity,44–51 age,46,49,52,53 sex,46,48,54 obesity,48,52,55 comorbidities46,48,51 and social deprivation50 are all related to patient-reported outcomes of pain and function. Less is known about predictors of complications of surgery, although we have previously shown that such complications are rare, with obesity associated with small but clinically insignificant effects. 41 Predictors of LOS are less commonly studied, with literature mostly relating to enhanced recovery interventions, although our work in Chapter 5 shows that age and comorbidity are associated with longer LOS. Much of this work on predictors of outcomes of THR and TKR is formally synthesised within large systematic reviews56 and in the published National Institute for Health Research (NIHR) Programme Grants for Applied Research report [Clinical Outcomes in Arthroplasty study (COASt)57].

We have previously demonstrated evidence of geographical variation and inequity in access to THR and TKR for patients operated on in 2002 (between 12 and 14 years before the patients in our study were operated on). 10 However, in those patients who navigate through the care pathway and obtain access to joint replacement surgery, there has been little research exploring geographical variations in outcomes of surgery. A previous NIHR Health Services and Delivery Research study by Street et al. 13 used HES data to explore variation in PROMs for THR and TKR across hospitals in England. Using multilevel regression modelling, they looked at whether or not patient factors (age, sex, comorbidity, deprivation) and hospital factors (volume, teaching hospital) predicted (1) health outcomes (EQ-5D, OHS, OKS) or (2) resource use (LOS, hospital costs). The key findings were significant unexplained variation among hospitals in both health outcomes and resource use. This is consistent with the findings of our study; however, our research takes this forward by looking at variation in other relevant outcomes (complications, LOS) and looking at a broader range of patient, surgical and hospital organisational factors as predictors of geographical variation in outcomes. Our findings suggest that such factors do not explain this variation. Hence, there are probably other unmeasured historical organisational factors and processes specific to individual local hospitals that may explain why such variation exists. Specifically, we have shown that it cannot be explained by ‘our population is different’, as we have well accounted for patient case-mix factors.

Data sources

The NJR, HES and PROMs linked data sets (see Chapter 2).

Outcomes

Length of stay, cost per bed-day, absolute change in OHS and OKS, and complications at 6 months after surgery (see Chapter 4 for details). The rate of revision up to 5 years after primary THR and TKR was also evaluated. This included revisions declared to the NJR by the surgeons and revisions reported to HES. We specified our analysis time in years, reporting the rate as number of revisions per 1000 implant-years. For a list of the OPCS-4 codes used to identify hip revision, see Report Supplementary Material 4; for knee revision, see Report Supplementary Material 5.

Intervention

The primary exposure (‘intervention’) is the period when the new ERP was introduced (between April 2009 and March 2011).

There is likely to be variation in the dates when individual hospitals introduced ERAS and variation in the type of ERAS service a hospital has adopted. Hence, in addition to looking at the national picture, we have looked at individual trusts in a region of England. In this region, adoption of enhanced recovery in orthopaedics was monitored by the Musculoskeletal Clinical Leaders Network when the pathway was being introduced to hospitals in 2009 and 2010. For this reason, information is available on the dates when individual hospitals in the region introduced their programmes. As outcomes of revision surgery and complications are rare, we focus just on outcomes of LOS and PROMs for the analysis of individual hospitals.

Potential modifiers

We evaluated whether or not trends in LOS and OHS and OKS over time differed by age (18–59, 60–69, 70–79, 80–84, ≥ 85 years) and evaluated the presence of comorbidities (yes/no). For a list of the ICD-1035 codes used to identify comorbidities, see Report Supplementary Material 6.

Statistical analysis

We used a natural experimental study design. 58 We evaluated the ERAS impact on trends before, during and after the implementation of the intervention. 59,60 The timing of implementation of ERP varied by trust and was assumed to span the 2 years of the implementation period (April 2009–March 2011). To do so, we described the trends by calculating monthly outcomes, being means (LOS, bed costs, OHS, OKS), proportions (complications) and rates (revision), together with their 95% CIs. We estimated a fractional polynomial over the study period and plotted the resulting curve along with the CI.

In interrupted times series studies, sample size calculations are related to the estimation of the number of observations or time points at which data will be collected. 61 According to the quality criteria of Ramsey et al. ,62 at least 10 pre and 10 post data points would be needed to reach at least 80% power to detect a change (if the autocorrelation is > 0.4). Our outcomes will be estimated at monthly intervals and, as autocorrelation is unknown, we will allow at least 2 years either side of the date of interest (24 pre and 24 post data points).

We used an interrupted time series approach to estimate changes in outcomes during and immediately following the intervention period, while controlling for baseline levels and trends. We modelled aggregated data points of each outcome of interest by month using segmented linear regression:59

The equation used on trusts in the South Central region excluded the intervention period because this was assumed to be a date point:

Here, Y_t is the mean LOS, mean OHS/OKS change, mean proportion of 6-month complications and mean 5-year revision rate (taking place in month t). ‘Time’ is a continuous variable representing number of months from the start of observation period (i.e. April 2008) at time t. β₀ estimates the baseline level of the outcome at the beginning of the time series (i.e. April 2008). β₁ estimates the trend before ERAS implementation. β₂ is the change in level immediately following the intervention. β₃ estimates the change in the trend in the monthly mean after ERAS started (i.e. ERAS implementation trend). β₄ is the change in level immediately following the end of the intervention. β₅ estimates the change in the trend in the mean monthly number or rate (depending of outcome) after ERAS ended (i.e. ERAS post-implementation trend). We checked the autocorrelation with the previous month, 2 months, etc., until the previous 12 months, using Durbin’s alternative test. 63 Autocorrelation would invalidate the interpretation of the model. For this reason, we estimated the linear regression models with Newey–West standard errors. 64 Parsimonious models were generated using the variables selected through backward regression.

Separate models are fitted for THR and TKR. All analyses were conducted using Stata.

National results

Length of stay

For THR, the mean LOS decreased from 5.6 days in April 2008 to 3.6 days in December 2016 (Figure 8), For TKR, the mean LOS decreased from 5.7 days in April 2008 to 3.6 days in December 2016 (Figure 9). Prior to ERAS, LOS was already decreasing significantly (Table 7). The rate of reduction in mean LOS was higher during the implementation period (April 2009–March 2011), and slowed afterwards (April 2011–December 2016).

FIGURE 8.

Trends in outcomes following primary hip replacement in England, 2008–16, by month. (a) LOS (days); (b) OHS 6 months – OHS baseline; (c) primary hip replacements with a complication in the following 6 months; and (d) primary hip replacements with a revision in the following 5 years. Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

FIGURE 9.

Trends in outcomes following primary TKR/UKR in England, 2008–16, by month. (a) LOS (days); (b) OKS 6 months – OKS baseline; (c) primary knee surgeries with a revision in the following 5 years; and (d) primary TKR/UKR with a complication in the following 6 months. Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

TABLE 7 - Temporal trends in patients undergoing planned primary hip replacement from April 2008 to December 2016

–, p ≥ 0.05; ERAS₀, start point of ERAS intervention in April 2009; ERAS_end, end point of ERAS intervention in March 2011.

Note

Parsimonious models with Newey–West standard errors.

Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

Parameter	Coefficient	95% CI	p-value
LOS
Intercept	5.674	5.655 to 5.693	< 0.001
Monthly trend	–0.020	–0.023 to –0.017	< 0.001
Level change ERAS0	0.176	0.120 to 0.232	< 0.001
Trend change after ERAS0	–0.013	–0.017 to –0.009	< 0.001
Level change ERASend	–0.102	–0.203 to –0.001	0.049
Trend change after ERASend	0.019	0.015 to 0.022	< 0.001
OHS 6 months – OHS baseline
Intercept	17.063	16.896 to 17.230	< 0.001
Monthly trend	0.158	0.130 to 0.186	< 0.001
Level change ERAS0	0.772	0.538 to 1.006	< 0.001
Trend change after ERAS0	–0.131	–0.161 to –0.101	< 0.001
Level change ERASend	0.564	0.208 to 0.920	0.002
Trend change after ERASend	–0.013	–0.025 to –0.001	0.039
Complication by 6 months
Intercept	4.044	3.465 to 4.624	< 0.001
Monthly trend	–0.078	–0.096 to –0.061	< 0.001
Level change ERAS0	–	–	–
Trend change after ERAS0	–	–	–
Level change ERASend	–	–	–
Trend change after ERASend	0.078	0.056 to 0.100	< 0.001
Revision rates by 5 years
Intercept	5.940	5.820 to 6.060	< 0.001
Monthly trend	–	–	–
Level change ERAS0	–	–	–
Trend change after ERAS0	–0.098	–0.105 to –0.090	< 0.001
Level change ERASend	–	–	–
Trend change after ERASend	0.103	0.068 to 0.139	< 0.001

Although older patients had a longer LOS, the secular trends in decreasing LOS were observed across all age groups (e.g. for THR, 4.7 days to 3.0 days in those aged 18–59 years; and 8.1 days to 5.3 days in those aged ≥ 85) (Figures 10 and 11). Secular trends also decreased in patients with and without pre-existing comorbidity (Figures 12 and 13).

FIGURE 10.

Trends of LOS at hospital following primary hip replacement according to age categories in England, 2008–16, by month. (a) All ages; (b) 18–59 years; (c) 60–69 years; (d) 70–79 years; (e) 80–84 years; and (f) ≥ 85 years. ERP implemented in England from April 2009 to March 2011. Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

FIGURE 11.

Trends of LOS at hospital after primary TKR/UKR according to age categories in England, 2008–16. (a) All ages; (b) 18–59 years; (c) 60–69 years; (d) 70–79 years; (e) 80–84 years; and (f) ≥ 85 years. ERAS programme implemented in England from April 2009 to March 2011. Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

FIGURE 12.

Trends of LOS at hospital following primary hip replacement by patients with/without comorbidities in England, 2008–16, by month. (a) Without comorbidities; and (b) one or more comorbidities. ERP implemented in England from April 2009 to March 2011. Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

FIGURE 13.

Trends of LOS at hospital after primary TKR/UKR by patients with/without comorbidities in England, 2008–16. (a) Without comorbidities; and (b) one or more comorbidities. Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

Oxford Hip Score and Oxford Knee Score change

Over the study period, there was an improvement in PROMs, with an increment in OHS 6 months after surgery of 17.7 points in April 2008 to 22.9 points in December 2016 for THR (see Figure 8). For TKR, there was an increment in OHS 6 months after surgery of 15.0 points in April 2008 to 17.1 points in December 2016 (see Figure 9). This secular trend was seen in patients with and without comorbidities, and in all age groups except in those aged ≥ 85 years, for whom the change in OHS and OKS was stable over the time period (Figures 14–17).

