Impact of interventions to improve recovery of older adults following planned hospital admission on quality-of-life following discharge: linked-evidence synthesis

Search strategy

For the systematic review of quantitative studies, the searches for our previous review were re-run with adaptations. 17 Search terms included terms for older people or interventions commonly undergone by older people, combined using the AND Boolean operator with terms for multicomponent interventions or terms that describe reducing length of stay, for example, ‘length’ adjacent to ‘stay’ adjacent to ‘reducing’. The full search strategy for MEDLINE ALL is included in Appendix 1. We adapted the search to include search terms for multicomponent interventions which were not relevant for our previous review, including supported discharge, and home or community rehabilitation. We also applied an adapted version of the study type filter used for our previous review, with an expanded set of search terms for non-randomised trials and controlled before-and-after (CBA) studies. 17 These terms were derived from the Cochrane EPOC study design filter (Paul Miller, EPOC, 23 August 2017, personal communication) and from inspecting the titles and abstracts of non-randomised and BA studies that were identified via supplementary searches for our previous review but which the bibliographic database searches failed to retrieve, thus ensuring that the bibliographic database searches had improved sensitivity. 17 In addition, we added search terms for quality-of-life studies.

The adapted search was developed by SB in conjunction with the review team and stakeholders in MEDLINE (via Ovid) and adapted for use in other databases. The full set of bibliographic databases searched included: MEDLINE ALL, Embase and the Health Management Information Consortium (HMIC) (all via Ovid), CENTRAL (via the Cochrane Library), and Cumulative Index to Nursing and Allied Health Literature (CINAHL) and Allied and Complementary Medicine Database (AMED) (both via EBSCO). Searches were run in May 2021 and updated in April 2022. For the update searches, to improve the efficiency of the searching and screening process we used a Search Summary Table to identify the minimum set of bibliographic databases required to retrieve all included studies identified by the initial set of searches, and limited the search to these databases. 28 Thus, we ran the update searches in MEDLINE (via Ovid) and CENTRAL only.

Search results for both initial and update searches were exported to EndNote 20 (Clarivate Analytics, Philadelphia, PA, USA) and de-duplicated using manual checking and the EndNote de-duplication tool. To expedite the study selection process, the 218 articles included in our previous review and 282 articles we previously excluded due to population, country or language (and thus failing to meet inclusion for this review) were removed from the search results prior to screening.

We checked reference lists of all included studies and carried out forward citation searching of included studies from the initial search using the Science Citation Index (Web of Science, Clarivate Analytics) and Scopus (Elsevier) (DC). No citation searches were carried out on included studies identified by the updated searches. The results of forward citation searches were exported to EndNote 20, and reference-list checking was conducted using Microsoft Excel spreadsheets to document potentially useful studies thus identified.

Inclusion criteria

We sought studies of multicomponent interventions to improve and/or accelerate the recovery of older adults undergoing elective surgical procedures requiring an overnight stay in hospital. Additionally, studies had to assess at least one PROM relating to patient recovery. Specific inclusion/exclusion criteria were as follows.

Data processing

Between-group differences were evaluated at post-intervention and longest follow-up. For dichotomous outcomes (e.g. readmissions, complications) odds ratios (ORs) were calculated in Microsoft Excel using standard equations described in section 9.2.2 of the Cochrane Handbook for Systematic Reviews of Interventions, version 6.3. 27 In addition, the statistical significance of the OR was assessed by calculating a p-value from the z-score for the difference and ascertaining 95% confidence intervals (CIs). 33

For continuous outcomes (e.g. scores on PROMs) Cohen’s d was calculated using the metan command in Stata (version 14.2, StataCorp, College Station, TX, USA). Effect sizes were interpreted in line with Cohen’s guidance (i.e. where d = 0.2 to 0.49, class as ‘small’; where d = 0.5 to 0.79, class as ‘medium’; and where d = 0.8 or above, class as ‘large’). 34 In addition, 95% CIs for the effect were calculated using the metan command in Stata. The p-value for the difference was obtained using the ttesti command in Stata, using data from the two-tailed analysis.

Where mean and standard deviation (SD) for an outcome were not provided, we used standard approaches to impute the required values27 (Cochrane ref section 7.7.3). Our approach to imputation is described elsewhere. 17 We did not impute data where studies only provided median and range for an outcome, due to high risk of skewness in the data.

Where similar procedure, intervention, comparator and outcome categories were present for two or more RCTs, random-effects meta-analysis was conducted. Forest plots were produced as part of the metan command in Stata. Pooled effects with 95% CIs and p-values were reported. Statistical heterogeneity was assessed using the I²-statistic, with greater a percentage indicating a greater proportion of total variance due to between-study variance as opposed to sampling error. 35 Meta-analysis of ORs was performed using log-transformed data.

When multiple measures of LOS or complications were presented within the same outcome category for a study included in the meta-analysis, one measure was chosen as the ‘best representative’. In the case of LOS, this meant the outcome that most closely accounted for the longest portion of the hospital stay, without consideration of readmissions. For example, ‘total LOS’ would be chosen ahead of ‘postoperative LOS’. For complication data, summary or composite outcomes were preferred, rather than incidences of specific complications. When only incidences of individual complications were presented, they were summed and the total number was used in meta-analysis.

Where multiple PROMs were available within the same outcome category, for example multiple mental health measures, we sought to calculate a composite score using standard approaches. 36

Studies that were not eligible for meta-analysis were described narratively. This included a description of the main characteristics and findings of each study.

Effectiveness of ERP interventions at improving clinical outcomes (RCTs only)

Report Supplementary Material 4, Table 6 displays clinical outcome data from the seven RCTs evaluating ERP interventions in older adults undergoing elective LLA. 50,60,66,68,69,78,86,89 After imputation, standardised mean differences between ERP and TAU groups for LOS were available in three studies. 60,68,69,89 A forest plot displaying the results of meta-analysis of these three groups is displayed in Figure 2, showing that ERP interventions were associated with a reduction in LOS when compared with usual care (d = −0.79, 95% CI −1.44 to −0.15; p < 0.001). Although a large effect was estimated, wide CI and statistically significant heterogeneity for this effect (I² = 87.7%, p < 0.001) reflect inconsistency in the evidence and suggest the true effect could be large or small. Effect sizes could not be calculated in two studies which provided variance data in the form of ranges. 50,66 In both cases, LOS was considerably lower in the ERP group – median 1 versus 6 days in the study by Borgwardt and colleagues,66 and mean 1.5 days versus 4.3 days in the study by Reilly and colleagues. 50

FIGURE 2.

Forest plot showing the results of meta-analysis of the effect of ERP vs. usual treatment on LOS following LLA.

Five studies reported readmission rates for those receiving ERP interventions after LLA. 50,66,68,69,86,89 There were no readmissions in the study by Borgwardt and colleagues66 and only one, in the group receiving usual care, in the study by Siggeirsdottir and colleagues. 86 In the remaining three studies, meta-analysis indicates that the odds of being readmitted to hospital following surgery were similar whether ERP or usual care was received (OR = 1.35, 95% CI 0.40 to 4.56; p > 0.05; Figure 3).

FIGURE 3.

Forest plot showing the results of meta-analysis of the effect of ERP vs. usual treatment on the odds of readmission following LLA.

Complications were reported in four studies. 50,78,86,89 Pour and colleagues reported no short-term complications in either group, and when the remaining three studies were pooled, there was no overall difference in the odds of complications between ERP and usual care (OR = 0.90, 95% CI 0.36 to 2.26, p > 0.05; see Figure 4). CIs were notably wide in two studies,50,89 reflecting the low numbers of cases entered into the analysis.

FIGURE 4.

Forest plot showing the results of meta-analysis of the effect of ERP vs. usual treatment on odds of complications following LLA.

Effectiveness of ERP interventions at improving patient-reported outcomes

Report Supplementary Material 4, Table 7 displays data for patient-reported outcomes for RCTs trialling ERP interventions to improve recovery from LLA in older adults. While PROMS were reported in all studies, it was not possible to calculate effect sizes using the data provided by Pour and colleagues. 78 Pour and colleagues reported favourable outcomes in the experimental group for quality of life, functional improvement, energy and mental health. 78

Effectiveness of ERP interventions at improving clinical and patient-reported outcomes (non-RCTs only)

There were eight papers from seven non-RCTs that compared ERP interventions with usual care. 39–43,46–48 The studies by Garriga and colleagues used an ITS design to report on patients receiving hip or knee replacements as two groups, with data accessed via The National Joint Registry of England, Wales, Northern Ireland and the Isle of Man. 39,40 Despite having a very large sample size for each group of patients (nearly 500,000 patients for each type of surgery) it was unclear how many participants were analysed before and after implementation of ERAS, and no variance data were provided alongside group means, so we could not calculate effect sizes. 39,40 However, the authors reported a positive picture for both groups, with a reduction in LOS, improvement in Oxford Knee Score or Oxford Hip Score, lower rates of complications and revisions reported for each cohort with implementation of ERAS. 39,40

The study by Hunt and colleagues was presented across two papers,42,43 the authors comparing outcomes between centres with differing LOS in Belfast, Liverpool and London. The paper by Hunt and colleagues42 reports most outcomes from the study, with the paper by Salmon and colleagues focusing on patient satisfaction. 43 Where standardised mean differences were calculated, there were no statistically significant differences between outcomes except for improvement on the Oxford Hip Score in the study by Hunt and colleagues for patients in the Belfast centre compared to both control sites. 42 There was also a slight improvement on the Western Ontario and McMaster Universities’ Osteoarthritis Index (WOMAC) function scale in patients in Belfast compared to those in south west London (d = 0.23, 95% CI 0.02 to 0.44; p < 0.05). In the Belfast centre, more patients reported no problems with care than in Liverpool (18.8% more) and south west London (26.5% more), while the number reporting no problems with recovery was similar in all three centres. 43

Maempel and colleagues conducted two studies, with hip46 and knee47 replacement patients receiving different ERP packages. The study conducted with hip-replacement patients46 was considerably larger (n = 1161) than the one conducted with knee-replacement patients47 (n = 165). Median LOS was 1 day shorter in the ERP group with knee-replacement patients, with no difference in scores on the American Knee Society Global Functioning outcome. 47 In the study with hip-replacement patients, LOS was 2 days shorter in ERP patients, associated with a large effect size (d = −1.35, 95% CI −1.47 to −1.22; p < 0.001), with no differences between groups for other outcomes, including the Harris Hip Score. 46

In the study by Higgins and colleagues, it was unclear how many participants were included in each group when evaluating outcomes, therefore it was not possible to calculate effect sizes. 41 The authors reported that an ERP intervention was associated with a significantly shorter mean LOS of two days, reduction in rates of reoperation within 60 days and improvements on all PROMs, which included the Oxford Knee Score, EuroQol-5 Dimensions (EQ-5D) and quality of life VAS. 41 There were no differences in the 60-day complication rate and 30-day readmission rate.

McDonald and colleagues reported that ERP was associated with a reduction in LOS of 2 days, with a large effect size and narrow CIs (d = −1.10, 95% CI −1.20 to −1.0; p < 0.001). 48 This reduction in LOS did not come at a cost to scores on the Oxford Knee Score outcome, which was almost identical between patients receiving ERP and those receiving usual care. 48

Effectiveness of Prehab interventions at improving clinical outcomes (RCTs only)

Six RCTs evaluated the effectiveness of Prehab interventions on outcomes following LLA, compared to usual care. 49,51,53,63,77,79 It was not possible to calculate effect sizes for LOS in two studies;49,63 however, the remaining four studies could be pooled in meta-analysis. Figure 7 is a forest plot showing that pooled data indicate that there was no overall association between Prehab and LOS, despite tendency towards a small effect (d = −0.13, 95% CI −0.32 to 0.07; p > 0.05).

FIGURE 7.

Forest plot showing the results of meta-analysis of the effect of Prehab vs. usual treatment on LOS following LLA.

Hoogeboom et al. reported complications, with two in the intervention group and none in the control group,63 but none of the other RCTs evaluating Prehab interventions captured data on readmissions, complications or mortality.

Effectiveness of Prehab interventions at improving PROMs

Report Supplementary Material 4, Table 9 displays patient-reported outcome data from the six RCTs comparing Prehab with usual care. 49,51,53,63,77,79 Assessments of physical functioning were made in five studies. 49,51,53,77,79 using multiple measures in all but one study. 79 Measurements were taken at multiple post-operative time points (3, 6 and 12 months) in the study by Beaupre and colleagues,53 otherwise they were taken between 4 and 12 weeks after surgery.

It was possible to perform a meta-analysis of the physical functioning outcomes used in the six studies. We selected the 3-month time point and combined the Physical Component Summary score of the SF-36 with the Function score of the WOMAC to provide an input for Beaupre et al. 53 For the data provided by McGregor and colleagues49 we combined the Harris Hip Score with the Function score of the WOMAC, and for the data provided by Williamson et al. we combined the Oxford Knee Score with the WOMAC total score. 51 Meta-analysis indicates that there was an overall association between Prehab interventions and improved physical functioning scores, with a small effect demonstrated (d = −0.36, 95% CI −0.56 to −0.17, p < 0.001). However, observation of the forest plot from this analysis (see Figure 8a) indicates that the result from the paper by Soeters and colleagues77 appears to be an outlier, and thus sensitivity analysis was performed. When the large effect observed by Soeters and colleagues77 is removed, the remaining studies consistently demonstrate no difference between scores when comparing those receiving Prehab and those receiving usual care (d = 0.05, 95% CI −0.16 to 0.26, p > 0.05; see Figure 8).

FIGURE 8.

Forest plot showing the results of meta-analysis of the effect of Prehab vs. usual treatment on physical functioning PROMs following LLA. (A) Forest plot includes data from all eligible studies; (B) outlier data from Soeters et al. (2018) removed.

Four studies captured pain reports after discharge from hospital. 49,51,53,77 Meta-analysis reveals a lack of consistency between the four studies reporting outcomes (see Figure 9). Two studies showed no effect of Prehab on pain scores at 3 months post-operatively,51,53 one showed a large improvement in pain scores 4–6 weeks after surgery77 and one showed that pain was worse in the Prehab group after 3 months. 49 The pooled effect suggested no overall influence of Prehab on pain, with very wide CIs and an I² value close to 100%, reflecting the inconsistency of data in the meta-analysis (d = −0.23, 95% CI −1.18 to 0.72; p > 0.05).

FIGURE 9.