FIGURE 14.

Trends of OHS change following primary hip replacement according to age categories in England, 2008–16, by month. (a) All ages; (b) 18–59 years; (c) 60–69 years; (d) 70–79 years; (e) 80–84 years; and (f) ≥ 85 years. Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

FIGURE 15.

Trends of OKS change (OKS 6 months – OKS baseline) after primary TKR/UKR according to age categories in England, 2008–16. (a) All ages; (b) 18–59 years; (c) 60–69 years; (d) 70–79 years; (e) 80–84 years; and (f) ≥ 85 years. Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

FIGURE 16.

Trends of OHS change following primary hip replacement by patients with/without comorbidities in England, 2008–16, by month. (a) Without comorbidities; and (b) one or more comorbidities. Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

Reproduced from Garriga et al. 65 © Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY. Published by BMJ. This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/. Minor changes have been made to the formatting.

FIGURE 17.

Trends of OKS change (OKS 6 months – OKS baseline) after primary TKR/UKR by patients with/without comorbidities in England, 2008–16. (a) Without comorbidities; and (b) one or more comorbidities. Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

The interrupted time series model shows that, prior to ERAS, OHS change significantly increased, by 0.158%, every month (see Table 7). During the ERAS implementation (April 2009–March 2011), the secular trend slowed down and continued to increase by just 0.027%. The subsequent period after the ERAS implementation (April 2011–December 2016) remained stable. For the OKS model, there is no impact of the intervention in the secular trend of OKS change (Table 8), although post-intervention trends improved in the 60–69 years and the 70–79 years age groups, by 0.025% and 0.024%, respectively.

TABLE 8 - Temporal trends in patients undergoing planned primary knee replacement from April 2008 to December 2016

–, p ≥ 0.20; ERAS₀, start point of ERP intervention in April 2009; ERAS_end, end point of ERAS intervention in March 2011.

Note

Parsimonious models with Newey–West standard errors.

Reproduced from Garriga et al. 66 © The Author(s). This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC-ND 4.0) license, which permits others to share this work, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

Parameter	Coefficient	95% CI	p-value
LOS
Intercept	5.713	5.688 to 5.738	< 0.001
Monthly trend	–0.029	–0.031 to –0.026	< 0.001
Level change ERAS0	0.151	0.100 to 0.202	< 0.001
Trend change after ERAS0	–	–	–
Level change ERASend	–0.088	–0.160 to –0.016	0.017
Trend change after ERASend	0.015	0.012 to 0.017	< 0.001
OKS 6 months – OKS baseline
Intercept	14.004	13.842 to 14.166	< 0.001
Monthly trend	0.029	0.024 to 0.034	< 0.001
Level change ERAS0	0.301	0.029 to 0.573	0.030
Trend change after ERAS0	–	–	–
Level change ERASend	–	–	–
Trend change after ERASend	–	–	–
Complication by 6 months
Intercept	3.928	3.727 to 4.128	< 0.001
Monthly trend	–0.056	–0.084 to –0.029	< 0.001
Level change ERAS0	–0.766	–1.345 to –0.187	0.010
Trend change after ERAS0	–	–	–
Level change ERASend	0.293	0.003 to 0.584	0.048
Trend change after ERASend	0.054	0.026 to 0.082	< 0.001
Revision rates by 5 years
Intercept	5.986	5.908 to 6.063	< 0.001
Monthly trend	–	–	–
Level change ERAS0	–	–	–
Trend change after ERAS0	–0.038	–0.045 to –0.030	< 0.001
Level change ERASend	–	–	–
Trend change after ERASend	0.050	0.022 to 0.079	0.001

A total of 6232 (1.6%) THR patients between April 2008 and March 2016 had one or more complication in the following 6 months. The proportion of complications at 6 months decreased from 4.1% to 1.7% (see Figure 8). The interrupted time series model shows that prior to ERAS, complication proportions decreased by 0.078% every month (see Table 7). During the ERAS, the secular trend increased by just 0.078%. The period after the ERAS intervention was stable.

For TKR patients, the proportion of complications at 6 months decreased between the first and the last month of the study period, from 4.1% in April 2008 to 1.7% in May 2016. The trend significantly declined before intervention (April 2008–March 2009), by 0.056% (see Table 8). The trend of 6-month complications slightly increased after the ERAS implementation (April 2011–December 2016), by 0.054%.

Observation sessions and job shadowing

Potential participants were identified by a staff member working in the orthopaedic department, or equivalent, of the individual hospital sites. Potential participants were then sent a study information pack that included information about the study, an invitation letter and a reply slip to return if they were interested in participating. Snowball sampling was also used, in which participants recommended other potential participants. 93

Using an observation checklist, observation sessions were conducted at each study site. Information was recorded in writing, as is standard in ethnographic research. The checklist was developed by the study team and provided structure for the researcher to collect information in a systematic manner so that field notes described the clinical setting, activities taking place, treatment protocols and factors that may have an impact on implementation. Informal interviews were also used. Initial notes were written up into full ethnographic ‘field notes’ during each day of data collection, or as soon as possible thereafter. To inform further data collection, memos or reflective notes were used to record emerging thematic ideas. A total of 19 staff agreed to be shadowed and approximately 160 hours of fieldwork was conducted (5 days of approximately 8 hours at each study site).

Semistructured interviews

Face-to-face semistructured interviews were undertaken with health-care professionals involved in service delivery. Interviews ranged from around 30 to 60 minutes in duration. Thirty-one health-care professionals participated in interviews, of whom 12 had also participated in observations. A ‘topic guide’ or list of themes to explore in the interviews was devised based on data collected during observation sessions. Interviews focused on participants’ views and experiences of delivering ERAS and factors having an impact on implementation. To ensure that the experiences of participants were not ‘forced’ into predefined concepts, data were not structured around implementation science theory. The topic guide was flexible to enable follow-up on issues raised. Interviews were audio-recorded, transcribed and anonymised.

Analysis

Analysis was iterative and ongoing and informed further data collection. Analysis was carried out in two phases, an interim and a final phase. Written field notes and transcripts of interviews were anonymised and imported into NVivo software for analysis (QSR International, Warrington, UK). Interview transcripts and field notes were analysed using an inductive thematic approach85 to identify themes and subthemes in the data. On account of the variation in service delivery between sites, data from each hospital site were analysed as a discrete data set. Twenty per cent of transcripts were double coded by another member of the research team (RG-H). Codes were then discussed and refined to reach a single code list. As part of the interim analyses, the CFIR was used to structure further analysis. This theory was identified as, unlike most other approaches, it suggests that successful implementation is dependent on how well the intervention meets patient needs, which accords with our findings. Using the CFIR as part of the analysis involved transposing themes that had been coded inductively onto the 31 constructs of the framework, grouped into the five domains: (1) ‘intervention characteristics’; (2) ‘outer setting’; (3) ‘inner setting’; (4) ‘characteristics of individuals’; and (5) ‘process’. This was an ‘abductive’ approach to analysis as described by Tavory and Timmermans. 94 Interpretive accounts of the data were then generated.

Recruitment

Potential participants were patients who had undergone THR or TKR at one of the four study hospitals. Study information packs were distributed by local hospital health-care professionals. Packs included an invitation letter, a booklet with study information and a reply slip. Reply slips were returned to a contact at each hospital. Potential participants were then either approached by the lead researcher while in hospital or contacted by telephone after discharge.

Semistructured interviews

Interviews ranged from around 30 to 60 minutes in duration. Thirty-seven patients participated in interviews, with the final sample size determined by data saturation. Interviews with patients 1 or 2 days postoperatively were conducted face to face in an inpatient setting. Those with patients who had been discharged took place by telephone owing to logistical challenges. A topic guide divided into themes was used to explore patients’ views and experiences of having a joint replacement throughout the care pathway.

Data analysis

Interviews were audio-recorded, transcribed, anonymised and imported into NVivo qualitative analysis software. An inductive thematic analysis was conducted, such that themes and subthemes were identified in the responses. 85 Memos were used to record emerging ideas and themes throughout the process of data collection and analysis. Ten per cent of interviews were independently double coded by a member of the study team and themes compared and contrasted to reach a single code list. Descriptive accounts of the data were then generated.

Ethics approval

Ethics approval was provided by the South-West Exeter Research Ethics Committee (reference 16/SW/0214). Participants provided their written consent before interview to confirm that they understood the aims and objectives of the research, that their participation was voluntary and that they were willing to let the study include anonymous quotations from them in reports of the study. Each NHS trust involved provided research and development approval.

Results

Characteristics of ERAS services for hip and knee replacement are displayed in Table 10. These present summarised information only to preserve the anonymity of sites.

Sample characteristics

The 38 participants comprised 10 physiotherapists or occupational therapists (OTs), 18 nurses, five orthopaedic surgeons, one anaesthetist, one matron, two therapy technician assistants and one theatre manager. Twelve staff participated in interviews and observations, 19 took part in interviews only and seven took part in observations only. Between 4 and 14 participants took part from each study site. Participants’ characteristics are displayed in Table 11, which presents summarised information by study site to avoid the potential for identification of individual participants. We use pseudonyms for study sites and, in this part of the report, when we use the term ‘participant’ we are referring to hospital staff.

Seventeen CFIR constructs were found to influence processes of implementation for ERAS programmes in all five domains of the framework. A summary of the themes identified in the data and their relationship to these constructs and domains is given in Table 12.

We now explore factors that have an impact on the implementation of ERAS services. These are structured with the five domains of CFIR.

Intervention characteristics

Intervention characteristics illustrative quotations are displayed in Box 1.

Staff expressed enthusiasm for the relative advantages of ERAS, as shortened LOS had resource and cost-saving implications. There was a sense of a ‘trade-off’ between reducing LOS and not increasing readmission and complication rates. At Shinebury, care pathways were developed internally by consultant surgeons who piloted ERAS and communicated findings to staff. This helped generate internal support. By contrast, a nurse sister at Elmfield described how ERAS practices had been ‘introduced on us’ and suggested that having someone to lead its development would have inspired enthusiasm.

Participants saw ERAS as adaptable, and this was key to it working in individual hospital contexts. None of those interviewed had seen any formal policies issued by the Department of Health and Social Care, although many were aware that ERAS was a government initiative. There were various views about which patients should be included in the pathway. At Shinebury, all patients were included in ERAS, whereas at Elmfield patients were selected only if it was felt that they were healthy enough for rapid discharge.

Outer setting

Outer setting illustrative quotations are displayed in Box 2.

As ERAS was viewed as a ‘partnership’ between members of staff and patients, participants thought that meeting patients’ needs would enable ERAS to work effectively. A number of patients could not be discharged because they were medically unfit, but some patients were also thought by staff to be somewhat resistant to ‘rapid’ discharge after their operation. A nurse at Elmfield thought that, to address these issues, it was crucial to adapt approaches to suit individual needs, adopting a recovery time that was manageable. Another emphasised the importance of education as a way of ‘breaking down’ attitudes that acted as a barrier to discharge.

Patient information was distributed in a range of formats. Written information helped to reinforce information provided at consultation and gave patients a source to refer back to. Sites operated hip and knee schools, regular classes designed to educate patients about their treatment, and participants thought that the ‘group dynamic’ created a safe space for asking questions and sharing experiences. Face-to-face contact was seen as an opportunity to clarify information. Participants thought that patient education had multiple aims, the most important of which was helping patients to have ownership of their recovery. Professionals thought that such education made it easier for them to provide postoperative support because patients had more knowledge about what to expect during recovery.

Participants conceptualised ERAS as a ‘message’, which had to be consistently communicated so that patients understood and adhered to recommendations. If the message and its consistency was ‘diluted’, then understanding and adherence could be reduced. Participants thought that new staff who were not familiar with ERAS and those who were ‘not buying into the process’ might provide inaccurate information.

Participants from all sites were concerned about post-discharge support. According to a nurse at Elmfield, this was important as patients ‘panicked’ without it. There was variation in post-discharge services provided by the sites. One participant felt that providing a telephone number as a point of contact was ‘the absolute minimum’. Staff thought that this made patients feel more ‘secure’.

Co-operation with external agencies was important, notably because the smoothness of connection with primary care at the point of referral and discharge had an impact on the successful implementation of ERAS. Staff suggested that GPs did not always understand the practicalities of ERAS, and referrals might be inappropriate or patients might receive information about recovery that may not accord with the information provided from within the hospital.

Participants highlighted some gaps in communication between primary and secondary care after patients’ discharge. One felt that GPs sometimes questioned patients’ readiness to return home; another felt that they got ‘cross’ when they thought that patients had been discharged by the hospital without sufficient pain relief. Participants at Shinebury and Elmfield were worried that patients had no point of contact in secondary care if they were experiencing difficulties before their follow-up review. This meant that such patients had to return to primary care first, which was seen as overly complex and placing an unnecessary burden on services. Furthermore, there were concerns that GPs were not provided with adequate post-discharge documentation. According to a nurse at Elmfield, there was a need to engage more closely with GPs and community services.

One of the key elements of success was effective networking and communication between staff. Multidisciplinary team members tended to operate in ‘silos’ with responsibility for delivering different components of care. To communicate patient information, knowledge had to be transferred along the care pathway as part of a ‘logical progression’. However, a nurse at Towerton was concerned that those undertaking pre-assessment were not consistently transferring information, meaning that potentially important details were missed.

Regular multidisciplinary team meetings to discuss ERAS were advocated, although these were challenging to organise. Informal communication was seen as being important and the location of staff in close proximity was seen to facilitate this. Nurses, physiotherapists and OTs at Shinebury and Lastmere ran ‘joint clinics’ together, and doing so encouraged collaboration. Multidisciplinary documentation was also valued, although the quality and consistency of this varied. For instance, at Elmfield, paperwork did not identify whether or not patients had been assigned to the ERAS pathway. A physiotherapist at Lastmere viewed documentation as a ‘back-up’, as staff were in ‘constant communication’ with one another.

A number of participants characterised ERAS as a ‘message’ that needed to be communicated across the multidisciplinary team to ensure that its components were being consistently delivered. However, this was not always achieved. For instance, at Elmfield, surgeons did not always agree with one another about which patients were eligible for ERAS. ERAS champions helped to ensure that the ‘message’ was successfully communicated and that staff were delivering components of care consistently.

Regarding the implementation climate, participants described the importance of a collective ethos and ‘belief’ in ERAS. ERAS champions helped to garner support from the multidisciplinary team. However, at Elmfield, there seemed to be no clear leadership. Furthermore, not all team members were invited to meetings to discuss ERAS development, and this made them feel less engaged.

There was variation in how compatible ERAS was seen to be with existing roles. A number of participants thought that ERAS involved expanding on existing working practices, making it easy for them to do the necessary work. However, some anaesthetists were reportedly resistant as they preferred using their own professional judgement over following protocols. Similarly, a participant at Towerton had found it difficult to change nursing practice, as colleagues were uncomfortable about encouraging patients to be so independent and were reluctant to send them home.

Targets and goals for LOS were established formally by the hospital or trusts and informally by ERAS services. Performance against formal and informal targets was fed back to staff and used as a basis to collaboratively improve service delivery. Failure to meet formal LOS targets led to fines, meaning that staff felt under substantial pressure to meet these goals.

The financial cost of maintaining ERAS was a concern, although the extent and nature of this varied. Elmfield staff were particularly worried about lack of current funding, which meant that they were not able to acquire sufficient staff or facilities, such as beds. Staff at Lastmere thought that funding cuts may prevent them from providing patient information booklets, which they saw as being central to effective rehabilitation.

A shortage of available staff and high staff turnover were seen as creating difficulties, as they meant that colleagues had to do additional work and struggled to find time to deliver care. At Shinebury, staff explained how time constraints made it challenging to arrange formal care packages after discharge and this had an impact on the quality of postoperative follow-up. A deputy nurse sister at Elmfield thought that follow-up reviews should be undertaken by nurses or physiotherapists, as they were at Shinebury, to relieve the ‘pressure’ on consultants. Patient numbers at Shinebury and Towerton were also seen to place pressures on services. Alongside this, operations could be cancelled at short notice when operating theatre space was needed for trauma surgery.

Staff at the four sites had received varying levels of access to knowledge and information about ERAS. Towerton appeared to have the most comprehensive training and education, and staff spent time with the nurse champion, who ran educational sessions and incorporated information on ERAS into ongoing orthopaedic training. Staff at Shinebury were also expected to attend joint school to help them educate patients more effectively. By contrast, a participant at Elmfield explained how the intervention had been introduced without any formal education and that this had not been effective. Multidisciplinary team meetings were used as a way of communicating information about changes in working practice, along with ‘learning on the job’.

Inner setting, characteristics of individuals and process

Participants emphasised the importance of individual commitment from staff. A strong belief in the relative advantages of ERAS meant that most staff were committed to delivering the service. Resistance to change existed where ERAS practices were seen as being incompatible with professional judgement.

Inner setting, characteristics of individuals and process illustrative quotations are provided in Box 3.

To plan processes of care, protocols were used to ‘streamline’ services and ensure that patients received key elements of care, although these were not always formally described. However, participants stressed that these should be sufficiently flexible to meet individual needs.

A consultant surgeon at Shinebury emphasised the importance of sustaining multidisciplinary commitment and advocated ‘top-down’ encouragement and close monitoring to do so. To facilitate this, staff at Shinebury held multidisciplinary meetings to ensure that key members of the team were cascading information to colleagues ‘to keep that momentum going’. However, a nurse at Elmfield explained that not all team members were invited to meetings to discuss ERAS development, and this made them feel less engaged.

Involving strong opinion leaders in the development of ERAS helped to generate internal support, whereas a lack of this at Elmfield was a barrier to engagement. The importance of having a recognised ERAS champion to ‘drive through’ changes was highlighted. Towerton had a designated nurse specialist who acted as the central point of contact. As a result, other members of the multidisciplinary team did not need to be familiar with all aspects of the protocol. Similarly, consultants at Shinebury were identified as a source of expertise. Clinical champions also helped to engender enthusiasm.

The ERAS programme had to be (re)activated on a continuous basis through reflection, evaluation and modification. To reconfigure care, staff at Shinebury used multidisciplinary meetings to review outcomes data and ‘brainstorm’ ways of improving services. Informal communication between team members, for instance in hip and knee schools, provided another opportunity for this. Patient feedback was used to shape patient education materials and joint schools. Feedback was collected informally or through patient satisfaction questionnaires. On account of these processes, ERAS was seen as having been improved or ‘refined’ at three study sites. ERAS at Shinebury was described as having a ‘core element’, which has grown outwards as the service has ‘tried to add bits on to try and improve the situation’. By contrast, staff at Elmfield talked about how ERAS was gradually being ‘nibbled at the edges’.

Sample characteristics

The 37 participants were 15 men and 22 women. Between 6 and 16 participants took part from each study site; 14 had undergone TKR and 23 had undergone THR. Twenty-two participants were interviewed in an inpatient setting and 15 were interviewed once they had been discharged. Of those who had been discharged, five were interviewed 2–4 weeks postoperatively and 10 were interviewed 5–13 weeks postoperatively. Participants’ characteristics are displayed in Table 13. In this section, we use the term ‘participants’ to refer to patients.

Referral process

Around half of the patients who were interviewed said that they had been given the chance to choose which hospital they would be seen at. Most participants made their choice of hospital based on previous experiences and ease or convenience. A smaller number of participants made their choices based on hospital performance. Others were guided by GPs in their decision-making, particularly those who had additional health-care needs. However, for most, choice of hospital was not a priority compared with the need to progress towards surgery.

Participants held varying views about waiting times for surgery. Some were surprised by the short period between referral, appointment and operation date, and others experienced delays.

Participants attributed delays to funding constraints and administrative difficulties, particularly at the interface between primary and secondary care. Several described how their operations had been cancelled and, although some were sympathetic to the pressures on the health-care system, for others this was emotionally difficult. The difficulty led to feelings of helplessness and frustration.

Preoperative education

Preoperative experiences illustrative quotations are displayed in Box 4.

Patients were referred from primary care (GPs) to secondary care and most participants were satisfied with the verbal information that they had received in secondary care. The verbal information was augmented by written information, and they thought that written information about postoperative exercises was particularly useful, as they kept it until after discharge.