Forest plot showing the results of meta-analysis of the effect of Prehab vs. usual treatment on pain scores 2–3 months after LLA.

Quality of life was assessed in three studies evaluating the effects of Prehab on outcomes after lower-limb arthroplasty. 49,53,79 The general health scale of the SF-36 from Beaupre et al. ,53 the EQ-5D VAS from Cavill et al. 79 and the General Disabilities score based on the Barthel Activities of Daily Living Index from McGregor et al. 49 were pooled. Meta-analysis indicates no overall effect of Prehab on quality of life (d = 0.08, 95% CI −0.19 to 0.36, p > 0.05; see Figure 10).

FIGURE 10.

Forest plot showing the results of meta-analysis of the effect of Prehab vs. usual treatment on quality of life scores after LLA.

Mental health was evaluated in two studies, with Beaupre and colleagues53 reporting relevant subscales of the SF-36, and Williamson and colleagues51 reporting anxiety and depression scores on the HADS. Of the five mental health measures reported, only anxiety in the intervention group in the study by Williamson et al. improved with Prehab, as shown in Figure 11. 51 This was associated with a medium effect size, although CIs reflect uncertainty in the true magnitude of effect (d = −0.56, 95% CI −0.92 to −0.19; p = 0.03).

FIGURE 11.

Forest plot showing the effect of Prehab vs. usual treatment on mental health scores after LLA for individual measures in two studies. No meta-analysis was performed.

Physical activity was evaluated by Hoogeboom and colleagues, with no difference in activity levels between groups within a week of surgery. 63 Social status was assessed using the social function subscale of the SF-36 in the study by Beaupre and colleagues, and there were no differences between groups. 53

Effectiveness of Prehab versus Rehab

A single RCT compared Prehab with Rehab. 70 Vesterby and colleagues reported LOS in terms of median and range, therefore an effect size was not calculated for the difference between groups. 70 The median LOS was 1 day in the Prehab group and 2 days in the Rehab group. 70 Readmissions and unplanned visits and calls to hospital were similar between groups. In terms of PROMs, there were no differences between groups in quality of life, physical function or mental health. 70

Effectiveness of Rehab versus TAU

One RCT evaluated a Rehab intervention, compared with usual care. 75 Mean LOS was 6.5 days shorter in patients receiving Rehab intervention with similar numbers of complications in both groups. 75 In addition, those in the Rehab group achieved lower scores on the WOMAC measure, this difference associated with a large effect size (d = −1.52 95% CI −1.89 to −1.14; p < 0.001 – per-protocol analysis). 75

Summary of evidence from trials seeking to improve recovery from lower-limb arthroplasty

There were 15 RCTs evaluating interventions to improve recovery following LLA in older adults. These were split into seven studies evaluating ERP interventions, compared to usual care,50,60,66,68,69,78,86,89 six comparing Prehab with usual care,49,51,53,63,77,79 one comparing Prehab with Rehab70 and one comparing Rehab with usual care. 75 There were also seven non-RCTs evaluating the effects of ERP interventions. 39–43,46–48

There was evidence from meta-analysis to suggest that ERP interventions were associated with reduced LOS (n = 5 studies), without negative impact on readmissions (n = 3 studies) or complications (n = 3 studies). The influence on patient-reported outcomes was evaluated using a range of measures, including pain, physical function, quality of life and satisfaction. Scores tended to be better in the ERP group, or there was no difference between scores for patients receiving ERP usual care. Meta-analysis of PROMs was only possible for pain (n = 3 studies) and physical function (n = 4 studies), with evidence of a small beneficial impact on pain.

The data from non-RCTs were not combined but reflected a broadly similar story to that from the RCTs evaluating ERP interventions, in that LOS was usually improved with ERP, without detriment to, and occasionally linked with improvements in, other outcomes. However, there were issues with reporting that meant effect sizes were not readily calculable in many instances.

Four of the six RCTs evaluating Prehab interventions provided LOS data that could be pooled, indicating that this type of intervention was associated with no change in LOS compared with usual care. Other clinical outcomes, including readmissions, complications and mortality, were not reported, except in one study, so we were unable to evaluate whether Prehab interventions influenced these outcomes. However, it was possible to perform meta-analysis of various PROMs, which were better reported in this group of studies. Physical functioning was shown to be slightly improved with Prehab, but this was skewed by an outlier showing a very large effect of the intervention. When this was removed from the analysis, the remaining four studies clearly indicated no difference between Prehab and usual care.

Meta-analysis also showed there was no difference in pain scores 2–3 months after surgery (n = 4 studies), although there was very high heterogeneity between studies, and quality of life was similar between groups (n = 3 studies). For other patient-reported outcomes, where meta-analysis was not possible, there tended to be evidence of no difference between Prehab and usual care.

The single study comparing Prehab with Rehab showed no differences between groups, and the single study comparing Rehab with usual care showed large improvements in LOS and physical function.

Overall, there was a handful of studies available for the evaluation of ERP and Prehab interventions. The evidence suggests that ERP interventions can provide improvement in clinical outcomes and can offer this without detriment to, and even with improvements in, some PROMs, although further evidence is required. The data for Prehab interventions indicate that there is likely to be little impact, either positive or negative, on clinical or patient-reported outcomes.

Effectiveness of ERP interventions at improving clinical outcomes

Report Supplementary Material 4, Table 2 displays clinical outcome data from the seven studies evaluating ERP interventions in older adults undergoing elective colorectal surgery. 45,65,82–85,88 After imputation, standardised mean differences between ERP and TAU groups for LOS were available in three comparisons from two studies. 65,84 A forest plot displaying the results of meta-analysis of these three groups is displayed in Figure 12, showing that ERP interventions were associated with a small reduction in LOS when compared with usual care (d = −0.32, 95% CI −0.56 to −0.08; p = 0.01). Heterogeneity was not statistically significant for this effect (I² = 44.1%; p > 0.05), reflecting consistency in the evidence.

FIGURE 12.

Forest plot showing the results of meta-analysis of the effect of ERP vs. usual treatment on LOS following colorectal surgery. LOS, length of stay; LAP, laparoscopic surgery group; Open, open surgery group.

In the studies presenting data which could not be used to calculate standardised mean differences, median LOS for patients receiving ERP interventions was lower in all cases. 45,82,84,85,88 Additionally, Pappalardo and colleagues reported that patients receiving ERP had been discharged by postoperative day six, but around half of those receiving usual care had not. 83

Three studies, evaluating four comparisons, reported readmission rates following ERP interventions. 45,65,88 Figure 13 is a forest plot displaying the results of meta-analysis of readmission data, indicating the odds of readmission were similar between patients receiving ERP and usual care (OR = 1.07, 95% CI 0.62 to 1.87). In addition, there were no readmissions reported in either trial arm in the study by Lee et al. 84 Seven studies, evaluating eight comparisons, reported incidence of complications following ERP interventions. 45,65,82–85,88 Meta-analysis indicates that there was no difference in the odds of experiencing a complication between patients receiving ERP or those receiving usual care (see Figure 14) (OR = 0.83, 95% CI 0.60 to 1.15).

FIGURE 13.

Forest plot showing the results of meta-analysis of the effect of ERP vs. usual treatment on the odds of readmission following colorectal surgery. Lap, laparoscopic surgery group; Open, open surgery group.

FIGURE 14.

Forest plot showing the results of meta-analysis of the effect of ERP vs. usual treatment on odds of complications following colorectal surgery. Lap, laparoscopic surgery group; Open, open surgery group.

Effectiveness of ERP interventions at improving patient-reported outcomes

Report Supplementary Material 4, Table 3 displays data for patient-reported outcomes for RCTs trialling ERP interventions to improve recovery from elective colorectal surgery in older adults. Although there were five studies collecting PROMs,65,82,84,85,88 actual outcome values were frequently unreported. It was possible to compare pain scores in two studies conducted by the same authors,84,85 with pooled data showing that pain 1 week after surgery was similar in the ERP and usual care groups (d = −0.03, 95% CI −0.31 to 0.2; p > 0.05). There were no other opportunities to combine PROMs of the same category between studies due to the absence of data provided by authors.

Outcomes that were not eligible for meta-analysis were as follows. Mean patient satisfaction scores were slightly higher in the ERP group in the study by Frontera and colleagues. 82 Khoo and colleagues45 reported that only 9% of patients randomised to the ERP group felt they would benefit from a longer inpatient stay, compared with 69% in the usual care group. Similar quality of life scores were reported between groups in the study by Pappalardo and colleagues. 83 Quality of life was similar at 10 and 30 days post-surgery in the groups in the study by Forsmo and colleagues,88 and in the study by Lee et al. 84 there was no difference in quality of life scores at 1 and 4 weeks post surgery when comparing ERP and usual-care pathways. Vlug and colleagues reported that patients fulfilled the discharge criterion of ‘accepting discharge’ sooner when they received ERP instead of usual care; this was the case in patients undergoing both laparoscopic (d = 0.56, 95% CI 0.28 to 0.84; p < 0.001) and open (d = 0.40, 95% CI 0.11 to 0.68; p = 0.003) surgery. 65

Effectiveness of ERP interventions at improving service utilisation after discharge

Khoo and colleagues reported the numbers of patients in each group who called the ward for advice or spoke to their GP and either received a prescription or advice, after discharge. 45 There were no statistically significant differences in the odds of these events taking place, despite a greater number of GP contacts in the usual care group (10 vs. 8 contacts).

Effectiveness of Prehab interventions at improving clinical outcomes

Five studies evaluated Prehab interventions, three compared to usual care54,56,67 and two compared to Rehab interventions. 55,59 Report Supplementary Material 4, Table 3 displays clinical outcome data reported in all studies. Meta-analysis was performed for two comparisons: Prehab versus TAU54,56,67 (see Figure 15 upper) and Prehab versus Rehab55,59 (see Figure 15 lower). The comparison of Prehab with TAU showed that LOS was similar between conditions (d = 0.20, 95% CI −0.06 to 0.47; p > 0.05). The pooled estimate showed no difference in LOS between patients receiving Prehab and Rehab interventions in the two studies exploring this comparison (d = −0.11, 95% CI −0.40 to 0.18; p > 0.05). 55,59

FIGURE 15.

Forest plot showing the results of meta-analysis of the effect of Prehab vs. usual treatment (top analysis) and Prehab vs. Rehab (bottom analysis) on LOS following colorectal surgery. TAU, treatment as usual.

The odds of experiencing a complication was similar in Prehab and usual-care groups in the three studies exploring this association (OR = 1.33, 95% CI 0.70 to 2.53; see Figure 16). 54,55,67 Carli and colleagues55 reported that the comprehensive complication index which accounts for the number and severity of complications was similar between Prehab and Rehab groups (p > 0.05). Finally, while Gillis and colleagues saw more complications in the control group (Rehab) than the intervention group (Prehab) this difference was not statistically significant.

FIGURE 16.

Forest plot showing the results of meta-analysis of the effect of Prehab vs. usual treatment on odds of complications following colorectal surgery.

Despite a trend towards lower rates of readmissions in the Prehab groups, meta-analysis of the two studies comparing Prehab to Rehab showed there was no overall difference in the odds of experiencing a readmission to hospital following colorectal surgery (OR = 0.57, 95% CI 0.25 to 1.27; see Figure 17).

FIGURE 17.

Forest plot showing the results of meta-analysis of the effect of Prehab vs. Rehab on the odds of readmissions following colorectal surgery.

Effectiveness of Prehab interventions at improving patient-reported outcomes

Effectiveness of Prehab versus Rehab

It was possible to pool some outcomes reported in the two studies comparing patients randomised to a Prehab intervention with those randomised to Rehab. 55,59 Both studies used the SF-36, HADS and CHAMPS outcome measures to assess aspects of mental health, quality of life, social function, physical function and physical activity and included data at 8 weeks post surgery. 55,59 All results for individual outcomes are available in Report Supplementary Material 4, Table 5. Meta-analysis of outcomes at 8 weeks post-surgery indicates that both anxiety (d = −0.07, 95% CI −0.40 to 0.26; p > 0.05) and depression (d = −0.11, 95% CI −0.44 to 0.21; p > 0.05) on the HADS questionnaire were no different between Prehab and Rehab groups.

Meta-analysis of outcomes from the SF-36 at 8 weeks post surgery showed that Prehab was associated with an improvement in the Mental Component summary score when compared with Rehab (d = 0.38, 95% CI 0.05 to 0.71; p = 0.022). Results from the two studies were very similar, with no statistic heterogeneity (I² = 0.0%). There was no difference in the Physical Component summary score between Prehab and Rehab groups (d = 0.08, 95% CI −0.25 to 0.41; p > 0.05).

The CHAMPS questionnaire was used in both studies to provide an indication of physical activity levels following surgery. Data were reported at 4 weeks post op by Carli et al. 55 and 8 weeks post surgery by Gillis and colleagues. 59 When pooling data from these studies, there was no overall difference in physical activity levels between those receiving Prehab versus those receiving Rehab (d = 0.07, 95% CI −0.25 to 0.40; p > 0.05).

There were no further opportunities to meta-analyse outcomes in the three studies comparing Prehab with Rehab after colorectal surgery. Additional outcomes were explored relating to additional domains of the SF-36 by Gillis and colleagues, with no differences between groups (see Report Supplementary Material 4, Table 5.)

Summary of evidence from trials seeking to improve recovery from colorectal surgery

There were 14 comparisons from 12 studies evaluating interventions to improve the recovery of older adults undergoing elective colorectal surgery. Interventions were either ERP (n = 7 studies) or Prehab interventions (n = 5) compared to either usual care or, in the case of two Prehab studies, Rehab. There were a number of examples of studies providing incomplete data or offering medians and ranges, which are not recommended to be used to impute means and standard deviations, and thus there were limited opportunities to pool data from these studies. 27

Where meta-analysis was possible, ERP studies were associated with a small reduction in LOS, without any influence on readmissions, and no effect on complications – despite a trend towards reduced complications. In terms of patient-reported outcomes, there was very poor reporting of outcomes that were purported to have been captured. There was also a heterogeneous array of outcomes across the available studies comparing ERP with TAU, such that only two studies measuring pain could be pooled. These showed no difference in pain between ERP and TAU groups. Despite this, a handful of analyses of PROMs in individual studies suggested that ERP was either associated with improved outcomes or was no worse than TAU. Only one study examined service utilisation by patients after discharge, finding no difference between groups.