Participants valued the opportunity to attend classes for patients who were about to have hip or knee surgery, called hip or knee ‘schools’. Patients felt that the classes helped prepare them for surgery and gave them ‘confidence’ in the care they would receive. Classes helped make them feel enthusiastic about the potential for early discharge and invested in their recovery, preparing them to take on the role of ‘active’ participants in their own care. Guidance about how to use crutches and perform exercises was considered particularly useful. Most thought that it was useful to practise exercises to prepare them for standing and walking after their surgery, although some said that they had not practised consistently.

Classes were also a source of emotional support, as they were an opportunity to meet patients ‘in the same boat’. Those undergoing their first replacement found it useful to talk with and listen to patients who had undergone previous surgery: they used the classes as a time to ask others questions about their experiences. An understanding of the roles of both patients and health-care professionals meant that expectations of care were aligned prior to surgery.

The timing of these classes was sometimes problematic. Those who had received a rapid referral owing to a cancellation missed out on this component of care. Others felt that they had attended too far in advance of their operation, meaning that they could not remember the information on admission. These challenges potentially undermined patients’ ability to participate fully in their care.

Participants wanted health-care professionals to provide them and their families with more information about their recovery and their expected progress to help them take care of themselves. Some also felt that they would have liked more emotional support. A small number of younger participants wanted more information about joint replacements in younger people, with a focus on any differences in their likely recovery after surgery.

Participants also sought information from family and friends who had undergone surgery, and from the internet. Nevertheless, one participant raised concerns about the availability of conflicting information and thought that hospitals should guide them to reliable sources.

Preoperative preparation

All participants found the preoperative assessment straightforward and comprehensive. They reported that they were able to do everything that was required of them before surgery (see Box 4).

Waiting for operation

A number of participants described that they had to wait several hours from the point of admission to hospital until their operation. They found this hard because they were not permitted to eat or drink while waiting. Patients said that they would have valued more information about the reason for their wait and about the precise time of their operation, and that absence of this information made them feel more anxious.

Anaesthesia

Inpatient experiences illustrative quotations are provided in Box 5.

Most participants were presented with the option of local or general anaesthesia, although they had different experiences of this. Although some were given this option during hip or knee schools or their preoperative assessment, others were asked on the day of surgery. These patients reported that it had made them feel ‘surprised’ and ‘alarmed’ and worried about ‘being awake’ during surgery. They also felt that receipt of information on the day of surgery meant they had not felt able to make an informed choice. Several explained that anaesthetists also expressed a clear preference for localised anaesthesia, which influenced their decision. Contrary to ERAS guidelines, one participant reported that she had not been offered localised anaesthesia, but had insisted on it against the advice of her anaesthetist.

Pain management

Patients were generally satisfied with how their pain was managed and felt that nurses were attentive to their needs (see Box 5). However, some patients at Shinebury felt under pressure to take their pain medicines, although they thought that the medicines provided were ‘excessive’ and made them feel ‘out of it’. Others wanted more information about the pain medicines that they had been prescribed. One patient expressed frustration and thought that nurses were reluctant to share this information, which meant that she was not able to make a fully informed choice about her pain medicines.

Inconsistencies in information

The majority of participants felt that they had been fully informed about what to expect during their inpatient stay and that, as a result, there were ‘no surprises’ (see Box 5). However, some participants identified inconsistencies between preoperative information and subsequent information, which they found unsettling.

Early postoperative mobilisation

Patients were surprised about being able to get up and walk so quickly after surgery (see Box 5). There was a sense of pride in being able to fulfil what they thought was required of them and they benchmarked their success against the content of the information that they had received before surgery, as well as feedback from health-care professionals. Positive affirmation was important to encourage patients to meet ERAS goals. Participants emphasised the importance of participating in exercises and working towards their own recovery, engaging in the emotional and physical work required of them as ‘active’ patients.

Most felt that mobilisation after their operation moved at an appropriate pace. When patients could not mobilise quickly, particularly older patients or those with complications, they felt that staff largely listened to their concerns. Likewise, a younger patient was satisfied that staff helped her to mobilise more quickly than anticipated. In this way, care was negotiated and plans were amended to meet patient needs.

However, other participants described how they had felt under pressure to mobilise and ‘conform’ to the demands of the programme. Such patients thought that health-care professionals did not listen to their concerns or respond to their individual needs. Others experienced a tension between giving their bodies time to ‘heal’ and the need to become active, expressing a concern that activity might be ‘damaging’. When expectations of patients and health-care professionals differed, this created frustrations for patients. A small number of participants thought that such tension undermined their right to autonomy and choice about their care.

Discharge

Most participants reported feeling ‘ready’ to return home by the time that discharge happened (see Box 5). They thought that it was better to go home than to remain in hospital. Feeling ‘prepared’ contributed to a feeling of confidence about discharge and return home. Participants valued the help and guidance that they had received before admission about how to prepare for discharge. Such preparations included reorganising the home, acquiring assistive devices and ensuring that family and friends were available to help. Participants at Towerton thought that assessing patients’ needs before admission and providing appropriate assistive devices were useful. Information empowered patients and instilled confidence. Those being discharged felt more reassured when they understood their recovery trajectories, how to perform post-discharge exercises and what they could and could not do. Part of the ERAS process is mobilisation (i.e. standing and walking) and participants said that their desire to go home motivates them to mobilise. Discharge delays due to staffing issues were a source of frustration and patients sometimes received conflicting information about which day they were being discharged, which made them feel unsettled.

By contrast, other participants reported that they felt under pressure to be discharged before they felt well enough. This was linked to their concerns that they were being made to mobilise too quickly. A small number of participants attributed this to a demand for beds. Among those who were concerned, worries about going home were wide-ranging and included not knowing what to expect, fear of falling and a lack of ability to sleep and perform everyday tasks. Those who lacked family support, particularly those who lived alone, were more anxious.

Postoperative

Post-discharge experiences illustrative quotations are provided in Box 6.

Participants said that their initial recovery at home after discharge from hospital was the most challenging period of their journey through joint replacement. Physical challenges included postoperative bruising, high pain levels and ‘unfamiliar sensations’ that they had not anticipated. Participants facing such challenges explained that they were unsure what was ‘normal’. Although many were relieved to be back in their home environment, participants described unexpected practical challenges, such as bathing and going to the toilet. Participants said that adaptive devices were useful, but also found that ones that they needed were not necessarily available.

Physiotherapy exercises

Participants emphasised the importance of actively participating in their care by carrying out exercises that the physiotherapy team had provided. However, several were unsure if they were performing the exercise in keeping with the guidance they had received. Again, there was a reported tension between being active and allowing their bodies time to heal. Some felt uncertain about how far to ‘push themselves’. Participants also said that they were not always certain about whether or not some daily activities were safe to carry out, for instance whether or not they should be sleeping on their side.

Pain relief

Many participants were satisfied with the pain relief medicines that they had been given on discharge. These included medicines to take home with them. However, if things were not going well then managing pain after discharge could be a source of worry. One participant described feeling ‘alone’ and ‘abandoned’ when experiencing severe pain 1 week after discharge, and not knowing who to contact for support. Most participants described feeling confident in reducing their pain medications without clinical input. However, one thought that her decision to do this too quickly had caused her pain to return.

Post-discharge support

Although patients were expected to be independent in their recovery after discharge, many struggled to maintain their role as active participants without clinical input. Among those experiencing challenges, there was a desire for more support to provide ‘reassurance’. This included guidance from professionals in secondary care and their own GPs. The provision of contact details on discharge helped provide a feeling of security. Knowledge that they were receiving a follow-up appointment was also a source of reassurance and participants saw this as a key point in their recovery.

Family

The importance of family and friends as a source of practical and emotional support during this period was repeatedly emphasised. When this was not possible, a small number of participants had help from community services, which they also valued.

The future

Despite some of the challenges encountered, participants were largely optimistic about moving forwards and they enjoyed reflecting on progress. Many discussed how their joint replacement would enable them to participate more fully in everyday life. Expectations included having ‘pain-free joints’ and ‘going for short walks’, and participants emphasised the importance of having ‘realistic’ expectations. For some, discussions of the future were dominated by the prospect of future joint replacements. Others remained uncertain about what to expect.

Overview of findings

Seventeen of the 31 CFIR constructs influenced the implementation of ERAS across all five domains. As described in Intervention characteristics, participants felt that ERAS afforded advantages over alternative solutions. Support was higher when ERAS was seen to have been developed internally rather than externally. Guidance was flexible and could be adapted to meet the demands of individual hospital services. In the ‘outer setting’, participants thought that ERAS should be tailored to patient needs and that education could empower them in their recovery.

There were concerns about a lack of post-discharge support and tensions between primary and secondary care. In the ‘inner setting’, one of the key elements of success was effective multidisciplinary collaboration. This was achieved by transferring knowledge about patients along the care pathway, through multidisciplinary team meetings and paperwork. ERAS was a ‘message’ that had to be communicated to all staff, but there were concerns about funding constraints, staffing levels and large numbers of patients. Access to information about the intervention was variable. The characteristics of individuals had an impact on implementation, and staff were reluctant to change working practices when ERAS was seen as being incompatible with professional judgements. Formal and informal targets were used to inform service delivery. Within ‘process’, protocols were used to streamline care, although these had to be flexible to meet individual needs. Participants thought that ‘top-down’ encouragement, monitoring and regular meetings helped to ensure team engagement. Involving strong opinion leaders in its development, and ‘champions’ who drove through implementation and acted as a point of contact, helped facilitate implementation. Reviewing outcomes data, informal communication to discuss progress and patient feedback helped to develop ERAS over time.

Findings on patients’ experiences of ERAS showed that time to surgery varied and cancellations were emotionally difficult. Patients valued preoperative education, but timings meant that not all patients were offered educational classes. Participants wanted more information on expected progress, along with emotional support. There were different experiences of being offered a choice of anaesthesia. For some, this was explained in advance of admission, and others were consulted on the day of surgery, which caused anxiety and prevented them from making an informed choice. Some participants identified inconsistencies in information between their preoperative education and inpatient care, which they found troubling. Most were enthusiastic about mobilising quickly after surgery, although others felt under pressure and wanted to leave their bodies time to ‘heal’. Most participants reported feeling ‘ready’ to return home, and this was enhanced by discharge preparation before admission and the support of family. However, a small number felt that they were under pressure to be discharged before they were well enough. Postoperative recovery was often most challenging in the first few weeks, and some were unsure how to perform exercises and daily tasks, or were worried about unfamiliar bodily sensations. They wanted more post-discharge input to provide ‘reassurance’. Follow-up appointments were therefore considered key points in recovery.