When exploring the effect of Prehab on clinical outcomes, meta-analysis showed no effect on LOS, readmissions or complications compared to TAU or Rehab. The studies comparing Prehab with TAU offered few opportunities for meta-analysis. It was possible to pool data for markers of physical activity but no difference was observed between groups. One study reported an improvement in anxiety with Prehab, with another reporting increased fatigue following Prehab.

Three similar studies evaluated Prehab against Rehab, using similar outcomes and thus facilitating meta-analysis. While the anxiety and depression subscales of the HADS did not differ between groups in meta-analysis, there was evidence of an improvement in the Mental Component Summary of the SF-36 with Prehab. The Physical Component Summary score did not differ between groups, however, nor did physical activity levels.

Overall, there were limited opportunities to perform meta-analysis, and usually this was with a low number of combined studies. The available evidence suggests that ERP may be associated with shorter LOS without detriment to other clinical outcomes, and may lead to some improvement in the patient experience. Prehab interventions did not have an influence on clinical outcomes but may have some positive impact on patient mental health. However, evidence is very limited and should be interpreted with caution.

Evidence regarding abdominal surgery

Report Supplementary Material 4, Table 1 displays clinical outcome data from the two studies evaluating ERP interventions in older adults undergoing elective abdominal surgery, both comparing effectiveness versus usual care. 81,87 After imputation, the standardised mean differences for LOS for these two studies were available for comparison. Meta-analysis of these two studies showed a large reduction in LOS with ERP interventions, when compared with usual care (d = −0.87, 95% CI −1.29 to −0.46; p < 0.05). Heterogeneity was not statistically significant for this effect (I² = 36.6%, p > 0.05), indicating consistency between the two studies.

Other clinical outcomes that were not eligible for meta-analysis for these two studies were as follows. Kapritsou and colleagues reported a 61% reduction in risk of experiencing a complication in the ERP group compared to TAU, but this difference was not statistically significant (OR = 0.39, 95% CI 0.14 to 1.04; p > 0.05). 81 Tagaki and colleagues saw a greater number of readmissions in the usual-care group although this difference was not statistically significant, and no deaths were reported in either group. 87

Although both studies described collecting PROMs, only Tagaki and colleagues provided data in a format allowing the calculation of an effect size. The authors reported a moderate improvement in quality of life for patients in the ERP group compared to those receiving usual care (d = 0.52, 95% CI 0.06 to 0.98; p = 0.03); however, wide CIs suggest there is uncertainty in the true effect. Kapritsou and colleagues reported decreased postoperative pain in the ERP groups, and no differences in stress or emotional response between groups. 81

Evidence regarding pelvic surgery

A single RCT was available which evaluated an intervention to improve recovery from pelvic surgery, and which collected data from PROMs. 58 This pilot study compared ERP with usual care, but reported results as medians with ranges, therefore it was not possible to calculate effect sizes for between-group differences. The authors reported that LOS was shorter and post-operative pain was lower with ERP. 58 Readmissions, complications, quality of life and satisfaction were reported to be similar between groups. 58

Evidence regarding thoracic surgery

One RCT compared Prehab with Rehab in patients undergoing thoracic surgery. 57 The authors found that LOS was similar between groups; however, there were improvements in the physical component (d = 0.61, 95% CI 0.19 to 1.02; p < 0.01), mental component (d = 0.43, 95% CI 0.02 to 0.84; p = 0.04) and total score (d = 0.51, 95% CI 0.1 to 0.92; p = 0.02) of the SF-36 when patients received Prehab. 57 There were no differences between groups when mental health, disease-specific quality of life and energy expenditure were assessed. 57

Evidence regarding upper abdominal surgery

Three RCTs evaluated interventions to improve recovery following upper abdominal surgery. 38,44,80 Two studies compared ERP with treatment as usual,44,80 and one evaluated Prehab compared to usual care. 38 Meta-analysis of the two studies evaluating ERP interventions showed that ERP was associated with a large reduction in LOS (d = −1.69, 95% CI −2.32 to −1.06, p < 0.001; see Figure 19A), while the odds of patients being readmitted following surgery was similar in both groups (OR = 0.61, 95% CI 0.27 to 1.38, p > 0.05; see Figure 19B). In the study by Jones and colleagues, there were two readmissions in the intervention group, none in the usual care group, and with one death in each group. 44 Quality of life was similar between groups in the study by Jones and colleagues, although no variance data were provided so we could not calculate an effect size. 44 Pain, sadness, stress and optimism were similar between groups in the study by Kapritsou and colleagues. 80

FIGURE 19.

Forest plot showing the results of meta-analysis of the effect of ERP vs. TAU on LOS (A) and odds of readmission (B) after upper abdominal surgery.

Complications

In the study comparing Prehab with usual care, both groups achieved similar scores for LOS, complications, readmissions, physical health, mental health and quality of life. 38

Search strategy

For the systematic review of qualitative studies, the same search terms for the quantitative search were used but we replaced the effectiveness study type filter with a qualitative study type filter. The qualitative study filter was adapted from the ‘Best Optimisation of Sensitivity and Specificity’ qualitative filter developed by Wong and colleagues, with adaptations to include additional qualitative terminology, specifically, indexing terms (e.g. Medical Subject Heading in MEDLINE) which were unavailable when Wong and colleagues developed the filter, and the free-text terms ‘focus group’, ‘perspective’, ‘perception’, ‘themes’ and ‘thematic’. 22

The search was developed by SB in conjunction with the review team and stakeholders using MEDLINE (via Ovid) and adapted for use in other databases. The selection of bibliographic databases was tailored for the identification of qualitative studies, noting that surveys and guidance recommend using MEDLINE and CINAHL,92,93 and that Embase is considered less useful. 92 The full set of bibliographic databases searched included: MEDLINE ALL and HMIC (both via Ovid), AMED and CINAHL (both via EBSCO) and ProQuest Dissertations and Theses (via ProQuest). Searches were run in June 2021. As this review was not attempting to derive a precise measure of effectiveness based on all the relevant evidence (i.e. an aggregative synthesis), updated searches were not run. Due to the high quantity of data identified through initial searches, the identification of further studies was unlikely to significantly alter the main messages derived from the synthesis. The synthesis was based on a configurative approach, which is more interpretive and based on an appropriate sample. Bibliographic database search results were exported to EndNote 20 (Clarivate Analytics, Philadelphia, PA, USA) and de-duplicated using manual checking and the EndNote de-duplication tool.

Reference lists of all included studies were checked and forward citation searching was completed for all included studies (DC, DK) using the Science Citation Index (Web of Science, Clarivate Analytics) and Scopus (Elsevier). The results of forward citation searches were exported to EndNote 20, and reference-list checking was conducted using Microsoft Excel spreadsheets to document potentially useful studies thus identified. We further extended the supplementary searches to include checking the reference lists of topically similar systematic reviews identified by the searches. Evidence suggests this to be a potentially fruitful approach due to the imprecise use of terminology and poor-quality indexing of qualitative studies in bibliographic databases. Due to the high number of studies identified through bibliographic database searches and supplementary searches, we did not search Google Scholar or relevant websites, which represented a deviation from the protocol.

Inclusion criteria

We included papers reporting primary qualitative research exploring experiences of, or attitudes towards, multicomponent interventions which aim to enhance recovery and/or reduce length of hospital stay of older adults following admission for a planned procedure. This included the views of patients, family, carers or health/social care staff. Further details of the definitions used to describe eligible populations and interventions can be found in Chapter 2.

Study selection

After completing a calibration exercise on a sample of articles (n = 100), the inclusion criteria were applied to the title and abstract of each article independently by two reviewers (DK, DC), with disagreements resolved through discussion or referral to a third reviewer as needed. Full texts were screened in the same way (DK, DC). The screening was supported by EndNote 20 (Clarivate Analytics, Philadelphia, PA, USA).

Familiarisation

All included studies were read and re-read so that reviewers could familiarise themselves with their content (DK, DC, LS). This process began during full-text screening and continued throughout data extraction and synthesis.

Determining how studies relate to one another

Summary descriptive data from all studies were tabulated, enabling comparisons to be made across included studies in terms of patient characteristics, who is being interviewed, intervention received, type of qualitative analysis, and quantity and quality of qualitative data available for synthesis. The summary of themes also allowed for early exploration of similarities across studies.

Reciprocal translation

First- and second-order construct data were extracted from the studies with a high quantity of interpretative data into a Microsoft Word document by one reviewer (DK), who developed a list of descriptive ideas and concepts seen within each study. These ideas were discussed and checked by other members of the review team (LS, DC). The lead reviewer (DK) then used this list to form a descriptive coding framework within NVIVO, which was used to conduct line-by-line coding of papers with high quantity of relevant data. The process of coding this subset of studies was divided between three reviewers (DK, LS, DC). The coding conducted in a sample of 12 studies was checked for consistency by a second reviewer (DK).

A process of purposive sampling was then conducted for the articles not yet entered into the synthesis. 99 This process was achieved through discussion amongst members of the review team (DK, DC, LS) and sought articles which provided experiences of:

• interventions or patient groups already represented in the quantitative systematic review but not yet included in the qualitative synthesis

• types of interventions or reasons for admission not yet seen

• participants groups whose views not yet seen

• themes, experiences, or ideas not yet seen.

The papers identified through purposive sampling were then coded using the existing coding framework by one reviewer (DK). The stages at which different papers were entered into the coding framework is summarised below in Table 4.

Similar concepts and ideas were merged in an iterative process to form themes, with existing coding being checked where this process yielded new interpretations or ideas (DK). Conceptually similar themes were then grouped together to form overarching constructs (DK). Throughout this process, data which refuted or challenged developing themes were explored and considered within the evolving synthesis. The synthesis was discussed with members of the review team (MN, LS, DC, SF) throughout its development.

Creating a line of argument

A line-of-argument is a model or theory which aims to summarise the relationship between the overarching constructs and contributing themes generated through the synthesis and the concepts contained within these. Here it is intended to capture the experiences of patients, their families or carers and staff of interventions intended to reduce LOS. The line-of-argument was developed by one reviewer (DK) and refined through discussion with the review team, clinical stakeholders and the patient and public involvement (PPI) group.

The stages of meta-ethnography outlined above, whilst conducted in a linear fashion, overlapped with one another in an iterative process. For further information regarding how stakeholder and patient/public involvement influenced this process, see Chapter 5.

Study overview: all included studies

Sixteen of the included studies were conducted within the UK,90,100,101,107,109,116,117,119,120,125,128,130,133–135,137 eight studies (nine papers) were conducted in Denmark,103–105,123,124,126,132,136,141 five studies were conducted in Australia,112,113,118,121,139 three studies in Norway,114,138,140 two studies were conducted in Canada,107,129 two in Sweden,102,122 two in the Netherlands,127,131 two in the USA,108,111 with one study being conducted in Italy,110 and one in Finland. 115

Enhanced recovery protocols (ERPs) were the most common intervention that participants were asked for their views or experiences on (n = 29 studies). 90,100,102,103,105–110,113–116,118,120,122–128,130,132,133,136,139–141 Other interventions for which views were sought included Rehab programmes (n = 4),101,129,135,137 supported discharge (n = 2)131,138 and early discharge (n = 2). 111,121 The following interventions were also each evaluated by one study: CGA,117 case management within a fast-track pathway,104 Prehab119 and staff mix,112 with one study not clearly defining their intervention. 134

Participants providing their views included patients (n = 31 studies),90,100,102,108–110,113–116,119–128,130–141 staff (n = 5),103,112,117,118,129 patients and staff (n = 2),101,107 carers (n = 3)104,105,111 and patients and carers (n = 1). 106 Interviews were the most common method of data collection (n = 37 studies);90,100–111,113,115–126,128,130–133,135–141 interviews were accompanied by observation in two studies. 105,117 Four studies collected data using focus groups112,114,127,134 and one study used interviews and focus groups. 129 Further summary details including methods of analysis, quantity of relevant data and summary of main themes for each of the included studies are provided in Appendix 3.

Below is a summary of the key characteristics of the studies prioritised for synthesis using meta-ethnography.

Participant characteristics

Thirty-four studies (35 papers) were prioritised for synthesis, including studies from the UK (n = 13),90,100,101,107,109,116,117,119,120,125,128,130,133 Denmark (n = 6),103–105,123,124,126,132 Australia (n = 3),112,113,121 the Netherlands (n = 2),127,131 Canada (n = 2),106,129 the USA (n = 2),108,111 Sweden (n = 2),102,122 Finland (n = 1),115 Italy (n = 1)110 and Norway (n = 1). 114 We prioritised all of the studies representing the views and experiences of staff (n = 7) and carers (n = 4) for synthesis using meta-ethnography. 96–98 We also prioritised 25 studies which represented patient experiences. 90,100–102,106–110,113–116,119–128,130,132,133 The total number of individuals who participated ranged from 5113 to 41. 117 The most common reason for patient admission was total hip arthroplasty (THA) and/or total knee arthroplasty (TKA); nine studies included patients admitted for hip and/or knee replacement,102,103,105,112,115,116,123–125,127,129 six studies included patients admitted for THA,90,104,106,113,119,121 and two studies included patients admitted for unilateral knee arthroplasty (UKA). 101,114 Seven studies focused on patients admitted for colorectal surgery,118,122,126,130–133 with one study representing patients admitted for the following reasons: gynaecological cancer surgery,100 hysterectomy,120 lung cancer surgery,107 saracolpopexy,108 bowel surgery,109 pancreatic surgery,110 coronary artery bypass graft (CABG) surgery,111 cancer surgery117 and liver resection. 128

Further details regarding the sample included in the prioritised studies, including mean age, percentage of female participants, ethnicity and socio-economic status and inclusion criteria, can be found in Appendix 3.

Intervention characteristics

Of the 34 studies (35 papers) prioritised for synthesis, the most common interventions participants received were fast-track, ERP or ERAS pathway (n = 26 studies),90,100,102,103,105–110,113–116,118,120–128,130,132,133 or rehabilitation (n = 2). 101,129 Single studies sought views on the following interventions: case management,104 CGA,117 early discharge,111 a perioperative care and ehealth programme,131 prehabilitation119 (Orpen) and staff mix. 112

Eight of these interventions were delivered alongside a comparator comprising usual care,90,101,104,106,108,128,131,132 key features of which included standard post-operative physiotherapy (n = 1),101 treatment as usual, case management and information (n = 1),104 usual care with placebo website (n = 1)131 and next day discharge (n = 2). 106,108 Two studies sought the views of participants within the treatment as usual condition. 106,133

Overall, interventions were poorly described, with 11 studies providing no, or minimal, description of the intervention received by participants. 100,102,106,109,111,113,122,125–127,129 Two studies provided details regarding flexibility of delivery of the intervention,101,121 four studies made reference to staff receiving training to deliver the interventions,101,104,130,133 and four studies made clear reference to the use of an intervention manual. 101,104,130,131 Information on adjuvant treatment was provided by four studies. 107,108,131,132

Intervention recipients included patients alone (n = 20),90,100,101,103,108,110,112,115–121,128–133 patients and carers (n = 5),105–107,114,123,124 or family/carers alone (n = 1). 104 Eight studies did not report this information. 102,109,111,113,122,125–127

Interventions were predominantly received in hospital (n = 20),90,103–108,110–112,114,117,118,121,128–133 including one intervention delivered via a website/mobile phone application,131 and one delivered in a pre-op hospital ward. 121 Seven studies reported the interventions had a pre-specified LOS,90,103,106,108,113,115,123,124 three studies reported that the intervention had no pre-specified LOS,112,117,132 with the remaining studies not reporting this information (n = 24). 100–102,104,105,107,109–111,114,116,118–122,125–131,133

One intervention was delivered jointly in the hospital or community,116 and one delivered both at hospital and in the patient’s home. 123,124 Ten studies did not report the location of intervention delivery,100,102,109,113,115,120,122,125,126 but it is reasonable to assume this was the hospital for eight of these. 100,102,109,113,115,120,125,127 Two interventions were delivered in the patient’s home. 101,119 Intervention characteristics for each study are reported in Appendix 3.

The majority of patients were discharged home, although this was not clearly stated in eight studies. 102,108,113,114,121,123,124,126,128 Alternative discharge destinations included temporary rehabilitation centre prior to home (n = 1),111 home or inpatient rehab (n = 1),112 and home or temporary nursing home (n = 1). 127 One study did not report this information. 129

Interviews were the most common method of data collection (n = 29). 90,100–111,113,115–126,128,130–133 Interviews were accompanied by observation in two studies. 105,117 Three studies collected data using focus groups,112,114,127 and one study used individual interviews and focus groups. 129 Quantitative data were collected alongside qualitative data collection in seven studies. 101,108,115,116,127,131,132 Further detail regarding the aims of each study, types of analysis used and recruitment method can be found in Appendix 3.

Construct summary: home preferred place for recovery

This construct illustrates how, for some patients, home can provide a safe and familiar environment which promotes recovery through enhancing feelings of control and increasing access to required support. However, discharge home also contributed towards feelings of vulnerability in patients, with some preferring to recover in hospital if their needs for security and support at home were not addressed. This construct links with all others developed in this analysis, so the idea of ‘home as preferred’ will be interwoven in the descriptions of other constructs.

Theme: meeting emotional and physical needs

Twenty-three studies (23 articles) highlighted that meeting patients’ (and carers’) emotional and physical needs was linked to their feelings of safety and their overall experience of the programme. 90,100,101,105,107–111,113,114,116,119,120,124–126,128,130,133

First- and second-order data from five studies (five articles) explored how meeting emotional needs not only helped patients and carers feel safe, but also supported the delivery of intervention components which aimed to promote physical recovery. 100,102,112,113,120 Phillips et al. 120 described how patients did not necessarily want to mobilise, but did so with encouragement from staff, with one patient from the study discussing how this physical activity had a psychological component:

Yeah, I think a lot of it though is mental, in your own mind, that right, I’m going to do this.

(Patient, non-cancer gynaecological surgery)120

This suggests appropriate support can increase patient confidence and overcome feelings of vulnerability, and that physical and emotional health cannot be considered in isolation from one another. This idea is explored in other themes within this construct and in the ‘Taking responsibility’ construct. The below quote from an advanced nurse practitioner highlights the importance that patients place on emotional support when giving feedback about their intervention:

Patients … who’ve been through the SOLACE project, they couldn’t say nicer things … The difference it’s made …. Not just physically but the social and psychological support that becomes part of it … I think that’s clearly been a great help to a lot of people.

(Advanced nurse practitioner, lung cancer surgery)107

First- and second-order data from four studies (four articles) suggested the extent to which patients felt their emotional needs, such as a feeling of being cared for, safety, or general wellbeing, had been met by interventions varied. 107,108,116,126 Evans et al. (2021)108 described that the ERP felt rushed to some patients and ‘wasn’t perceived as an advance in care, but as an absence of care’, which is also illustrated by the quote below from a patient from Thomsen et al.:126

There was no room for questions and one is also a bit fearful. I don’t just need general information or that the doctor gives his specific message. I also need to be asked: how are you feeling.

(Patient, colorectal cancer surgery)126

The above quote suggests that, in order to support emotional needs, the intervention needs to be paced according to the needs of the patients and give them an opportunity to ask questions. Meeting emotional needs is therefore also linked to the theme of ‘Increasing confidence and preparedness through information and planning in advance’ within this construct, and the construct of ‘Individualisation of a standardised programme’.

While some patients felt their emotional needs were met, data from three studies (three articles) suggested patients could mask negative experiences, such as feeling their needs were not met, in their initial feedback (see construct: ‘Outcomes’). 90,109,125 All the patients on an accelerated discharge pathway following THA in the study conducted by Hunt et al. (2009)90 had criticisms regarding feeling overlooked by staff or uninformed, which they also tried to justify, a problem which was echoed by patients in Fecher-Jones et al.’s study:109

I suppose really the nurses have their own lives to lead and then they often, you think they’re neglecting me you know, I wish they’d come and do something.

(Patient, THA)90

I know they were ever so busy, they did plenty medically, but little things, you know, for comfort, they didn’t.

(Patient, colonic resection surgery)109

The second quote also suggests that staff may find it difficult to identify when patients are struggling (and provide appropriate support) when their workloads are high and they do not have time to check in with them in this way.

Sixteen (16 articles) indicated how a good relationship with medical staff did make a difference in whether patients or carers felt cared for emotionally as well as physically. 90,100–102,104,105,107,109,110,113,114,116,120,124,125,133 At times this was merely knowing who it was that would be medically caring for them, suggesting some basic familiarity with staff could contribute to feeling safe. This is illustrated by the below quote from a LLA patient:

When the doctors gave information I became more confident, just to put faces to some of the staff ....

(Patient, LLA)124

Alternatively, data from four studies (four articles) suggest patients and/or carers found it beneficial having a specific staff member they could build a strong relationship with, like a co-ordinator. 101,104,107,113 This concept relates to those discussed within the theme of ‘Appropriate, available and accessible follow-up’, as patients knew who to contact when they had concerns:

And she always said that if you have any problems just call me, right? And that has been very comforting. I will admit to that. We are not spring chickens any more.

(Carer, spouse of THA patient)104

However, data from six studies suggested it was not the member of staff’s seniority or particular role in the team which was of importance, but their ability to empathise, listen and connect to the patients and/or carers. 101,104,107,113,120,133

He’s [not a physiotherapist] bless him … but he’s also had injuries himself so he knows, he knows what it’s like to have a lot of pain … he’s an extraordinarily empathetic young man.

(Patient, knee arthroplasty)101

Theme: increasing confidence and preparedness through information and planning in advance

Data from 28 studies (29 articles) support this theme. 90,100,102–108,110–116,118–120,122–126,128,129,131–133 Below, the importance of pre-op care and provision of accessible and timely information is that it helps patients and carers to feel more confident and better prepared, and aligns their expectations with medical expectations. This theme also discusses the role of consistency and format of information, and discharge planning in supporting patients and carers to feel safe.

First- and second-order construct data from 13 studies (13 articles) highlighted how clear information helped patients and/or carers to feel more confident about the approaching procedure, and could reduce anxiety, linking to the idea of feeling safe. 100,102,105,108,110–112,114,116,119,122,124,129 This is illustrated by patients describing their experiences of receiving information before sacrocolpopexy surgery:

Patient 1: They made [patient instructions] clear orally as well as in the written directions …. So, I felt very comfortable going into the surgery.

Patient 2: I did not feel tense, I did not dread, I was not afraid.

(Patients, sacrocolpopexy surgery)108

Here, one patient indicates their appreciation for repeated information in different formats. Clear information was not only about managing the patient’s feelings of safety and confidence, but also to ensure they understood why they needed to do certain tasks, which meant they may be more likely to do them. This is illustrated by one female patient who was unclear as to why certain parts of her care plan were in place:

Ummm. [clicks her tongue] You see, the drinks, I don’t quite know what they do, to you. Are they full of proteins?

(Patient, non-cancer gynaecological surgery)120

The above quote indicates how lack of understanding could prevent patients from fully engaging with their recovery plan and being ‘active’ in their recovery (see theme; ‘Active patient’). In contrast, the patient in the below quote had been informed clearly about early mobilisation, and was prepared to get up and walk:

The fact that I knew what I was going to have to do when I came round. I knew I was going to have to get up … I knew that I had to get up and walk. The preparation is good.

(Patient, gynaecological cancer surgery)100

Connected to being an ‘active patient’, shared decision-making between staff and patients was difficult for patients without access to consistent and adequate information. 116,123,125 For example, this patient in Strickland et al.’s 125 study did not have enough information on what options were available, and subsequently was unable to make informed choices:

I was asking for something I couldn’t have.

(Patient, LLA)125

Being informed could also help with worries about whether a particular symptom was ‘normal’ when recovering at home, which will be discussed further in the theme: ‘Appropriate, available and accessible follow-up’.

Overall, data from four studies (four articles) suggest that the level of information provided impacted confidence or feelings of safety post-discharge. 102,114,116,122 For example, whilst Berg et al. (2019)102 described most patients as accepting of discharge due to having clear information during the pre-operative phase, Samuelsson et al. (2018)122 reported that some patients had concerns about the post-discharge phase as they felt ill-prepared.

Patients and carers from 10 studies highlighted how missing information was a particular issue for the post-discharge phase, and linked with the availability of follow-up. 90,102,104–106,111,116,126,132,133 One caregiver discussed the impact a lack of appropriate information can have on patients ability to manage their pain post discharge:

The meds, too … no one told us to make sure we take [them] …and he got really in hot water [not taking enough], so the pain came back full force, and it took longer to get it down again.

(Carer for THA patient)106

The work of carers supporting the patient at home was made more difficult when they could not get access to the information they needed to perform their home caring role, which was also echoed by our PPI group. For example, one carer in Berthelsen et al.’s (2017)104 study was described as having difficulties changing the patient’s dressing at home as she was not given enough information prior to discharge. Patients also felt this concern for those caring for them, as illustrated by the patient who had undergone hip replacement in the quote below from Hunt et al.’s (2009) study:

If you could spend a half day with us and this is what we’re going to show you what you’ll be able to … what you’ll have to do after … at least the carer would know exactly what’s going to happen.

[Patient, THR]90

Six studies suggested inconsistency of information between different staff members impacted patient confidence:105,113,116,124,126,133

… whether or not your pillow is under your knees or not at night, whether your legs should be apart or not. … because it’s your hip and your future and your recovery, I found those sort of things a bit unnerving.

[Patient, THR]113

The same patient in Heine et al.’s (2004) study said that it did not matter if the inconsistency was regarding a more minor matter, suggesting that care must be taken to ensure all aspects of information are consistent:

It’s often the small things that unravel the most. I’m inclined to agree with that.

(Patient, THR)113

First- and second-order data from four studies (five articles) discussed how having too much information, or the right information in the wrong format, could be difficult for patients or carers to handle. 105,120,123,124,133 For example, Phillips et al. 120 discuss one gynaecological surgery patient who particularly struggled with pain and linked her pain to difficulty with recalling the amount of information she was given:

… after I had some lunch I remembered that … I was told … to start off with water and liquids and I didn’t …. So I knew at that point I was in … that I’d done something wrong …. So I started to have some pain.

(Patient, non-cancer gynaecological surgery)120

However, the amount of information desired varied between patients, linking to the idea of needing to individualise care, as illustrated by two quotes from different patients in the same study below:100,102

I’m that kind of person, so that if they hadn’t given me I would have forced them to give me … I want to be prepared for what they’re going to do … I want to know about the details. (Patient 1)

In fact I want to know as little as possible about the procedure. No, I’m not really so fond of these kinds of operations. (Patient 2)

(Patients, LLA)102

The information needed to be tailored to the individual patient or carer, with adequate time or opportunity for questions. 115,119,122–124,133 Some found written information useful as a resource to refer back to,108,113–116,119,120,122,131 whereas others valued verbal instruction. 108,116,119,133

Patients from 11 studies (12 papers) indicated the timing of the information also impacted their ability to remember it, with some pre-op information classes too distant from the procedure itself, and some post-op information given to the patient when they were still affected by analgesia. 100,106,113,114,116,119,120,122–126 This is highlighted by a patient who had a longer wait between the pre-op information clinic and her procedure, and then struggled to remember instructions for the post-operative exercises she had been given:

I can’t remember [when the hip school was] because of the delay. That’s really why I’ve forgotten some of the exercises really.

(Patient, LLA)116

Data from 12 studies (13 articles) indicated that an advance plan made pre-operatively or prior to discharge could enhance feelings of safety. 100,102,104,106,113,116,118,119,121,123,124,129,133 Firstly, prehab or a pre-op visit at home helped prepare patients for their procedure and the recovery period post-discharge. 100,102,103,116,118,119,124,129,133 This pre-operative planning and thinking through could help to identify potential issues in advance and support patients to feel more confident that they know what they practically needed to do:

I was confident about coming out of hospital because I had actually walked through in my mind at the home visit … issues that hadn’t been in my mind at all … and wouldn’t have come to mind until 3 days after the op.

(Patient, THR)119

Pre-op care was important in terms of the patient gaining realistic expectations of their recovery, which could impact experience of pain (see theme: ‘The active patient’). This is illustrated by one caregiver of a patient admitted for total hip arthroplasty, who explains how they had both under-estimated the degree of post-surgical pain, which also points to the importance of realistic expectations helping patients and carers to plan:

Yeah, [the pain] was a little bit more … than I thought, because everyone said ‘Oh, the surgery is perfect, it goes great, you will be up and at ’em in no time flat’, but it’s still surgery, so it wasn’t that ….