How findings relate to current literature

Findings characterise differences in how ERAS services for hip and knee replacement are delivered by identifying barriers to and enablers of their successful implementation. This may help to account for variation in health outcomes for these surgeries. 13 For instance, meeting patient needs may help them to work more successfully towards their own recovery. Findings reflect those from previous studies that have explored processes that influence implementation of ERAS for other conditions. 88,96,97 These found that multidisciplinary collaboration was essential and that this could be threatened by the need to co-ordinate working practices across different departments. 98 Likewise, components of ERAS were seen as incompatible with working practices of some members of the multidisciplinary team; this meant that some staff were resistant to change. 99 A need to engage staff was emphasised, and ERAS ‘champions’ were seen as a means of achieving this goal. 26,96 The importance of providing education and information to patients, and providing realistic expectations of their recovery, was discussed. 28,88 Temporality or strategies to embed ERAS over time were discussed in a small number of studies. 59 Studies have been synthesised in a recent systematic review. 87 Our study contributed to this literature by emphasising the importance of meeting patient needs in service design and for effective collaboration between primary and secondary care services.

Previous work has highlighted the value of preoperative education as a way of helping patients to prepare emotionally and practically for colorectal surgery. 100 In other surgeries, inconsistencies between preoperative education and inpatient experiences in colorectal and colonic surgeries have also been previously identified and our findings accord with these. 25,101 Patients undergoing colorectal surgery expressed enthusiasm about working towards their own recovery,100 a finding reflected in our work. In addition, some colorectal patients echoed the tension we identified between meeting the demands of the programme and allowing themselves time to recover. 102 Several of these also worried that they were being discharged too early. 24,25 A recent review of patient experiences of ERAS for a range of surgeries reflects one of the core findings of our study about the importance of post-discharge support. 24,103 Other issues identified in our study relate to challenges in the wider health-care system, such as the cancellation of elective surgeries owing to funding constraints104 and discharge delays. 27

Our study identified issues that may relate specifically to ERAS as delivered in hip and knee replacement. For instance, some patients felt unable to make an informed choice about anaesthesia, although both local and general anaesthesia may be used. 105 Challenges were identified in the postoperative period, particularly around difficulties in performing physiotherapy exercises that are a crucial component of rehabilitation17 and a desire for more assistive devices.

Patient experiences of ERAS for hip and knee replacement can be understood using Mol’s work, which explores how health-care professionals interact with patients and negotiate processes of care. 86 Processes of negotiation became apparent when patients began engaging in efforts to mobilise post surgery. In the immediate post-discharge period, most patients were enthusiastic about working towards their recovery or reported how decisions were taken in collaboration with health-care professionals who ‘tinkered’ with their care to address individual needs. However, others reported tensions between the expectations of health-care professionals and their own preferences, and expressed a desire for a greater element of control. Issues were echoed in discussions around timing of discharge. In contrast, once they had been discharged, patients struggled with their autonomy and wanted more support to guide their rehabilitation. The study also highlights the physical and emotional work that patients did as ‘active’ participants in their recovery. 86

Strengths and weaknesses

Using ethnographic research methods involved spending extended periods of time at study sites, using multiple research methods that provided a rounded account of practice. Analysis included information about what people did, as well as what they said, and their reasons for their actions and decisions. 106 Different numbers of participants were drawn from each of the study sites, meaning that experiences at some hospital sites may be over-represented in the analysis. This was mitigated by analysing data from each hospital as a discrete data set and then comparing and contrasting findings. We conducted recruitment of patients through busy health-care professionals working within services, and we did not ask them to keep records of those who were approached but who did not take part out of respect for the professionals’ busy workloads. This means that there is the chance that those who participated may not represent the experiences of all patients accessing services. However, inclusion of four hospitals, the relatively large sample size of health-care professionals (n = 38) and patients (n = 37) with diverse characteristics, achievement of data saturation92 and robust analysis provide confidence that findings may be transferable to other settings. 107 In addition, interviewed patients were either in hospital awaiting discharge or already at home.

Discussions around preoperative and inpatient experiences represent the views of all patients who participated in the study (n = 37). We found that experiences in both groups were broadly similar and therefore we do not make a distinction between them in the analysis. Postoperative accounts relate only to those who had been discharged from hospital at the time of interview (n = 15).

The CFIR provided a theoretical basis to our analysis. We used CFIR because of its emphasis on meeting patients’ needs in service design. Our study highlighted that meeting patients’ needs was central to the successful implementation of ERAS into everyday practice. By using inductive coding and transposing themes onto the theory that we thought was the best ‘fit’ for the data, we ensured that data were not forced into predefined constructs. A challenge that we encountered in analysis was how best to make decisions about where themes fitted best, particularly when it was possible that these could be mapped against more than one construct. When this was the case, themes were mapped onto the construct that was considered to be the best ‘fit’ or coded into more than one construct.

Future research

Study participants reflected on the role of primary care in delivering components of ERAS, including processes of referral and post-discharge support. Further research could explore how primary care interacts with ERAS protocols, providing a more comprehensive understanding of the delivery of ERAS. In addition, further research should explore patient experiences of referral from primary to secondary care services, to provide a more comprehensive account of experiences through the care pathway.

Whole enhanced recovery pathway

Larsen et al. 109 report an economic analysis alongside a randomised trial, with 56 hip replacement and 31 knee replacement participants in Denmark. The pathway used in the treatment arm (‘accelerated care’) was not different to that of the control group (‘conventional rehabilitation’) in terms of intraoperative management, analgesia, nausea control or bowel regulation. Differences in the treatment protocols between the two arms involved patient education, nutrition, admission times, staffing and mobilisation (described in Report Supplementary Material 7). The accelerated care pathway was less costly and more effective than the control, both overall and in the subgroup of hip replacement participants. For knee replacement patients, the authors reported the accelerated care pathway to be cost-saving but less effective than the control, albeit the difference was not statistically significant. They reported a cost saving of 618,075 Danish krone per QALY lost with accelerated care compared with conventional rehabilitation, which made it cost-effective (threshold of 160,000 krone per QALY in Denmark).

Brunenberg et al. 98 report an economic evaluation of a before-and-after trial with 98 hip and 62 knee replacement participants in the Netherlands. The intervention (‘joint recovery programme’) consisted of a 20-minute pre-assessment screening 6 weeks before the operation, for physical assessment and analysis of the home situation to aid discharge planning, patient education sessions 1–2 weeks before surgery, group rehabilitation sessions and supervision by physical therapists and nurses (described in Report Supplementary Material 7). Patients in the ‘usual care’ group underwent conventional physiotherapy for 1 hour daily, did not receive pre-assessment screening or information sessions, and discharge arrangements were addressed during admission to hospital. The joint recovery programme intervention was less costly and more effective than the control for both hip and knee replacement, resulting in a cost saving of US$1261 per patient for hip replacement and US$3336 per patient for knee replacement, albeit with no statistically significant difference in QALYs. The probability that the joint recovery programme was the most cost-effective option was > 80% for willingness-to-pay thresholds up to US$45,000.

Preoperative components of the enhanced recovery pathway

McLawhorn et al. 110 used a Markov model to assess the cost-effectiveness of bariatric surgery 2 years before total knee arthroplasty for morbidly obese (BMI ≥ 35 kg/m²) patients who were candidates for both operations because of end-stage knee osteoarthritis and failed non-operative weight loss interventions. The strategy including bariatric surgery was cost-effective at the stated willingness-to-pay threshold of US$100,000 per QALY in 98.8% of probabilistic simulations.

Courville et al. 97 compared three strategies of screening for and treating S. aureus colonisation to prevent deep surgical site infections following total knee arthroplasty and total hip arthroplasty. Courville et al. 97 found that decolonisation of all preoperative patients with mupirocin, without testing for S. aureus, was the dominant strategy (cheaper and more effective) when compared with treating only patients testing positive for S. aureus, or no screening or decolonisation for S. aureus.

Intraoperative components of the enhanced recovery pathway

Two studies investigated strategies to reduce allogenic blood transfusions: collecting autologous blood prior to surgery100 or aseptic collection of wound drainage. 99 Autologous blood prior to surgery was found to be cost-effective, at US$2750 per QALY gained, whereas aseptic collection of wound drainage was not, at a cost of US$5.7M per QALY gained.

Marques et al. 101 conducted economic evaluations alongside two randomised controlled trials of adding local wound infiltration with bupivacaine to usual anaesthetic care for total hip arthroplasty and total knee arthroplasty. 101 The infiltration of local anaesthetic was found to be less costly and more effective than standard anaesthesia in both hip and knee replacement patients. Three studies used Markov models to investigate measures to reduce surgical site infection in the USA,103 UK102 and Australia. 104 The cheapest and most effective measure in all three included use of antibiotic-impregnated cement. The two studies that looked at other factors102,104 each found the use of prophylactic systemic antibiotics to be less costly and more effective than non-use, and the use of conventional ventilation in operating theatres to be less costly and more effective than laminar airflow ventilation. Graves et al. 102 considered use of body exhaust suits and found them to be dominated by strategies that did not include use of these suits.

Postoperative components of the enhanced recovery pathway

Fusco and Turchetti105 used a Markov model to evaluate a strategy of 10 face-to-face rehabilitation sessions followed by 10 telerehabilitation sessions after knee replacement, compared with 20 face-to-face sessions. They found the strategy including telerehabilitation to be cost-saving and to improve range of movement (knee flexion). However, they found no utility data for patients following a telerehabilitation programme, so for their base case assumed it to be non-inferior to face-to-face rehabilitation. In a sensitivity analysis, if telerehabilitation conferred an improvement in quality of life of at least 2.5%, the strategy’s probability of being cost-effective was 1 (at a willingness-to-pay threshold of €30,000/QALY).

Kauppila et al. 86 performed an economic evaluation of a 10-day outpatient rehabilitation course between 2 and 4 months after knee replacement, which included clinical assessments, physical activity, sessions with a psychologist, and lectures from an orthopaedic surgeon, nutritionist and social worker. 107 They found that patients who completed this course had higher costs and slightly worse quality-of-life outcomes over the 1 year of follow-up (albeit not statistically significant) than those receiving conventional orthopaedic care.

Bolz et al. 106 compared three follow-up strategies: 2-yearly routine follow-up, follow-up twice (at 3 months, and between 1 and 2 years after surgery) or no follow-up. The model assumed that no revisions would be delayed in either strategy that included follow-up and the outcomes for these two strategies were identical in each analysis. The no follow-up strategy was dominant for any assumed rate of delayed revision between 1% and 50%.