(Carer, THR)106

Similarly, four studies discussed how advance planning could help patients and carers and feel more prepared for discharge. 113,119,121,123 This is important given that patients (and carers) needed to feel prepared for discharge. 100,113,119,121,123 For example, Specht et al. (2018)123 describe how patient confidence in going home was linked to ‘being involved and the feeling of control in the discharge planning process’.

Patients did not feel prepared when they felt rushed or their emotional needs were not met, linking with the idea that both physical and emotional needs need to be considered:

The discharge, it went fast … it was messy … I was not sure about the pills … but they also were very busy at that time.

(Patient, LLA)123

As discussed previously, consistent and adequate information was important for patient and carer confidence about post-discharge recovery, but it was also important for confidence in the discharge process itself. The below quote from a patient in Strickland et al. ’s (2018) study highlights not only the potential confusion for patients, but also the need for co-ordination between staff members and services (discussed in more detail in the construct ‘Taking responsibility’).

The surgeons tell you, you can come out the next day, the nursing staff say you are not ready to go home and you can’t go home till the physios say … so everybody you speak to tells you a different story.

(Patient, LLA)125

In addition to the importance of patients (and carers) feeling prepared for discharge, there was also a need to consider the patient’s home circumstances when planning discharge (see construct ‘Individualisation of a standardised programme’). For example, Hovik et al. 114 described how those living alone had to prepare their homes and support mechanisms for the recovery period in advance of hospitalisation.

Seven studies highlighted how carers needed to be involved in discharge and its planning to that they could effectively support the patient at home. 90,104–106,111,119,124 As will be discussed further in the ‘Essential care at home’ construct, if carers are required to support the patient in their recovery, then they need the information and support to do so:

I wasn’t there when [the physiotherapist was] there …. When we got home, he’s in his walker, there was a bit of confusion … like, put this leg first, well, what did [the physiotherapist] say?

(Carer, LLA)106

Consistency of information was important to patients, as inconsistency could make them feel more vulnerable (which is also linked to the need for service co-ordination). The format of this information could also impact understanding and memory, and it was also important to consider the needs of family carers. Planning discharge, providing information about expectations and available services, and involving carers as relevant, helps patients and carers to feel safe and less worried about discharge.

Theme: appropriate, available and accessible follow-up

Thirty studies (30 articles) contributed to this theme, which discusses how accessible and available follow-up care helped patients and carers to feel more safe and secure, even if they did not need to use it. 90,100–109,111,113–116,118–123,125–132 The difficulties in accessing follow-up, including not knowing whom to contact for particular difficulties or how, are also explored and support content within the ‘Staff and service co-ordination’ theme. Accessible and available follow-up was one of the key issues described by members of our PPI group.

Eighteen studies highlighted that patients really valued follow-up care, and that it was reassuring. 90,100,102–104,106–108,113–116,123,125,126,128,130,131

I … even just a little follow-up on it all makes you feel a lot better.

(Patient, colorectal cancer surgery)126

Four studies (four articles)109,113,119,129 particularly highlighted that this follow-up by hospital or primary carer services was even more important when patients did not have home carers, which is also supported by the idea that having a carer at home is essential for patients (as described in greater depth in the construct ‘Essential care at home’). One patient highlighted how the death of family and friends had left her isolated:

I’ve got no-one else, no-one. My friends have died, all my friends are gone. I had some lovely friends but they’re all gone. That’s the trouble isn’t it and my brother has gone who would have helped.

(Patient, THR)119

As well as follow-up care being available to patients, nine studies indicated that patients found having follow-up care booked prior to discharge helped them to feel safe:90,100,103,104,113,116,123,125,126

I didn’t want to sort of be a nuisance … [and ring them], and I was a bit worried that what would happen just in case they were any problems … it was nice to know [nurse] was going to call on Monday.

(Patient, gynaecological cancer surgery)100

The above quote also highlights how some patients or carers did not want to bother busy healthcare staff, or did not know when a particular symptom met the threshold of needing to call so did not call when needed. 100,105,106,109,126,132 Given that some are reluctant to contact of their own accord, a proactive approach by the hospital would give patients permission to voice their concerns.

… I didn’t think that it was anything worth calling about. Now I remember the nurse telling me, don’t call if you have a swollen finger. Then I thought, is a bladder infection worth calling about? I just waited until … [outpatient appointment].

(Patient discharged following colonic cancer surgery)132

Even though they said ‘you can call us’, I didn’t, because I felt that by calling on a Tuesday morning I would disturb their work.

(Patient discharged following colorectal cancer surgery)126

However, if hospitals offer a follow-up service, they must follow-through on this, as Archer et al. 100 describe in interpretation of their findings – patients may be relying on that follow-up to discuss difficulties or ask questions, and not receiving an expected service could lead to a negative experience.

Nine studies (nine articles) highlighted patient worries about difficulties in accessing follow-up care once discharged. 101,107,108,111,115,116,118,128,130 One example of this was practical concerns such as living alone or far from the hospital:

I told them I might need to stay at least one night … because we live so far out and I needed to be [at the hospital] because I did not know if I was going to react again to this medication.

(Patient, sacrocolpopexy)108

Here the patient indicates that she was involved in the process of deciding when she should go home. Thirteen studies indicated that some of the concerns experienced by patients and carers prior to discharge were well-founded, as the availability or accessibility of follow-up care could vary, even where it was part of a programme or intervention. 90,101,107,111,114–116,118,119,122,126,129,130 For example, Hovik et al. 114 described how a patient who lived alone had difficulty booking follow-up physiotherapy despite being told this was available. The accessibility of follow-up care could be a particular problem for those who were not able to travel to a service for support, as illustrated by the quote from a surgeon in Westby et al. ’s study below:

The other thing that’s non-existent for the most part is home physical therapy for … the patient who is unable to get transportation somewhere or has social issues that would preclude them from [getting there]. Those patients fall through the cracks ….

(Staff, surgeon, LLA)129

A lack of easily accessible follow-up care was also especially difficult when dealing with complications or unexpected symptoms at home (see theme ‘The active patient’), as illustrated by a patient who had been discharged following total hip arthroplasty:

The first [episode] was that night and [the second] the next morning … it scared both of us … I did not know what to do, they said to call the hospital, but you still don’t know, because I hadn’t read anything about fainting ….

(Patient, THA)106

Seven studies (seven articles) highlighted how patients and/or carers did not always know whom or what service to contact for support after discharge, though this did vary between studies, dependent on the intervention:101,111,115,116,122,126,130

It seemed to me that once you get home you felt a bit out on a limb when you tried to get back in …. This the problem. I mean you don’t really know exactly who you’ve got to contact.

(Patient, colorectal cancer surgery)130

No regrets about getting home because you could always ring physiotherapy and if there’s any issue you know drop in.

(Patient, LLA)125

Data from eight studies (eight articles) indicated that patients and carers appreciated having access to someone to whom they could ask questions or share concerns. 90,100,102–104,107,115,125 This was particularly the case when the patients or carers had a named service co-ordinator, which will be discussed further in the theme ‘Staff and service co-ordination’. This was illustrated by a study by Berthelsen et al. ,104 who specifically examined the provision of a case manager, and found carers highlighted the opportunity to telephone the manager to talk or ask questions as the most important element of the programme.

First- and second-order construct data from five studies (five articles) discussed how some patients also found it harder to measure their progress or self-motivate without some guidance through follow-up, and appreciated explicit goals or guidance which they could measure themselves against:102,113,116,123,131

Sometimes when I’m depressed, I think that there’s something wrong with me … perhaps one needs to have small goals … so that you see that things are going in the right direction, am I too slow or too fast ….

(Patient, LLA)102

But that [activity tracker] does stimulate you at the end of the day, to see where I am and oh tomorrow I have to do a bit more.

(Patient, colorectal cancer surgery)131

Construct summary: feeling safe

This construct explores the different factors which support patients and carers to feel safe throughout the hospital stay and following hospital discharge. The needs of both patients and carers need to be viewed holistically, with pre-op assessments, discharge planning and provision of information individually tailored to consider their emotional and physical needs. This can help increase patient and carer confidence by allowing them to put a post-discharge plan in place, reassuring them they have the skills and support they need to cope once they leave hospital. Follow-up care was essential for both patients and carers after hospital discharge. Both patients and carers found the opportunity to ask questions reassuring and valued being asked how they were feeling. Follow-up care also allowed medical concerns to be identified and/or addressed. Feeling safe may promote patient recovery through increasing patient and carer confidence at returning home and enabling carers to provide the support required for patients to recover at home, thus reducing the stress associated with this transition.

Theme: being inside or outside the programme

Twenty studies (20 articles) explored how patients whose needs were met within the programme and could meet the standardised timeframes can have very different experiences and perceptions to those who are unable to, for example due to developing complications or lacking support at home. 90,102,103,107,108,110,114–118,121,122,128–131,133 This theme also explores issues related to comorbidities and complications, and weekend care.

Second-order data from two studies describe their respective patient participants as having contrasting experiences due to their individual circumstances. 108,110 Galli et al. (2015) highlighted that patients following a protocolised programme perceived they were ‘inside the right path’ whilst those who were unable to participate fully in an intervention due to more complex needs saw themselves as ‘outside’ the programme. 110

For some patients and carers, the nature of the programme being structured could be comforting or reassuring. For example, Evans et al. 108 noted that patients reported the ERP gave them a ‘sense of continuous care’ as it covered their full ‘journey’ from pre-op to post-op recovery at home, and that this made their experience more personalised, with one of their participants stating:

… you know, it just feels like, more like, there is care out there ….

(Patient, sacrocolpopexy)108

However, not all patients had this experience. Some patients were ‘outside’ the programme,110 in that they could not follow the protocol or pathway. First- and second-order data from nine studies suggested that comorbidities and/or complications were the most common reason for patients being ‘outside’ the programme. 103,108,110,117,118,121,122,128,130

Patients could feel discouraged or worried if they were not meeting the standardised milestones, or saw other patients meeting them more easily. 103,110,116,128 One patient discharged following liver resection surgery described her anxiety about not meeting expected milestones in her recovery:

The only problem with that (milestone) of course is if that doesn’t happen, because then you can get anxious because you’re thinking … if I didn’t match up to it I could feel myself thinking, oh maybe there’s something wrong.

(Patient, liver-resection surgery)128

Whether a patient experienced complications impacted whether they wanted a longer hospital LOS. Evans et al. 108 interviewed patients who had been discharged on the same day as their procedure, and those who had chosen to stay overnight, and found that those who felt more confident about their health were happier to be discharged sooner:

I was ready to come home … I knew my blood sugar had come back down to a normal level, so I was fine and felt very comfortable.

(Patient, sacrocolpopexy)108

I was really very thankful … they let me stay … I know part of the reason for that was they wanted to watch my creatinine level. But … they [also] had to treat [my blood sugar] with insulin … if I was home, I would not have … known.

(Patient, sacrocolpopexy)108

The above quote also indicates how having comorbidities could complicate the home recovery process, making it more challenging for home carers, or making patients feel more vulnerable. One patient described, in the below quote, how having comorbidities made the home recovery process more challenging. She had also discussed her partner (carer) having health issues which meant he was unable to support her as needed.

It was difficult the first few weeks … you felt you had nobody to help you, you were just on your own … in my case having [spinal-stenosis] as well … for three nights I couldn’t … get in and out of bed, just sat in the chair.

(Patient, THR)121

The above quote also suggests that the need for home support is vital (further in the theme ‘Essential care at home’), particularly given that this patient group may be more likely to have comorbidities, complications, or frailty. 6,9–16

First- and second-order data from three studies (three papers) suggested that having a procedure, or being discharged, around the weekend was another key reason for being ‘outside’ the programme, due to less availability of services both within the hospital and in the community post-discharge at the weekend. 111,112,118 Although only represented within a few studies, our PPI group highlighted it as a key difficulty.

Two of the studies112,118 highlighted that when patients had an operation on a Friday, their discharge could be delayed not because of complications, but simply because the required teams were not working over the weekend so the next step of their pathway could not be completed. Staff members from one of these studies112 discussed lower staffing levels over the weekend meaning they had to spend less time with each patient, prioritising those who would be sooner discharged:

So … they missed two full days of seeing a stoma therapy nurse and then on Monday … they should be getting out, whereas they haven’t even met the stoma therapy nurse.

(Staff, colorectal surgery)118

On weekends, we have less staffing for the same number of patients. The exact quantity is slightly different for different professions. And the way we normally prioritise is dependent on whether they (the patients) are going to go home over the weekend.

(Staff, occupational therapist, LLA)112

Patients and carers could struggle to access or arrange timely follow-up care when discharged on a Friday or at the weekend. Planned discharge on a Monday could also be challenging, as patients and carers could not contact community services on a weekend to make arrangements. The following quote from a carer illustrates the difficulty of trying to make arrangements with post-discharge services at the weekend:

… 3 o’clock on Friday that I get this bombshell of ‘find someplace’ … they wanted her out on Monday … so I made some telephone calls and went on Saturday … of course nothing could be done until Monday because the people had the weekend off.

(Carer of patient discharged following CABG)111

As suggested in the availability of follow-up theme, there could be variation in the availability of community services, regardless of when the patient was discharged (supported by 13 studies). 90,101,107,111,114–116,118,119,122,126,129,130 Where a patient did not have access to, for example, community physiotherapy which was part of the standardised plan, it could delay discharge or make the recovery process at home more challenging. This was discussed by staff in Lyon et al.,118 who cared for patients admitted for colorectal surgery within an ERP programme, and commented on how it was difficult to arrange follow-up care for patients in more rural parts of Australia and that this could delay discharge.

Data from nine studies (nine papers) suggested that when patients did not fit the standardised programme, there was variation in whether patients felt staff met their individual needs, even within the same study/programme. 108,110,114–116,121,122,130,131 For example, Judge et al. 116 described how some patients who could not mobilise quickly felt that staff listened to their worries, but others felt they were not listened to and instead were pressured to ‘conform’ to the prescribed milestones of the pathway.

Similarly, staff in four studies discussed struggling to fit all patients into the prescribed programme or meet discharge targets:103,117,118,129

It depends on their age and what other comorbidities they have.

(Staff member, colorectal surgery)118

Three studies suggested that busy staff workloads could prevent individualisation of care, as they did not have the time for longer discussion or additional care tasks. 103,122,129 For example, Berthelsen et al. 103 described how nurses reported that the only obstacles for them in providing individualised care such as additional discussion with the patients was ‘a lack of time and a busy daily schedule’. This was noticed by patients, who could then avoid requesting this kind of care, such as the below patient in Samuelsson et al. ’s study:122

You can see that the staff has a lot to do, so you feel reticent even though you need to ask a question.

(Patient, colorectal cancer surgery)122

In one study,103 staff emotionally resolved this difficulty by placing the blame on the patients themselves for being unprepared, weak, or problematic (discussed in theme ‘Differing priorities’):

Those ‘I live alone’ (voice quivering) … well you have known about this for a long time. Don’t you tell me … you haven’t put food in your freezer and you haven’t talked to your neighbour and family? Don’t you have any friends?

(Nurse, LLA)103

The above quote shows not only that some staff can struggle to empathise with the feeling of vulnerability that can come with living alone, but also that patients may benefit from support to think about the resources that they do have available to them.

One study101 suggested that a pre-op visit in the patient’s own home could help to individualise care, as staff could better understand and empathise with each patient’s context and specific needs and adjust plans accordingly:

In a … sterile clinic or environment … it doesn’t bear that much resemblance to somebody’s house … I think seeing people in their own home … being able to relate more to what the patient’s saying ’cause you can see it, it’s not just a theoretical problem ….

(Staff, physiotherapist, knee arthroplasty)101

Another study119 suggested that individualised home care could help patients to prepare for their procedure and the recovery:

it’s difficult … to translate anything that’s said [in hospital] into your home environment … I wouldn’t be able to imagine everything that I would need at home while I was sitting in hospital with somebody just talking about it.

(Patient, THR)119

However, individualising care in this way takes time and resources, which, as described above, may not be available within busy staff and service workloads.

Theme: differing priorities

Six studies (six papers) contributed to this theme, which describes how organisational, staff and patient priorities were not always aligned, and this could cause difficulties in managing individualisation of patient-care or resources. 103,112,117,119,121,129

First- and second-order data from six studies indicated the overall priorities of the organisation or healthcare systems could be mis-aligned with what staff working with patients wanted to prioritise. 103,112,117,119,121,129 Rapid discharge was cited by staff as an organisational priority which meant that their time and activities had to be directed towards this, rather than other priorities which may support other kinds of patient care:

From my perspective, allied health services are more directed at discharge planning to allow patient flow. Resources currently are more focused toward getting people home than providing a rehabilitation service.

(Staff, physiotherapist, LLA)112

As discussed in the theme ‘Being inside or outside the programme’, staff sometimes struggled to fit all patients within the programme. When LOS was the primary goal, it was not always possible to individualise care where this would increase time spent in hospital, or time spent with the patient if there were heavy workloads. This meant, for some staff, sticking as rigidly as possible to the structured plan and patients occasionally being discharged too early, and for others trying to weigh up the consequences of missing targets against the patient’s individual needs. The following two quotes from staff members highlight some of the difficulties in trying to balance these competing priorities:

He was in worse shape than we thought. He … needed to function better than he did when he was discharged … So the individual considerations can sometimes disappear.

(Staff, nurse, LLA)103

it’s weighing up those completely incongruous goals … Bugger the breach targets, but spend the time and get the patient ready, so their operation recovery are better. Or do we do it as quick as we can … hit the targets, bugger the patient?

(Site lead, CGA implementation)117

Staff from the same study as the latter quote indicated that organisational priorities could prevent innovation that did not serve these priorities, were aimed at a minority patient group who were not well-served by the existing programme, or due to time pressures. It could also mean that staff had to carefully consider how to best use resources when they were constrained by healthcare systems or insurance companies:

So you play this game with the insurance company and you get caught in the middle of the game as a patient …. One of the biggest changes we’ve seen is with rehab.

(Staff, surgeon, LLA)129

Although services in the UK do not have to contend with insurance companies, resource allocation is an increasingly pertinent one in the NHS, so staff and services must consider to whom and what resources to allocate, and how, such as through limits on number of physiotherapy sessions. This was an issue mentioned by the authors of one included UK-based study. 119

Construct summary: individualisation of a standardised programme

This construct highlights the different experiences of patients who can follow a protocolised programme versus those who are ‘outside’ the programme. It can be more difficult to accommodate patients with more complex needs, and those who experience complications or are discharged at weekends, within a standardised programme. This can cause stress for patients, who may worry that they are missing key milestones, and the people supporting them. Staff may find it hard to identify with patient feelings of vulnerability, while some patients may struggle to translate information and advice received in hospital into their home environment. Priorities may differ between patient/carer, staff and organisational groups, which may influence the extent to which personal, individualised care can be provided.

Theme: caregivers as essential

Twenty-two studies (23 articles) support this theme, which details how having a home caregiver, such as a spouse or friend, was seen by nearly all patients, carers and staff members as essential to earlier discharge and recovery at home. 90,100–106,108,110,113,114,116,119,121,123–130 The role of carers in providing reassurance and support with practical tasks is explored, alongside patient concerns around being a burden to those supporting them.

First- and second-order data from 14 studies (15 articles) indicate that the majority of patients in studies viewed having a home carer as essential for recovery at home. 90,102,103,106,108,110,114,116,119,121,123,124,126–128 This is illustrated by patients from two studies who experienced a fast-track protocol during their hospital stay:

I definitely could not have came home and been by myself.

(Patient, sacrocolpopexy)108

I think it’s important to have help in the home … one needs help with shopping and preparing food and so forth … and you are not allowed to drive a car.

(Patient, joint replacement)102

The second quote above indicates some of the practical tasks which patients require support with following hospital discharge (see theme ‘Direct, indirect and other work of caregiving’ below). However, patients also found the presence of someone to support them at home reassuring. First- and second-order data from 10 studies90,102,111,113,114,116,123,127,128,130 indicated that the uncertainty surrounding discharge provoked unease, whether this was regarding not knowing what to expect in terms of outcomes or pain following the operation,127 or finding someone who was able to provide the support required:128

My husband was very ill …. The only help available to me is my daughter and she’s got her two girls and husband at home and family to look after … I can’t really ask her to do any sort of full-time care.

(Patient, liver resection)128

[My wife]’s not a nurse by any means. And … obviously it’s a worry if there’s anything. Whereas when I’m in hospital, you’ve got teams there, if there is a problem …. So from that point of view there’s always that concern isn’t there?

(Patient, liver resection)128

These quotes also suggest that patients are aware of the additional stress providing support would place on their caregivers. Four studies (four articles) indicated that some patients felt a burden to their home carer, or worried about how they would cope with the tasks of caregiving. 113,121,128,130 One patient discussed her concerns around how both she and her partner would cope following discharge after surgery for colorectal cancer:

I felt a little bit lacking in confidence of coming home … I was just a little bit worried about how I was going to cope … how (my partner) was going to cope with this … I was afraid of putting pressure on him.

(Patient, colorectal cancer)130

This quote indicates that providing support to carers is essential, not only to ensure that they can continue with their role as caregivers, but also to reassure patients that the people they care about have adequate support. This may help alleviate any guilt experienced by patients113,128,130 associated with their reliance upon them and facilitating feelings of safety for both parties (see theme ‘The direct, indirect and other work of caregiving’).

First and second-order construct data from four studies indicated both carers105,106 and staff also viewed the role of home-care as essential:100,129

I do what I have been told to do and I keep to the routine: breakfast, do the washing up, make the beds, close the windows …. When I have finished doing that, I … shop …. It’s not difficult and I can easily manage.

(Relative of patient discharged following joint replacement)105

And the consultant explained again that if everything was OK he’d check again on the circumstances at home … he said that if [husband] was happy to have me home, then there would be no reason why I couldn’t go home.

(Patient, gynaecological surgery)100

The second quote100 suggests that the presence of a partner at home can also be reassuring for staff members involved in ensuring the discharge of patients home is appropriate and safe.

Theme: the direct, indirect and other work of caregiving

This theme explores the wide range of work undertaken by home carers, which they may not have had to undertake if the patient had a longer hospital stay. This theme is supported by data from 14 studies,90,100,102,104–106,111,113,115,125,126,128,130,132 and considers the work undertaken by caregivers according to the three categories suggested by Ganske. 111 The first category details the direct work carried out, such as nursing tasks like wound care; the second category explores indirect work such as scheduling and household tasks and the final category details other responsibilities such as caring for children or maintaining employment. This theme also considers the emotional and physical impact of caring for a relative or friend on the carer.

First- and second-order construct data from five studies (five articles) highlighted that the work category of ‘direct caregiver work’ included nursing tasks like wound care, support for bathing, or managing medications. 104–106,111,113 As illustrated by the carer of a CABG surgery patient over 80 years of age, this often required carers to learn new skills, which were often physically and emotionally challenging:

I thought I had it, and I didn’t … get good blood. I think I was afraid of hurting him.

(Carer of CABG patient)111

In addition to carrying out the task itself, direct caregiver work also necessitated gaining access to, understanding and remembering medical information related to the patient’s care,100,104–106 as discussed in more detail within the ‘information’ theme. These tasks represent work which is not immediately observable but place additional stress on caregiver resources. Six studies (six articles) provided first- and second-order construct data which support the work category of ‘indirect caregiver work’,104–106,111,128,132 which included tasks which could not be described as nursing tasks, but encompassed responsibilities the carer had to perform on behalf or, or in order to care for, the patient:

I have three doctors’ appointments [to arrange] … so I called [the urologist]. And then we have to go to [cardiologist] and [cardiac surgeon].

(Carer of CABG patient)111

my schedule is full of appointments for my parents … it’s kind of a juggling act to keep everything [straight].

(Daughter of CABG patient)111

Second-order data from three studies discuss the other tasks and responsibilities beyond the patient that the carer had in their life, such as childcare or a job, which could be challenging to balance. 106,111,128 This care work had a physical cost, causing exhaustion and the exacerbation of existing health problems, particularly in elderly carers:106,111

… sometimes I’d be so tired and numb … especially at 2 or 3 or 4 o’clock in the morning and then … [only] sleep for an hour … when you’re so exhausted … your eyes hurt you’re so tired.

(Carer of CABG patient)111

In addition to this physical stress, four studies (four articles) highlighted the emotional impacts of caring, which included difficulty in witnessing a loved one in pain, stress, frustration and worry. 104–106,111 Two carers discuss their emotional responses to providing support for their family members at different stages of the recovery process following CABG surgery, and one following THA surgery:

this is twice and I don’t want to go through it again … oh I can’t even talk about [burst into tears] … seeing [him] in intensive care … oh it’s horrible. It was like going in and seeing a corpse … I wouldn’t go back again.

(Carer, CABG patient)111

I was worrying …. ‘Am I going to have to help him in and out of bed?’ because he is so much taller, and I know I could do it, but no one has ever taught me how to do it.

(Carer, THA)106

… you know you’re supposed to walk, you know you’re supposed to drink water and you know that you’re supposed to eat something to get your strength back … I’m sick of saying did you, did you, did you ….

(Carer, wife, CABG patient)111

In addition to managing their own emotional reactions arising from the trauma and challenges associated with a hospital stay and discharge, caregivers were also the recipients of anger, frustration and sadness from the people they are supporting:

There was one time when I wasn’t quick enough to provide assistance and then she was mad at me … I wasn’t close by to help her … but we had just agreed that the house needed cleaning so I was … hoover[ing] … I didn’t hear her.

(Carer, spouse, THA replacement)104

The above quote highlights how the challenge of managing multiple tasks at any one time can be frustrating for patients, when their family members are unable to provide them with timely support. Five studies highlighted that in order to fulfil their caring responsibilities and maintain their wellbeing, the carers themselves also needed support. 90,103,104,106,111 Support for home caregivers included receiving relevant information and training from the hospital,106 and support from other family members, friends and healthcare staff. One woman expressed gratitude towards her husband for the support he provided to her:

I’ve been thanking him [her husband] because he’s just been so great in going with me every day … it truly is a great help.

(Carer for CABG patient)111

First- and second-order construct data from three studies highlighted how there could also be negative emotional impacts on the carer or patient, such as conflict due to changes in their respective roles at home. 100,104,126 One patient admitted for colorectal surgery talked about the impact of role changes following her transition home:

We don’t talk much about illness in our house, and I have always been the one to take care of my husband and kids when they were ill. Now it’s me who needs looking after …. That has been a little hard.

(Patient, colorectal cancer surgery)126

However, data from two studies104,111 suggested that the emotions experienced by carers changed over time, with carers expressing more positive emotions as the patient recovered:

We have been very successful. Mom has had a much better recovery than I was worrying about.

(Carer, CABG patient)111

Construct summary: essential care at home

This construct highlights how the direct and indirect work by caregivers is vital to support patient recovery at home post-discharge. The tasks directly associated with caregiving, such as providing support with medications, changing dressings, or providing transportation to appointments, are also associated with work which is not always observable or predictable, such as making appointments and learning new skills. The transition home can be difficult for both the people providing and those receiving care, due to role changes at home and the stress associated with the additional workload of caregiving and/or being dependent on others for support. Older caregivers may have their own health needs and struggle to maintain their own wellbeing whilst supporting the patient to recover at home. Follow-up appointments may provide a way to provide practical and emotional support to caregivers, as well as patients, enabling them to manage their caregiving role.

Theme: the active patient

This theme is supported by 30 studies and explores how patients had to be active, rather than passive, and take charge of their own care, and the ways the staff or programme components signal this to them. 90,100–104,106,107,109–111,113–116,118–132 This theme encompasses three important concepts: ‘patient attitudes and experiences’, ‘pain management’ and ‘empowerment versus vulnerability’.

The enhanced recovery programmes, by their nature, required patients to take an active rather than passive approach, as they would largely be responsible for their own recovery at home. However, two studies102,125 suggested there was variation in how much of a say patients wanted in medical decisions. For example, Berg et al. 102 described how some patients wanted to leave the decisions about surgery to medical staff, whereas others, such as the patient quoted below, wanted to actively participate:

I participated and decided when I wanted the operation. … I felt that I was participating more when it was time to do it, and it was unavoidable.

(Patient, LLA)102

But one study125 suggested this choice to be involved in decision-making may be dependent on patients receiving adequate information in empowering patients to be active patients, as discussed in the ‘Increasing confidence and preparedness through information and planning in advance’ theme.

Two studies114,123 suggested that early mobilisation emphasised the patient’s role as being active in the process of recovery. For example, the below quote highlights that not being ‘allowed’ to stay in bed was a way of communicating independence, though the language of ‘chased’ may also suggest patients can be reluctant to mobilise early and link to the idea of blame being placed on patients who do not conform to the set milestones discussed in the construct of ‘Individualisation of a standardised programme’.