Assessment of study quality and reporting quality

Using the CHEC list, the quality of the studies was generally good (Figure 23a). However, the time horizons in the four trial-based studies (1 year86,98,101,109) and two model-based studies (1 year97 and 7 years106) were too short to capture all relevant outcomes or costs. Furthermore, most studies did not report whether or not they had conflicts of interest.

FIGURE 23.

Assessments of study quality based on tools from (a) CHEC; (b) ISPOR questionnaire; (c) AdViSHE tool; and (d) the Cochrane Collaboration.

Using the Cochrane Collaboration’s tool (see Figure 23b), the risk of bias of the trial-based studies was rated as being low for items such as incomplete outcome data and selective reporting, but high or uncertain for the remainder. Three trial-based studies stated that participants were allocated at random;79,101,107 the fourth was a before-and-after trial whereby patients were recruited consecutively from a waiting list for hip and knee replacement. 98 Larsen et al. 79 discussed stratification and Kauppila et al. 107 reported a computer-generated sequence. The risk of bias could not be assessed accurately, as the random sequence generation was not described precisely in any of the studies. The staff administering the interventions were not blinded to allocation in any trial, and in three trials98,101,107 outcomes were assessed by researchers aware of the treatment allocation.

Using the ISPOR questionnaire, the quality of the model-based studies was generally good (see Figure 23c). However, none of the model-based studies reported a detailed process for internal and external validation. Three studies99,100,105 were based on previously published models. 111–113 Of these, only Briggs et al. 111 (the basis for Fusco and Turchetti’s model105) provided details of model validation. Further limitations in model validation were highlighted with the AdViSHE tool (see Figure 23d). 93

Methods of estimating costs of joint replacement

Results of costs of joint replacement

We identified 397,119 and 457,747 patients as having had a primary THR or TKR, respectively, in the NJR, HES and PROMs linked data set. Table 16 reports the baseline characteristics of the two cohorts. There was a lower proportion of women in the THR cohort than in the TKR cohort (40.4% vs. 57.0%). More younger individuals (69.1 years vs. 69.5 years) with lower BMI (28.8 kg/m² vs. 30.7 kg/m²) underwent THR than TKR. Furthermore, the Oxford and EQ-5D scores were slightly lower in the THR cohort than in the TKR cohort (17.4 vs. 18.2 for OHS/OKS and 0.33 vs. 0.37 for EQ-5D).

Hospitalisation costs

The follow-up for the THR and TKR cohorts was 3.9 years (SD 2.5 years) (Table 17). In the first year after the joint replacement, 1.2% (4071/344,721) and 0.8% (2965/394,118) of individuals died in the hip and knee cohorts, respectively. An improvement at 6 months was reported for both cohorts using the EQ-5D and OHS/OKS instruments.

Hospitalisation costs associated with index admission were £6208 (median £5824, SD £969) for THR compared with £6122 (median £5692, SD £967) for TKR. The mean LOS in the index admission was 4.82 [median 4, SD 3.8, interquartile range (IQR) 3–6] days and 4.77 (median 4, SD 3.5, IQR 3–5) days for THR and TKR, respectively. These costs and LOS represent averages across the whole observation period. However, we found no clear trend in costs over the observation period when resource use was valued using the recommended NHS Reference Costs 2016–17114 (Table 18).

The mean 1-year hospitalisation costs were estimated to be £7817 (median £6258, SD £4618) and £7784 (median £6226, SD £4520) for THR and TKR, respectively, of which the index admission accounted for 79.4% and 78.5% of the total. Hospitalisation costs and LOS within 1 year were highly correlated for both types of joint replacement (Spearman’s correlation coefficient 0.84). Comparing 1-year costs over the observation period, we found that joint replacements decreased in each year relative to the previous year (see Table 18). In 2015, TKR and THR 1-year costs were, respectively, £194 (95% CI £136 to £252) and £253 (95% CI £191 to £315) lower than in 2008, adjusting for other covariates.

Musculoskeletal diagnosis (ICD-10 chapter 1335) accounted for 32–35% of readmission costs within the first year of joint replacement, followed by injury diagnosis (ICD-10 chapter 19:35 21%) and circulatory system diagnosis (ICD-10 chapter 9:35 8–9%) (Table 19).

Individuals undergoing partial knee replacement had, on average, lower 1-year costs and LOS [mean £6897 (SD £3263) and 4.4 (SD 5.9) days, n = 29,066] than individuals undergoing patellofemoral joint replacement [£7381 (SD £3892) and 5.4 (SD 7.7) days, n = 4370] and TKR [£7860 (SD £4606) and 7.5 (SD 11.5) days, n = 360,682]. Individuals undergoing metal-on-metal (MoM) resurfacing had, on average, lower 1-year costs and LOS [mean £6553 (SD £2900) and 4.8 (SD 5.8) days, n = 7000] than individuals undergoing THR [£7842 (SD £4643) and 7.4 (SD 11.3) days, n = 337,721].

After adjusting for censoring, the mean 1-year costs were similar to the complete-case analysis, at £7827 (95% CI £7813 to £7842) and £7805 (95% CI £7790 to £7818) for THR and TKR, respectively. For THR, the mean cost in the first 2 years following joint replacement (2-year) adjusted for censoring was £9258 (95% CI £9233 to £9280), compared with £9277 (SD £6927) using only the complete cases (n = 293,618). For TKR, the cost in the first 2 years following joint replacement (2-year) adjusted for censoring was £9452 (95% CI £9430 to £9475), similar to £9446 using only complete cases (n = 333,123). Comparing 2-year costs over the observation period, we found that joint replacements decreased each year relative to the previous year (see Table 18). In 2014, TKR and THR costs were £373 (95% CI £283 to £463) and £323 (95% CI £233 to £413) lower than in 2008, respectively, adjusting for other covariates.

Predictors of hospitalisation costs in the first year following joint replacement

Data were missing for 50% and 70% of patients concerning Oxford and EQ-5D scores (before surgery and at 6 months), BMI or other predefined variables to inform the prediction of hospitalisation costs for hip and knee replacement, respectively. The cohorts with incomplete data had a lower proportion of total joint replacements, higher mortality rates at 1 year, higher hospitalisation costs and higher LOS than complete cases. Following multiple imputation, the predictors of hospitalisation costs for THR and TKR are shown in Tables 20 and 21, respectively. A GLM with gamma family and identity link function was appropriate and had a good fit.

Holding all else constant, conventional THR was more expensive, on average, than MoM resurfacing (£451; p < 0.001). Women had higher mean hospitalisation costs than men (£167; p < 0.001) and increasing age was associated with higher costs (£28 per additional year). Individuals with higher quality-of-life values (EQ-5D and OHS) at baseline and reporting improvements at 6 months were associated with lower hospitalisation costs. Larger femur head size was associated with higher costs, with additional costs of up to £226 for head sizes of > 53 mm compared with the cost for head sizes of ≤ 28 mm. We found significant variation in hospitalisation costs concerning bearing surfaces with strong evidence (p < 0.01) that ceramic-on-ceramic, ceramic-on-metal and metal-on-ceramic bearings were associated with lower mean 1-year hospitalisation costs than metal-on-polythene bearings (the most common bearing type in the cohort). Costs were lower in recent years (–£31/year; p < 0.001), holding all else constant. Complications and revisions within the year were significantly associated with higher mean costs, with an additional £6601 (p < 0.001) and £11,255 (p < 0.001), respectively.

Holding all else constant, TKR was significantly associated with higher 1-year hospitalisation costs than unicondylar knee replacement (£391; p < 0.001). Women had higher mean hospitalisation costs than men (£248; p < 0.001) and costs were positively associated with age (£31 per additional year; p < 0.001) and higher deprivation, with individuals living in more deprived areas having higher costs. Individuals with higher quality-of-life values (EQ-5D and OKS) at baseline and those reporting improvements at 6 months had lower hospitalisation costs. Higher deformity and lower range of flexion were also significantly associated with higher costs. Costs were lower in recent years (–£10/year; p < 0.001), holding all else constant. Complications and revisions within the first year were significantly associated with higher costs, with an additional £6075 (p < 0.001) and £7753 (p < 0.001), respectively.

The mean 1-year hospitalisation costs were higher for patients who died (an additional £4682 for THR and £4597 for TKR), mostly as a result of higher LOS than those who survived the first year (mean of 30.4 days vs. 7.0 days for hip patients who did/did not die and 30.0 days vs. 7.0 days for knee patients who did/did not die).

Costs before and after joint replacement

Adding primary, outpatient and inpatient hospitalisation costs, the mean NHS costs associated with THR amounted to £9295 in the year of surgery (Figure 24) compared with £9483 following TKR (Figure 25). Using the annual number of UK primary joint replacements in 2017, the NHS primary and hospital care costs were estimated at £897M (n = 96,717) and £1007M (n = 106,334) in the year of the THR and TKR, respectively.

FIGURE 24.

Costs in the years before and after THR.

FIGURE 25.

Costs in the years before and after TKR.

Hospitalisation costs accounted for the highest proportion of the total 1-year cost for both THR and TKR (82–84% of total costs in the year of joint replacement), followed by primary care (10–11% of total costs) and outpatient care (6–7% of total costs). In the second year after joint replacement, total costs decreased to £3095 for knee and £2692 for hip cohorts, with inpatient costs again being the largest component (53% for both knee and hip).

Figure 26 reports the hospitalisation costs in the months before and after joint replacement. The annual hospitalisation costs in the year of joint replacement were £6753 (95% CI £6732 to £6774) and £6563 (95% CI £6544 to £6583) higher for THR and TKR, respectively, than those of the previous year (Table 22). However, hospitalisation costs decreased in the 5 months prior to surgery, reflecting lower hospital admissions leading up to the elective admission. In the year of joint replacement, about 81–83% of hospitalisation costs occurred in the first month post joint replacement. Costs in the second year after joint replacement were £389 (95% CI £370 to £407) and £349 (95% CI £329 to £368) higher than costs in the year prior to surgery for knee and hip replacement, respectively.

FIGURE 26.

Hospitalisation costs in the months before and after joint replacement.

A similar pattern was observed with outpatient and primary care costs (Figures 27 and 28). Primary care costs in the year of the surgery were higher than in the previous year, at £64 (95% CI £37 to £91) and £5 (95% CI –£17 to £26) for TKR and THR, respectively (Table 23). The highest costs occurred in the first month after joint replacement. Outpatient costs were also higher in the year of surgery than in the previous year, with an additional £80 (95% CI £53 to £107) and £15 (95% CI –£31 to £61) for TKR and THR, respectively, with the highest costs occurring in the second month (Table 24). However, outpatient costs in the second year after surgery were significantly lower than in the year preceding the surgery for both types of joint replacement [–£105 (95% CI –£78 to –£133) and –£126 (95% CI –£109 to –£143) for knee and hip, respectively]. In contrast, primary care costs were lower in the second year after surgery for THR (by –£53), but higher for TKR (by £37) than in the year preceding surgery.

FIGURE 27.

Primary care costs in the months before and after joint replacement.

FIGURE 28.

Outpatient care costs in the months before and after joint replacement

Cost-effectiveness of enhanced recovery

We identified 11 studies presenting cost–utility data for components of an ERP. 86,97,99–106,110 However, these cover only a few of the potential ERP components and the available evidence was mostly based on single studies. These studies covered a variety of interventions, across different health-care systems, and used different cost perspectives. The studies identified in this review were generally of good quality, according to the CHEC list,91 with a short time horizon identified as a key limitation in six studies. 86,97,98,101,106,109 When the models were assessed against the ISPOR questionnaire,123 there were concerns about the lack of model validation work, potentially questioning the reliability of 9 of the 13 studies identified. 97,99,100,102–106,110 The trials were generally of good quality.

Our review had some limitations. We may have missed relevant evidence by limiting our search to reports published in the English language and excluding studies that did not report QALYs. Of the studies excluded because they did not report QALYs, one investigated a complete ERP. 124 This was a retrospective review of total knee arthroplasty patients in Turkey who followed either a rapid-recovery protocol or a standard care protocol. The rapid-recovery protocol was found to be cost-saving and associated with statistically significant differences in knee flexion and extension at 6 months.

Costs of joint replacement

Consistent with Chapter 5, we found 1- and 2-year hospitalisation costs to decrease over recent years. However, we found no clear trend concerning the costs of the elective surgery itself, despite a significant decrease in LOS between 2008 and 2016. This can be explained by the nature of costing in the NHS, which is based on HRGs for hospital episodes rather than days in hospital. Therefore, it is key to note that such a costing approach was shown here not to be sensitive to capture what we would expect to be a decrease in costs.

We identified a reduction in hospitalisation costs in the 5 months prior to surgery for both types of joint replacement, reflecting fewer hospitalisations leading up to the elective admission. This suggests that the selection of patients for joint replacement takes into account their recent history of hospital admissions. Furthermore, primary care costs were slightly lower in the second year after surgery for THR, but slightly higher for TKR, than in the year preceding surgery, possibly reflecting differences in recovery times between the two procedures.

The main predictors of costs were similar for THR and TKR. As in previous work,122 we found preoperative quality of life to be associated with hospitalisation costs: 1-year costs were higher for individuals with worse preoperative quality of life, even after adjusting for other covariates. In addition, 1-year costs were lower for individuals reporting improvements in quality of life at 6 months.

Having a revision or complication was found to be associated with very high additional costs in the first year after surgery, after adjusting for other covariates. Given the high annual primary care and hospital costs of £897M and £1007M for THR and TKR, respectively, there is a considerable economic incentive to fund research aimed at identifying cost-effective ways of improving the quality of life of patients following joint replacement and reducing the risk of revisions and complications.

Our study had some limitations. NJR data were obtained for individuals undergoing joint replacement, with osteoarthritis as an indication for surgery. Hence, individuals without osteoarthritis as one of the indications were not available for analysis (e.g. rheumatoid arthritis or hip fracture). A large proportion of the cohort had missing data for one or more key covariates of the hospitalisation costs, in particular EQ-5D or OHS responses and BMI, which necessitated the use of missing data methods, specifically multiple imputation. A key assumption whenever multiple imputation is utilised is that the missing data may be classified as missing at random. This assumption is always untestable, but owing to the large number of relevant covariates in our linked data we judged it to be reasonable in this case.

Natural experimental study

Our data show that outcomes of primary THR and TKR have been gradually improving over time, with a decrease in LOS, reduction in estimated average inpatient bed-day cost, improvement in patient-reported outcomes of pain and function, decrease in complications and reduction in 5-year revision risk. These trends of improving outcomes were seen across all age groups, and in those with and without comorbidity. These findings are positive, in highlighting that reductions in LOS have been achieved without adversely affecting patient outcomes. However, these trends of improving outcomes had begun prior to the formal roll-out of ERPs through the national NHS ERAS implementation programme and were not temporally associated with the national ERAS programme. It is of note that when we looked at the effect of individual hospitals in a region of England, a significant impact of ERAS on LOS was seen at Southampton and the Isle of Wight, where, prior to implementing ERAS, LOS was stable and not changing, and, after ERAS implementation, LOS declined substantially. However, in many hospitals, formalising ERAS implementation saw no improvements in LOS, predominantly as mean LOS had already been declining, with little further reductions in LOS once ERAS was formally implemented.

Health economics

We found limited cost–utility evidence, either for an entire ERP or for individual components of a pathway, for patients having THR or TKR. There are also concerns regarding the ability of short time horizons in trials in this area to capture relevant outcomes, and regarding a general lack of reporting of model validation. However, our findings support the use of ERPs as a whole, prophylactic systemic antibiotics, antibiotic-impregnated cement and conventional ventilation.

We estimated the hospitalisation costs associated with index admission for THR to be £6208, compared with £6122 for TKR. Within 1 year of the joint replacement, the total hospitalisation costs were estimated to be £7827 and £7805 for THR and TKR, respectively, of which the index admission accounted for 79.4% and 78.5% of the total. Although complications, revision and reoperation are uncommon following surgery, they are associated with high costs. For THR, complications and revisions were significantly associated with higher costs, at £6601 and £11,255, respectively. Given that there are > 200,000 patients receiving THR and TKR each year in the UK, the costs associated with these operations are high, and remain higher in subsequent years after the operation. However, THR and TKR are the most cost-effective options in individuals with osteoarthritis. There is then a strong economic incentive to further improve the quality of life of individuals following joint replacement and reduce revision surgery and complications of surgery.

Process evaluation (qualitative research)

The ethnographic study of health professionals involved in the delivery of enhanced recovery services found that support for the care pathway was higher when the pathway was developed internally rather than externally. Guidance was flexible and could be adapted to meet the demands of individual hospital services. There were concerns about a lack of post-discharge support and tensions between primary and secondary care. One of the key elements of success was effective multidisciplinary collaboration, achieved through multidisciplinary team meetings and paperwork. Involving strong opinion leaders in its development and ‘champions’ who drove through implementation and acted as a point of contact helped facilitate implementation. The study highlighted the importance of ensuring that protocols are sufficiently flexible to meet individual patient needs. Services should also prioritise strategies to empower patients in their recovery through education.

The study of patients’ experiences of the ERP found that hip and knee schools (classes) for patients undergoing the surgeries were highly valued. Patients felt that they helped prepare them for surgery and provided ‘confidence’ in the care they would receive. They were enthusiastic about the potential for early discharge and invested in their recovery, preparing them to take on the role of ‘active’ participants in their care. Most thought that it was useful to practise exercises to prepare them for postoperative mobilisation. Most participants were presented with the option of local or general anaesthesia, although they had different experiences of this. Although some were given this option during hip or knee schools or at their preoperative assessment, others were asked on the day of surgery, meaning that they felt unable to make an informed choice. Patients were surprised about being able to mobilise so quickly after surgery. There was a sense of pride in being able to fulfil what they thought was required of them and they benchmarked their success against their preoperative education and feedback from health-care professionals. Participants emphasised the importance of participating in exercises and working towards their own recovery, engaging in the emotional and physical work required of them as ‘active’ patients. Most of those interviewed reported feeling ‘ready’ to return home and considered it preferable to remaining in hospital. Some participants reported feeling under pressure to be discharged before they felt well enough and this was linked to concerns that they were being made to mobilise too quickly. A small number of participants attributed this to a demand for beds. Those who lacked family support, particularly those who lived alone, were more anxious. The study highlights how patients might value more information and support from health-care professionals, particularly in the post-discharge period, to enable them to work more effectively towards their own recovery.

Research in context

Although at a national level we did not observe any impact of the ERAS pathway on changing trends in outcomes of surgery, hospitals will have implemented this at different time points, and may be early adopters having already implemented many of the ERP components before the national programme had begun. We therefore looked at the effect within individual hospitals within a region of England, and conducted the statistical interrupted time series analysis within these individual hospitals, focusing on the outcomes of pain/function and LOS. Three of these NHS hospitals were part of the qualitative process evaluation, allowing us to set the findings of the statistical analysis in context of what we observe from the qualitative data.

For all three of these hospitals, it is clear that there is no impact of the date they implemented ERAS on the OHS and OKS pain and functional outcome scores, these generally being continuations of existing trends. However, all three hospitals observed reductions in LOS over the study period.

In Shinebury, LOS for THR and TKR was already declining prior to the official ERAS start date, and it continued to decline over the study duration. This hospital was an early adopter and innovator and hence the official start date of 2009 from the NHS England planned care programme just reflects that ERAS was already implemented prior to the official national programme. Enhanced recovery and early discharge started as a research project for a hip arthroplasty database that the trust had been running since 1998. ERAS was piloted in 2000, and was originally initiated by a key consultant surgeon and a lead physiotherapist who has since moved on. Physiotherapists gradually took on more ERAS duties, in the form of 2- and 6-week joint review clinics from about 2002, and then hip school from about 2004. Digital versatile discs (DVDs) were first recored in 2005 and became available about 2006. Joint review clinics originally included both hips and knees, but, as the number of patients increased, two separate clinics were introduced (circa 2008). There were two nurses (sisters) who were the ‘main kind of driving force behind the clinic’, and were also central points of contact for doctors and consultants. The hospital won an innovation award in 2006 for implementing the DVD, booklet and clinic.