Being ‘chased out of bed’ emphasised the expectation of self-care.

(Patient, TKA)114

Five studies (five papers) suggested it could be difficult for patients managing their own recovery at home to strike the balance between activity and rest. 106,114,116,123,132

You just wonder how much that you should do because I’m that sort of person you know, ‘did I overdo it? Should I have rested more with my leg up?’ and then you worry about developing problems … if you don’t mobilise.

(Patient, LLA)116

The above quote links to the ideas within the construct ‘Feeling safe’, as patients needed and wanted information (and follow-up) to manage their recovery effectively and without anxiety. If patients are ‘responsible’ for their recovery, then they need the tools to be able to do so. It also links to the importance of managing expectations, and patients having a realistic understanding of the likely trajectory of their recovery to have a better idea of, for example, the degree of pain they may be feeling in relation to rehabilitation exercises at different stages.

Seven studies indicated that both staff and patients felt that patient expectations had an impact on engagement, satisfaction, or recovery itself,106,109,114,118,125,129 which links to the earlier theme of ‘Increasing confidence and preparedness through information and planning in advance’ and the need for accurate preparations to manage expectations. Some patients’ expectations could be too high, or they could anticipate that they would feel less pain due to a planned earlier discharge than they may have experienced previously, whereas others had enough information to develop a realistic expectation, as highlighted by the contrasting quotes below.

I don’t know how many people [with TKA] I’ve had in the last little while that come in and they’re stunned that they have pain postoperatively …. They’re so not prepared for the amount of pain they have.

(Staff, physical therapist, LLA)129

My pain was as I expected it and I know how to manage it.

(Patient, LLA)125

Two studies118,129 suggested that staff felt it was important that patients had clear expectations, but that these were not always clearly communicated in a timely manner, linking to the importance of good pre-operative information (see theme ‘Increasing confidence and preparedness through information and planning in advance’). A surgeon highlighted one of the potential challenges for staff in meeting this need, linking this theme with the challenges of individualisation in the ‘Individualisation of a standardised programme’ construct:

… depending on how much time you have to spend with people and so on. You may miss the boat in terms of what they’re expecting.

(Staff, surgeon, LLA)129

The same two studies118,129 highlighted that some staff members felt that the patient’s personality or mental health was important in how well they engaged as an active patient in a programme, as illustrated by the two quotes from staff members below:

One of the most common [concurrent] diagnoses … is depression in the patients … which hugely affects motivation, adherence to the protocols, and follow-up, and it doesn’t get addressed frequently because primary care physicians don’t take the time to diagnose it appropriately.

(Staff, nurse, LLA)129

There is a group of patients who, um, basically do what they want no matter what we say.

(Staff, colorectal surgery)118

In addition, nine studies indicated both patients and staff suggested that the patient’s attitude had an impact on how well they responded to the programme. 100,101,109,113,118,125,128,129,131 This is illustrated by the patient below, who describes optimism, as opposed to pessimism, being a key part of how well you recover.

I think if you went into something with [a pessimistic] sort of mind set it must be very difficult to recover, because you’re not helping yourself. So I think the mind set is hugely important, as important as the body.

(Patient, liver-resection surgery)128

Two studies121,129 suggested that age differences could impact experiences, with Westby and Backman129 suggesting that there were differences in outcome expectations and rehabilitation needs between younger and older patients, and Reay et al. 121 suggesting that older patients were better able to cope with social restrictions caused by their recovery. One older adult discussed how she occupied her time post-discharge:

The first couple of days, uh, mainly sat down, mostly exercising; it’s in the book (the hip replacement information booklet provided). I wasn’t bored; I read quite a lot.

(Patient, THR)121

However, this does not mean to say that all older patients coped well with recovering alone, as this was also dependent on other factors such as recent loss:

I think my mental thing is more related to the loss of my wife …. Because I’m in a house on my own …. Now whether that’s had an impact on the way … I’ve recovered I don’t know … but I feel it has had an effect.

(Patient, TKA)101

The previous quote also suggests a link back to the construct of ‘Individualisation of a standardised programme’, as individual circumstances, such as recent loss, may impact experiences, and may require a different level or kind of support in home recovery.

The experience of pain and pain management was an important specific feature of being an ‘active patient’, and was a salient issue for our PPI group. This links to the need for accessible follow-up, the work of carers (when they are managing the medication), and how complications can impact perspectives.

Fourteen studies (15 papers)90,102,104,106,111,114–116,120,123–126,129,132 suggested that a number of patients and carers found it challenging to know how to manage post-operative pain when at home, including uncertainty in how to manage their pain or medication, needing more guidance or information to do so effectively. This is illustrated by the quote below from a carer who describes the impact of not understanding the need for regular pain medication to prevent pain from escalating to a more unmanageable level:

The meds, too … no one told us to make sure we take [them] … and he got really in hot water [not taking enough], so the pain came back full force, and it took longer to get it down again.

(Carer of THA patient)106

Some feared becoming addicted to the medication, or taking too much, so were reluctant to take it as described, as suggested by six studies102,106,116,120,124,129 and illustrated by the quote below.

I only took them for about a few days …. But they said you must take them because you get better quicker with pain relief. But, I just ended up taking a couple of paracetamol … because I don’t have a lot of tablets.

(Patient, non-cancerous gynaecological surgery)120

Others struggled with side effects of the medication, and tried to balance between the amount of pain they were in and the side effects they would experience (described by three studies; four articles). 106,116,123,124

It’s the painkillers you become so tired from … a total lack of energy, so it’s not easy to do anything.

(Patient, LLA)124

Being in pain could impact patients’ quality of life, including their ability to sleep and mood.

I was awake all hours of the night … not to be able to get any sleep at night is a horrible thing. (Participant 2)

Yes, it was painful it didn’t stop me sleeping but it was painful. (Participant 10)

(Patient, LLA)125

Being in pain could also impact patients’ ability to do prescribed exercises or rehab, or they struggled to find a balance between pain and movement, as indicated by five studies (six articles),106,114,120,123,124,129 and illustrated by the hip arthroplasty patient quoted below from Churchill et al.:

I pushed myself, my muscles were inflamed, and there were nights I didn’t sleep as well because I was doing too much.

(Patient, THA)106

However, there was variation between patients, even within the same studies examining the same programme or regimen, again suggesting a need to individualise information or protocols to the patient (see construct: ‘Individualisation of a standardised programme’).

I’ve hardly had any pain, I’ve really only taken the tablets that I was forced to take.

I needed more, I never had enough. I didn’t want to take an overdose either. That was the thing that felt hopeless and disconsolate … that I didn’t experience any relief so that I could relax and feel hope.

(Patient, LLA)102

These issues around pain management could impact a patient’s ability to be actively involved in their own care, due to the physical and cognitive impacts of, for example, medication side effects or pain itself. It links to the construct of ‘Feeling safe’ – patients need clear information and support to manage their pain at home, as well as a consideration of their emotional needs (such as anxieties over addiction to medication).

Another key issue for being an ‘active patient’ linked to the construct of ‘Feeling safe’ is the finding that being an ‘active patient’ could be empowering to some, but could make others feel vulnerable, suggested by 15 studies (15 papers). 90,100,106,107,109,110,114,119,122,124,126,128–130,132 For example, Thomsen et al. 126 described how some patients and carers were forced to be ‘active’ due to a lack of follow-up care, or a reluctance to contact medical staff. For some, this nurtured a feeling of being in control, but for others it could increase feelings of vulnerability and a lack of safety, particularly if they experienced complications or unexpected side effects, or promised follow-up care was not available:

I was not prepared for being continually tired. It makes me very afraid.

(Patient, colorectal cancer surgery)126

Theme: staff skills – expertise versus generalists

Four studies (four papers) contributed to this theme, which describes that there were differing views amongst staff about whether you needed a specialist for particular components of a programme, or if generalists could be trained to do them. 101,112,117,129

Three studies112,117,129 suggested that some staff members had different perceptions of the bounds of different staffing groups’ roles, and who was primarily responsible for particular tasks or even the patient themselves. For example, in Haas et al. ’s study,112 doctors and nurses generally believed it was a physiotherapist who was responsible for early mobilisation, but physiotherapists felt they were only responsible for more complex patients and nurses were otherwise well placed to mobilise patients.

Having a physiotherapist to get them out of bed in the morning, they build the confidence in the patient, they know exactly what they’re doing … and know they’re capabilities, know exactly what to do with that patient.

(Nurse, LLA patients)112

It’s the difficult ones to get up and moving that’s most challenging. That’s where physiotherapists are best placed to use their expertise.

(Physiotherapist, LLA patients)112

Similarly, in Kocman et al. ’s study,117 which aimed to introduce the CGA into the care pathway, geriatricians saw their role as supporting the general team to develop specific skills and knowledge in caring for older people, whereas those in the general team saw the geriatrician as the one to hold those skills and knowledge.

One study101 examined an intervention in which more junior rehabilitation assistants worked in patients’ homes under the (remote) supervision of a trained physiotherapist. At times, the assistants could feel out of their comfort zone, but good support and communication helped to ensure both the assistants’ and the patients’ confidence, and physiotherapists aimed to find a balance between being supported and being independent.

It’s quite nice from a physio[therapist] assistant point of view to feel like you’re making an impact independently and [the physiotherapist] trust[s] you …. But at the same time, I didn’t feel at all like I was … abandoned … it was a really nice balance.

(Assistant, TKA patients)101

Theme: staff/service co-ordination

Nine studies (nine papers) contributed to this theme about the co-ordination of staff and services. This includes within hospital teams, between hospital services, and between hospital and community-based services. 104,105,107,111,114,115,118,124,129 It also includes the importance of a co-ordinator, from the perspectives of staff and patients/carers.

Four studies105,114,118,129 suggested that, where staff teams within or between care settings did not co-ordinate, it could cause confusion for patients when they received differing instructions, linking to the construct of ‘Feeling safe’, in which inconsistent information made patients feel less safe. For example, Westby et al. 129 described how their focus group of patients spent a great deal of time discussing issues of staff and service communication, with most patient examples describing ‘how poor or a lack of communication decreased efficiency, effectiveness, and collaboration’.

Communication issues could also cause confusion for the staff themselves, as illustrated by the quote below, which could also indicate a potential lack of understanding of the ERAS protocols in place by some staff members:

Sometimes they’ll say, ‘ERAS’, and then, ‘Nil by mouth’. So sometimes it’s not always very clear in the documentation.

(Staff, colorectal surgery patients)118

Five studies (six articles) also suggested that a lack of communication between services could result in sub-optimal, less individualised, care for patients:105,111,115,123,124,129

we have this parade of people with total hips … coming through as though they’re all the same and they’re not. I think this lack of information leads to rote [physical-therapy] procedures that don’t have very much thinking going on with them.

(Staff, physical therapist, LLA)129

But three studies107,118,129 indicated that when there was a small team or a named contact person for staff, it was easier to obtain the information needed:

… because we’re small, we can call up one person … so it’s easy. I think it works well, the link from the communication we have, acute hospital stay to community back into the outpatient department.

(Staff, occupational therapist, LLA)129

Within hospital, a care co-ordinator can help ensure that the different programme components are completed in a timely way to ensure that other components are not delayed, ensure an early discharge, and be available for staff or patients should they have questions (as indicated by six studies104,107,115,118,124,129). For example, Specht et al. 124 described how patients’ discharge could be delayed if they couldn’t take part in their training because they were waiting for a dressing to be changed. Staff in Lyon et al. ’s study118 felt that an ERAS co-ordinator helped staff to follow the protocol and to answer any patient’s questions:

I suspect without an ERAS coordinator it would be futile.

(Staff, colorectal surgery patients)118

When patients have no care co-ordinator, it can fall to them or their carers to do the work of trying to chase appointments, which could be especially challenging when professionals they needed to see were not co-ordinated, linking to the theme of ‘The direct, indirect and other work of caregiving’.

I have three doctors’ appointments [to arrange] … so I called [the urologist]. And then we have to go to [cardiologist] and [cardiac surgeon].

(Carer for CABG patient)111

Somehow, I felt that medical imaging nurses had their own schedule and physiotherapists had their own and they were not at all synchronised.

(Patient, THA)115

Two studies104,107 specifically investigated patients having a care co-ordinator, and it was her availability to patients which made the biggest difference, linked to the idea of making follow-up care easily accessible. For example, Berthelsen et al. 104 described how spousal carers specifically mentioned their case manager’s constant availability, explanations of information and ability to arrange appointments with other health professionals as what they appreciated about their involvement.

There was sometimes disagreement about who had responsibility for patient care following discharge, linked to service co-ordination. This made it more difficult for patients to know whom/where to contact if they had questions or difficulties. Within the ‘Appropriate, available and accessible follow-up’ theme, it was noted that patients do not always know whom they should contact when they have a problem or an issue, suggesting that it is not clear to patients who is taking responsibility for their care post-discharge. This was also reflected in four studies103,115,122,129 related to staff, who on occasion disagreed about who had primary responsibility for the patient.

I called the … nurse who said … I should ring the primary care centre, but they say that the operation was the hospital’s responsibility. Then they said: let’s not concern ourselves about this any more; from now on you can fix this yourself.

(Patient, colorectal cancer surgery)122

Construct summary: taking responsibility

This construct focuses on responsibility for patient care and recovery, particularly after discharge from hospital. Patients had to be active, rather than passive, and take charge of their own care. The experience of pain and pain management was an important specific feature of being an ‘active patient’, with some patients reluctant to take medication, and others (and some carers) unsure of how to manage medications. For some, the experience of being an ‘active patient’ could be an empowering experience, but for others it made them feel vulnerable. Patient expectations could impact their recovery or their views on recovery. Many patients and staff felt that a positive attitude was important in recovery, but this could also become blaming towards patients who could not fit the programme or did not recover ‘well’. This construct also highlighted the importance of good staff and service co-ordination, particularly after discharge, so that patients (and carers) received consistent information and were clear on whom they needed to contact if they had difficulties or questions. However, it also highlighted that some staff/services disagreed about who had primary responsibility for the patient following discharge, and that there could be differing views on whether generalist or specialist medical staff were best placed to provide care.

Theme: masking the negative

Eight studies (eight papers)90,100,105,106,109,126,128,132 contributed to this theme, which describes how some patients give high satisfaction ratings, but this can mask concerns or things they would improve, which are then missed in quantitative evaluations.