Fitness for surgery is evaluated at pre-assessment and includes full medical checks, and patients sometimes attend hip/knee school on that same day as well. Follow-up takes place at the joint review clinic at 2, 6 and 52 weeks; this takes place with the same team and in the same place as knee/hip school. If there are extra difficulties, the patient is seen by the consultant as an outpatient. All patients have to attend hip/knee school and they receive a booklet and a DVD as part of this: ‘they are as educated as they could possibly be’. The DVD includes consultants talking about THR surgery and previous patients talking about their experiences, while demonstrating some of the exercises. Hip and knee patients follow identical care pathways. Patients see OTs at pre-assessment and complete questionnaires (home environment sheet) about their home environment, to assess the level of support that will be needed post discharge. OTs have a printed list every week of joint patients coming to pre-assessment, so that they know in advance who they are expecting and when. Patients do not necessarily see an anaesthetist at pre-assessment, unless requested by the surgeon. For complex patients, notes are given to the anaesthetist in between pre-assessment and surgery. Patients are seen by physiotherapists and mobilised on day 1 after surgery. Joint review clinics are run by the lead physiotherapist, OTs and specialised orthopaedic nurses.

What is surprising is that, although an ERAS service had been in place for many years, LOS for this hospital trust (Shinebury) was still high at the start of the study in 2008, and was still able to decline so much. This may partly reflect coding of hospitals in the HES database (see Appendix 1, Figure 29). Shinebury Hospitals NHS Foundation Trust came into being in January 2012 as a result of the integration of two different NHS trusts, which achieved foundation trust status in 2006. Separating these sites out, we can see that LOS was already very low for THR at this site, whereas for TKR LOS was much higher in both hospital sites.

In Towerton Hospital, LOS for both THR and TKR was relatively high, with a stable trend prior to the ERAS start date; however, after the start date there was a significant reduction and decline in mean LOS (TKR: 5.6 days in April 2008 declining to 4.3 days in December 2016; THR: 6.2 days in April 2008 declining to 4.7 days in December 2016). In this hospital setting, there is a clear impact of ERAS on reducing LOS. The process evaluation shows that Towerton Hospital had a very strong and comprehensive ERAS service. There was a ‘nurse champion’ who, along with the matron, ran multidisciplinary team groups to decide how to implement ERAS at the hospital, and ran education classes for staff. The surgical pre-assessment involves a clinical assessment to establish patients’ fitness for surgery, taking next of kin details and who to call post operation. Pre-assessment includes looking at patients’ social history and current living arrangements, and home measurement forms are given to patients for them to complete and bring to hip/knee school, where a physiotherapist or OT can pick them up. Patients see OTs one to one at the end of the hip/knee school sessions if they are eligible for help (i.e. if they live in a certain postcode area). Equipment can be delivered to eligible patients’ homes preoperatively. Physiotherapists see patients on the ward as soon as possible post operation (sometimes day zero if appropriate, but no later than day 1 after surgery). Postoperative, OTs do a washing/dressing and kitchen assessment, to see whether or not extra support from social services is required. An ERAS nurse champion attends ward round meetings every morning, as do all multidisciplinary team staff. The hospital uses a ‘step-down’ ward (part of a maternity unit at a nearby hospital) for patients who are not quite ready to go home but are ‘too good’ to stay on the elective ward. Follow-up knee classes are held in the outpatient department, which is at another local hospital. Hip patients are not routinely scheduled for outpatient physiotherapy, unless they need it for a particular reason. The hospital provides education sessions and a ‘rolling education programme’, so ERAS is an ongoing and evolving concept.

In Elmfield Hospital, a different pattern was observed for THR and TKR. For THR, LOS was already declining. However, after the ERAS start date, there was no further decline and, if anything, LOS started to increase afterwards. For TKR, LOS was in decline and continued on the existing downwards trend after the official ERAS start date. From the information obtained about the service, there does not appear to be an officially running ERAS programme; rather, ERAS methods are used to reduce waiting times for hip replacement but not for knee replacement. There are separate care pathways for hip and knee patients. The initial implementation of ERAS for hips was through funding for a specialist physiotherapist, and involved streaming patients to include those considered appropriate for it. This was decided on the basis of a pro forma completed by patients, giving details of medical and social ‘fitness for it’. Patients whom Elmfield identified as suitable were flagged at the time they were first seen as outpatients. However, this funding for the physiotherapist was lost after about 18 months. Currently, it has a twice-monthly hip arthritis clinic set up to reduce waiting time, and a same-day assessment clinic so patients attend only once for all preoperative assessments. The anaesthetists introduced an anaesthetic protocol that included preoperative, perioperative and postoperative pain control and management. There is a hip school that is run twice weekly. There is not a formal ERAS for knees, but they do day-case UKR, and knee patients are also streamed according to comorbidities and analgesic requirements, but there is no knee school. There is nothing in the patient pre-assessment clinic paperwork to indicate that patients have been streamed to ERAS. At the beginning of ERAS implementation, home visits were undertaken as standard; this no longer happens. OTs used to see patients in pre-admission, but, since same-day assessment was introduced, they now see patients at a pre-admission clinic only if a ‘trigger’ has been flagged during the assessment; otherwise OTs do not see patients until hip school or post operation. It is ‘not the norm’ for patients to be specially referred to an OT.

Implications for practice

There is variation in patient outcomes of primary THR and TKR. The determinants of this variation remain unknown and cannot be explained by a hospital treating healthier or fitter patients. In those hospitals with poorer outcomes, there is a need for hospital mangers and clinicians to understand why their outcomes are worse than others and take action to address this.

Reductions in LOS can be achieved without having a negative impact on patient outcomes. ERAS services are valued by both patients and health professionals involved in running these services. Patients valued hip and knee schools, feeling that they helped to prepare them for surgery and provided ‘confidence’ in the care they would receive. Patients emphasised the importance of participating in exercises and working towards their own recovery, engaging in the emotional and physical work required of them as ‘active’ patients. Although we found no observed impact of the national implementation of ERAS on patient outcomes, this probably reflects early adopters and innovators having already introduced ERAS components into their service, with the continuing decline in LOS nationally reflecting additional hospitals introducing ERAS. Exploring the effect of ERAS in individual hospitals, we can see that in sites that are implementing ERAS for the first time (those with a relatively high and stable LOS prior to implementation date), the introduction of ERAS can have an impact, with subsequent trends of reducing LOS. Those hospitals that are still observed to have a high LOS in our current years of registry data would probably see benefit from implementing ERAS principles.

Scope for future work

• Although national trends towards improved patient outcome are encouraging, there is still substantial variation in outcome between hospital trusts and CCGs, so there is work to be done to understand why these potentially unwarranted variations in patient outcome exist, particularly those with consistently worse outcomes over time.

• Further research should explore patient experiences of referral from primary to secondary care services, to provide a more comprehensive account of experiences through the care pathway. In addition, future research may also consider GPs’ views and experiences of ERAS, particularly at the points of referral and discharge. This would provide a ‘system-wide’ understanding of the delivery of ERAS.

Posters and presentations

Garriga C, Prats-Uribe A, Price A, Prieto-Alhambra D, Carr A, Arden N, et al. Repercussion of a National Intervention on Patient Outcomes of Primary Knee Arthroplasty to Enhance the Recovery Pathway: 2008–2016 England Trends. World Congress on Osteoporosis, Osteoarthritis and Musculoskeletal Diseases (WCO-IOF-ESCEO), Kraków, 19–22 April 2018. Poster number: P674.

Garriga C, Prats-Uribe A, Price A, Prieto-Alhambra D, Carr A, Arden N, et al. Impact of a National Enhanced Recovery Programme on Patient Outcomes of Primary Total and Unicompartmental Knee Replacement: 2008–2016 England Trends. Osteoarthritis Research Society International (OARSI), Liverpool, 26–29 April 2018. Abstract number: 387. Poster.

Garriga C, Price A, Prieto-Alhambra D, Carr A, Arden N, Leal J, et al. Impact of a National Enhanced Recovery Programme on Patient Outcomes of Primary Hip Replacement: An Interrupted Time Series Analysis from England. 19th European Federation of National Associations of Orthopaedics and Traumatology (EFORT) Congress, Barcelona, 30 May–1 June 2018. Oral. Abstract number: 1614.

Drew S, Judge A, Cohen R, Gooberman-Hill R. Enhanced Recovery After Surgery Implementation in Practice: An Ethnographic Study of Services for Hip and Knee Replacement. British Sociological Association, Medical Sociology Annual Conference, Glasgow, September 2018.

Garriga C, Leal J, Price A, Peat G, Prieto-Alhambra D, Carr A, et al. Geographical Variation in Outcomes of Primary Hip Replacement: Study from the National Joint Registry of England and Wales. World Congress on Osteoporosis, Osteoarthritis and Musculoskeletal Diseases, Paris, 4–7 April 2019. Poster presentation: P376.

Garriga C, Leal J, Price A, Prieto-Alhambra D, Carr A, Rangan A, et al. Geographical Variation in Patient Outcomes of Primary Knee Replacement Across Clinical Commissioning Groups: Study from the National Joint Registry of England, Wales, Ireland and the Isle of Man. Annual European Congress of Rheumatology (EULAR 2019), Madrid, 12–15 June 2019. Oral. Abstract number OP0306.

Articles

Murphy J, Pritchard MG, Cheng LY, Janarthanan R, Leal J. Cost-effectiveness of enhanced recovery in hip and knee replacement: a systematic review protocol. BMJ Open 2018;8:e019740.

Cohen R, Gooberman-Hill R. Staff experiences of enhanced recovery after surgery: systematic review of qualitative studies. BMJ Open 2019;9:e022259.

Drew S, Judge A, Cohen R, Fitzpatrick R, Barker K, Gooberman-Hill R. Enhanced recovery after surgery implementation in practice: an ethnographic study of services for hip and knee replacement. BMJ Open 2019;9:e024431.

Garriga C, Leal J, Sánchez-Santos MT, Arden N, Price A, Prieto-Alhambra D, et al. Geographical variation in outcomes of primary hip and knee replacement. JAMA Netw Open 2019;2:e1914325.

Garriga C, Murphy J, Leal J, Price A, Prieto-Alhambra D, Carr A, et al. Impact of a national enhanced recovery after surgery programme on patient outcomes of primary total knee replacement: an interrupted time series analysis from ‘The National Joint Registry of England,Wales, Northern Ireland and the Isle of Man’. Osteoarthritis Cartilage 2019;27:1280–93.

Garriga C, Murphy J, Leal J, Arden NK, Price A, Prieto-Alhambra D, et al. Assessment on patient outcomes of primary hip replacement: an interrupted time series analysis from ‘The National Joint Registry of England and Wales’. BMJ Open 2019;9:e031599.

Murphy J, Pritchard MG, Cheng LY, Janarthanan R, Judge A, Leal J. Enhanced recovery following hip and knee arthroplasty: a systematic review of cost-effectiveness evidence. BMJ Open 2020; in press.