As highlighted in the follow-up theme within the construct ‘Feeling safe’, patients and carers may not always contact their care team because they did not want to bother them. 100,105,106,109,126,132 This may lead to a skewed view of how the patient is progressing, or how the programme is working overall, as the care team then does not hear about problems or issues the patients were facing:

… even though they say that if you’ve got any problems you can ring us, well I know … I don’t like to bother people, and I probably wouldn’t have phoned unless I was really, really worried.

(Patient, gynaecological cancer surgery)100

Hip arthroplasty patients from one study90 explicitly highlighted that though patients did have criticisms of the service or care that they had received, most masked this by trying to justify the problem on the health team’s behalf:

I suppose, though really they’ve not got time and there aren’t enough physios probably, for this. But er, that’s what I feel. I think physio is very, very important, proper physio.

(Patient, THA)90

Studies contributing to this theme indicate that patients may not always raise concerns when they occur or try and excuse services which did not meet their expectations, which may make it hard for services to identify patients and/or caregivers who are struggling or dissatisfied and offer appropriate support. This highlights the importance of active follow-up by staff at key points in the patient’s recovery pathway.

Theme: the right questions at the right time

This theme explores how people’s experiences of recovery and of caring for patients can change over time, meaning that asking evaluation questions at one time point may not provide a complete understanding of patient experiences. Six studies (six papers) contributed to this theme. 110,111,115,123,129,130

One study115 noted that their LLA patients did not know how to give feedback on their experiences, particularly specific feedback for specific aspects of the service. Our PPI group also emphasised that not only did they want the opportunity to provide feedback or information, but that they wanted this communication to be acknowledged or responded to, particularly in terms of the impact it may have on changes to the service.

One study129 highlighted that it was not clear how outcomes should best be measured, with staff participants highlighting that outcomes measurements were not standardised across services, or given at all time points, which made it difficult to make comparisons. Some staff in this study suggested questionnaires were not very useful in any case, and others highlighted that the patient’s goal was often getting back to ‘normal’, and the individualised goals which mean ‘normal’ were not well captured by these tools:

… it would be nice for people to actually use the same outcome measures pre-operatively, immediately post-op … so you could actually see a difference.

(Staff, physiotherapist, LLA)129

I don’t ask patients to fill out questionnaires. That’s highly inefficient.

(Staff, surgeon, LLA)129

The point in the care pathway / duration since discharge at which patients or carers are asked for feedback can make a difference, as their experiences and views on programmes can change over time, for the negative or positive. Galli et al. 110 described the experiences of one patient who felt positively at discharge, but one month later felt angry and unhappy with her experience as she did not feel she had an effective recovery. In contrast, Ganske111 suggests that carers initially have more negative emotions and perceptions of the experience at discharge, but that they became more positive in time as they gain more confidence in their caring tasks and the patient recovers. This suggests that evaluation may need to be longitudinal to capture the full range of experiences of both patients and carers.

Construct summary: outcomes

This construct highlights how services may receive an overly optimistic picture of a patient’s recovery journey due to patient reluctance to report concerns or difficulties, or appear critical of the care they receive. Patient and carer views and experiences may also change over the course of the recovery journey, necessitating the need for repeated follow-up by services. However, this follow-up will only be useful to services and patients if the discussion reflects the potential needs of services and individual patients and/or carers at that particular stage of the intervention pathway.

ERP interventions following abdominal surgery

We first created a model for the ERP abdominal-cavity studies, which we intended to form the basis of subsequent models. We initially looked at four conditions reflecting whether the intervention (1) provided information in different formats to participants, (2) had an emphasis on goal setting, (3) had a consistent co-ordinator role (e.g. a nurse) throughout the intervention and (4) collected a high number of PROMs.

An initial truth table was created that suggested that studies that had an emphasis on understanding patient experiences through collecting a high number of PROMs tended to be successful (see Table 6). However, on further inspection it became apparent that some configurations within the truth table included a mixture of successful and unsuccessful studies, and an alternative model was developed. This is summarised in the truth table in Table 7, which examined configurations according to whether the study (1) provided information in different formats to participants (Diff_Format), (2) had an emphasis on goal setting (Goal_set), (3) had a consistent co-ordinator role (e.g. a nurse) throughout the intervention (Nurse_Led) and (4) included efforts to support discharge (Supp_Discharge).

The truth table reassuringly showed that studies that did not have any of the four conditions belonged to an unsuccessful configuration (a configuration composed of three studies). 83–85 However, the truth table also revealed that a study with all four elements conducted by73 was also unsuccessful, which challenged the assumptions outlined in the logic model (configuration F below). Further examination of the characteristics of this study showed that the study also included a distinct set of key intervention features, with the intervention’s components including light and music therapy alongside psychological intervention; all other interventions included in the dataset were evaluating changes in the type and implementation of clinical practice (e.g. changes in fluid management) as well as changes in the implementation of care practices.

As the intervention appeared to be based on a distinct programme theory that was unlike other interventions, the remainder of the analysis is based on 11 studies, with a revised truth table shown in Table 8. Four configurations are observed as triggering a successful outcome, giving an indication that there are multiple pathways to triggering a successful outcome, although only one of these is supported by multiple studies. We can see that in this truth table, seven out of a possible 16 potential configurations of conditions are observed in the data. A study conducted by Forsmo and colleagues88 which trialled an ERAS protocol for patients who had experienced colorectal surgery was the only study in the data coded as including efforts supporting discharge; this study improved length of stay but not PROMs and was allocated to the partially unsuccessful set.

Prehab interventions for abdominal surgery

We started with a model that mirrored the conditions for the ERP studies which required further iteration due to inconsistencies within configurations. A truth table (see Table 9) consisting of four conditions was developed that reflected whether interventions (1) provided information in different formats to participants (Diff Format), (2) had an emphasis on goal setting (Goal set), (3) included additional time being spent with patients (Add time) and (4) included an emphasis on the early mobilisation of patients (Early Mob). Of 16 possible configurations, seven were observed within the data, with the largest configuration consisting of six studies that were unsuccessful which provided additional time for patients and goal-setting but did not focus efforts on early mobilisation or provide information in different formats. This suggests that configurations involving these latter two conditions may be particularly important in determining intervention success; this is explored further in the minimisation process.

Interventions for lower-limb arthroplasty surgery

Data from 14 studies are included in the truth table (see Table 10), which explores whether interventions (1) included an emphasis on the early mobilisation of patients (Early Mob), (2) had an emphasis on goal setting (Goal set), (3) had a consistent co-ordinator role (e.g. a nurse) throughout the intervention (Nurse_Led) and (4) collected a high number of PROMs (high PROMs). Of 16 possible configurations, nine were observed within the data. Four configurations were observed to trigger successful interventions; configuration D was close to the consistency threshold of 0.79 and included a partially successful study where there was an improvement in at least one PROM but not in LOS;51 configuration C also included a partially successful study. 78 Two of the configurations suggest that studies that had an emphasis on understanding patient experiences through collecting a high number of PROMs were also more likely to be successful, albeit with the presence of other conditions. The truth table in this case helps to separate out configurations that include fully as well as partially successful studies from those that are partially or fully unsuccessful.

ERP interventions following abdominal surgery

The solution shows two pathways to triggering a successful intervention (see Table 11) – the first pathway suggests that goal-setting alone is sufficient (although not necessary) to trigger a successful intervention; the second suggests that providing information in a different format in the absence of taking efforts to support discharge leads to a successful intervention (not supporting discharge, represented by the ~ sign in the table). The interpretation of this solution is discussed later in the results. The solution has a high coverage value, and exploration of individual cases shows that the solution accounts for all cases that were full members of the effective intervention set.

Prehab interventions supporting abdominal surgery

This dataset included four studies that were in the successful intervention set, and the solution shows two pathways to triggering a successful intervention (see Table 12) – the first pathway suggests that providing information in a different format and spending additional time with patients is sufficient to trigger a successful intervention (the ‘and’ relationship represented by the * sign); the second pathway suggests that efforts to support early mobilisation but not providing information in different formats and not setting goals leads to a successful intervention (‘not’ represented by the ~ sign below). The interpretation of this solution is discussed later in the results. The solution has a high coverage value, and exploration of individual cases shows that the solution accounts for all cases that were full members of the effective intervention set, although it does not account for two studies that were partial members of the successful set whose characteristics were indistinguishable from less successful studies. 54,57

Interventions supporting lower-limb arthroplasty surgery

This dataset of 14 studies included six studies that were full members of the successful outcome set and two that were partial members, and the solution accounted for all eight of these studies (see Table 12). Two pathways were identified that triggered a successful intervention – the first accounted for four studies and suggested that a combination of early mobilisation and goal setting triggered a successful intervention; the second also accounted for four studies and suggested that goal setting, along with a consistent point of contact (e.g. a nurse) leading the intervention, along with a high number of PROMs being measured, were sufficient to trigger a successful intervention.

Additional quality checks implemented across all models

In obtaining the solutions above, some untenable assumptions may have been made about the logical remainders and we implemented an algorithm developed by Dusa151 to remove untenable and contradictory logical remainders that could be otherwise be used to generate the solution, generating an ‘enhanced intermediate solution’. Further checks on the solutions were also undertaken to ensure that the solutions did not also trigger the negation of the outcome (a possibility in QCA due to causal asymmetry). Finally, any discrepancies in coverage were also checked, with the models for both sets of abdominal studies accounting for all studies that were full members of the successful outcome set, and the model for LLA accounting for full and partial members of the successful outcome set.

Enhanced recovery protocols for abdominal surgery

Two pathways were identified here, with one pathway suggesting that goal-setting alone was sufficient to trigger a successful intervention, represented by two studies, one of which also appeared in the second pathway. In both studies, patients were informed of protocols and goals for different stages of their recovery, for example daily mobilisation and nutritional goals. 44,45 Such goal-setting was found to be an important way of allowing patients to measure their progress and self-motivate in the qualitative synthesis. Both the qualitative synthesis as well as the results of recent trials153 emphasise that goal-setting should be person-centred and should be viewed as shared by the patients and the clinicians. Given that both successful studies describe the clinical elements of the studies in greater detail than some of the elements around implementation (i.e. how the intervention was conducted), it is unclear the extent to which goal-setting was implemented as a joint enterprise and was person-centred in nature. 44,45

A second pathway suggested that providing information in different formats, and not implementing efforts to support discharge, triggered a successful intervention. The importance of providing information in different formats (e.g. written and orally) is congruent with the qualitative synthesis, which showed that providing information clearly helped improve the confidence of patients in self-management tasks, but also improved understanding of why tasks were important to their recovery. Efforts to support discharge were not reported in 11 ERP studies, or were similar in intervention and comparator groups, and were only present in an unsuccessful study conducted by Forsmo and colleagues. 88 The intervention involved stoma education, which was intended to allow patients to be better prepared for discharge and subsequently managing a stoma at home. This intervention was associated with reduced LOS but there was no improvement in PROMs measured.

Given that there was only one study reporting differences in efforts to support patients at or after discharge, there is insufficient evidence in the studies to fully evaluate the impact of this condition on the outcome. It may be that while there were some measures to support discharge put into place, in the case of Forsmo these did not fully prepare patients, although the description provided does not allow us to assess this further. 88 While the second pathway in this group of studies indicates that not implementing efforts to support discharge triggers a successful outcome, this suggestion is incompatible with our logic model and the reflections of PPI members and clinical stakeholders.

Prehab interventions supporting abdominal surgery

For Prehab interventions, providing information in different formats was identified as a trigger for successful intervention in one pathway, but only in combination with spending additional time with patients. This pathway once again suggests that a person-centred approach to the implementation of the intervention generates a successful outcome; here it suggests that providing information and allowing time for questions and the development of broader relationships can lead to more individualised care and improved patient outcomes. For example, in the case of one successful study supporting this pathway, in addition to education and reinforcement, monthly nurse-led telephone calls not only allowed patients time to ask questions, they also ‘provided reassurance’ to patients, emphasising that information in conjunction with creating a safe space to process information and ask questions is important in reducing length of stay and improving PROMs. 52

One study found that early mobilisation but not providing information in a different format and not incorporating goal-setting triggered a successful outcome. The intervention evaluated by Kapritsou and colleagues focused heavily on progressive mobilisation out of bed 4 hours after surgery, with four daily bouts of ambulation in the first day post-surgery, whereas conventional protocols involved far fewer bouts of ambulation from the outset. 80 Such efforts align with the qualitative evidence that early mobilisation helps crystallise the patients’ role as active patients (which may reduce the length of stay) and successful mobilisation can send a message to patients that they are on the path to recovery. While early mobilisation in this study appears to trigger success in the absence of the other two components, this may be a feature of reporting more than a form of conjunctural causality in this study given that early mobilisation may itself be viewed as a shared goal between patients and clinicians and the study reported provision of information as a core intervention feature (although this was not reported as being provided in different formats).

Interventions supporting lower-limb arthroplasty

Goal-setting was a feature of both pathways identified as triggering success for interventions supporting LLA. Among four studies, success was triggered when goal-setting was accompanied by early mobilisation, with one study noting that efforts to support early mobilisation were set around specific milestones (goals), including getting up on the day of the surgery and climbing stairs two days afterwards, and was accompanied by positive reinforcement by a case manager. 75

In a different set of studies, goal-setting was accompanied by a consistent point of care and coordination as well as a high number (4+) of PROMs measured (even if not all changed positively). We interpret this latter condition as being symbolic of a commitment by triallists to promote patient-centred care and to understand patient experiences in a holistic way; while all trials included measured a PROM as part of the inclusion criteria, there was variation in the extent to which triallists were interested in multidimensional measures. For example in one successful study included in this configuration,86 a key aim was to address a gap in evidence about the broad functioning and quality of life and levels of pain among patients who have a shorter hospital stay. In addition, the role of a consistent coordinator was also identified as crucial in this pathway alongside goal-setting and a consistent point of care, which is congruent with the qualitative research where a coordinating role was observed to be a vital one among patients and clinicians to ensure adherence to protocols and to facilitate the patient journey. Returning to the example of Siggeirsdottir and colleagues, a physiotherapist and/or OT was involved in the care of each patient, providing education from one month before the surgery, through to accompanying the patient home if needed, and conducting a number of visits after discharge (a median of four within the first 2 weeks post discharge) to ensure that the regime was being followed. 86 In the qualitative synthesis, such a regime that involves a consistent point of care helped to ensure that patients felt that their care was individualised and that follow-up care was easily accessible.