Evolution of enhanced recovery program in French university hospitals: a before-after retrospective survey

Background: Enhanced recovery program (ERP) after surgery needs development in Assistance Publique Hôpitaux de Paris. Methods: a retrospective before-and-after study was performed in 2015 and 2017 on three surgical models (total knee arthroplasty (TKA), colectomy, and hysterectomy) in 17 hospitals including 29 surgical departments. For each elective surgery, data were collected in one control intervention (Total hip arthroplasty (THA), gastrectomy and ovariectomy).To favour development of ERP strategy, in 2016, institutional mobilization developed with one day meeting of information, development of a Massive Open Online Course on ERP, diffusion of national update on ERP by Haute Autorité de Santé and inclusion in a regional professional partnership program. Primary outcomes were length of stay (LOS) and complications after surgery. Data on ERP items were collected in the patients’ chart and in anaesthetist and surgeon interview. 70 % application rate reects application of ERP procedure. Results: 1321 patient’s les were analysed (812 in 2015 and 509 in 2017). The length of stay (LOS, mean (SD)) is reduced by 1.6 day for TKA (2015: 8.7 (6.7) versus 7.1 (3.4) in 2017; p<0.001) but increases of 0.5 day for THA (2015: 8.2 (4.1) versus 8.7 (15.5) in 2017; p=0.011). LOS is stable for colectomy and hysterectomy as in their control groups (i.e. gastrectomy and ovariectomy). Incidence of severe complications after surgery is unchanged in all type of surgical models except in THA patients (11.1% in 2015 and 1.5% in 2017, p=0.029). For TKA and hysterectomy respectively applied items of ERP (i.e: >70% application) increased respectively from 5 to 7 out of 17 and 16 in 2015 and 2017. For colectomy, they were stable at 6 out of 21 in 2015 and 2017. All centres progressed in ERP between 2015 and 2017 for TKA (7/7 centres) and hysterectomy (6/6 centres). The LOS was negatively correlated with ERP items application when data collected in 2015 and 2017 were analyzed together. Conclusion: ERP application improved between 2015 and 2017 for three surgical models after an institutional information and diffusion of recommendations in 29 surgical departments of seventeen


Introduction
Enhanced recovery program (ERP) for surgical patients is an approach for perioperative multimodal management of the patient aiming at a rapid restoration of previous physical and mental capacities allowing the reduction of perioperative morbidity. It also results in shorter hospital stays. This approach corresponds to a speci c organization of care around patient-centered clinical paths. A recent metaanalysis con rms its effectiveness in reducing the length of stay (− 1.14 days on average, 95% CI − 1.45 days to − 0.85 days) and the overall rate of complications (risk ratio (RR) 0.71, 95% CI 0.60 to 0.86), all specialties combined (1). However, several authors point out the di culties or failures of the method.
These guidelines remain di cult to adopt essentially because they require the simultaneous and complex involvement of all members of the perioperative team (2). A Canadian team, having deployed an ERP on colonic surgery in more than 2 years in 15 academic hospitals, concludes to the importance of a demand that emerges from the eld and the need for the cohesion of the institution (3). However, other organizational aspects are to be credited with this approach: the bene ts observed, in particular in terms of care load, would allow the movement to be maintained and there is a diffusion effect towards other interventions initially not concerned (2).
In 2016, the French Haute Autorité de Santé has proposed an update in the form of an orientation report and a memo sheet encouraging the development of the ERP (4), while the Assistance Publique Hôpitaux Thus, the primary objective of this study is to measure the effects on the length of hospital stays and on safety of the implementation of ERP, throughout several hospitals in the APHP institution.

Methodology Population
Inclusion criteria: adult patient, without opposition to the extraction of data from her/his le and the follow-up on day 30, scheduled for one surgical model interventions or a control intervention. The following interventions were studied as ERP models since speci c guidelines exist for ERP in these models, they are frequent (≥ 600 annual stays at the APHP) and carried out in around ten different services (which will allow an analysis of the determinants at the service level. We chose total knee  (5); 5. Preparation of participation to a regional program organized by Regional Health Agency (Agence Régionale de Santé) offering inter institutional collaboration on ERP program. This was the initiative of ten surgical departments out of 29 involved in the survey participate in 2017 to the Regional Health Agency training program offering inter institutional collaboration on ERP program. These centres were involved as learning centres or teaching expert based on the existence of ERP protocols and previous evaluation with GRACE group. Four participate in orthopaedic surgery (2 expert, 2 learning centres), 3 in visceral surgery (1 expert, 2 learning) and 3 in gynaecological surgery (1 expert, 2 learning).

Outcomes Patients
The main demographic (age, sex), clinical (main pathology, comorbidities, ASA score, autonomy) and sociologic (living alone) characteristics were collected simultaneously. The Charleson score was calculated for all patients.

Length of stay, complications an rehospitalisation
Patients' les were used to obtain duration of stay and incidence of complications after surgery. A list of eight complications was used for identi cation in the patient's le (i.e. transfert to intensive care unit, new surgery, bleeding, infection requiring antibiotic, pulmonary embolism or venous thrombosis, allergy, aggravation of existing medical condition, other). It was de ned in the patient's chart whether this complication has prolonged the hospital stay or if a new hospitalization in the same institution occurred within 30 days after surgery.

Patient management and ERP
For each ERP model intervention, a speci c grid was designed based on existing ERP guidelines in 2016 for each surgery to collect data describing the patient management process pre, per and postoperatively (6). The cut off value of 70% for ERP item application (i.e. 70% of a particular item is applied on the patient population) was considered as re ecting su cient appropriation (7).

Interview of professionals
It was performed when research assistant visit the hospital for data collection with a representative of surgical and anaesthesiology department to describe level of development of ERP, resources issue and di culties. Concerning services, the characteristics collected concerned: structure (equipment) and resource (personnel, quali cation) data; commitment to quality improvement approach (Morbi mortality review; quality professional improvement), existing clinical pathways. The same independent dedicated staff trained to data collection, collected all data and performed interview in each centre

Statistical analysis
Primary outcome was the length of stay of the different surgical models selected, over a 2 periods of one year, before (year 2015) then after (year 2017) the institutional awareness-raising phase on ERP. The secondary objectives were: effects on the length of hospital stays in each centre following an awareness phase on the practice of ERP; evaluation of the conformity of practices with regard to ERP before and after awareness-raising phase; evaluation of the incidence of 30 days post-intervention complications, in each intervention, before and after the implementation of the program We postulated that the implementation of the ERP would modify the compliance rate from 60% of noncompliant les "before" to 40% of non-compliant les "after", then requiring, for a two-sided alpha risk of 5% and 80% power, 2 groups of 110 les, ie 220 les in total (110 before and 110 after) per model intervention.
However, there is probably a cluster effect at hospital level, all the interventions of a model being supported by a single anaesthesia service and a single surgical team, and therefore the observations of the same hospital tend to resemble each other and provide less information than the same number of independent observations. Considering the literature on cluster effect for process-related variables, we retained a 0.02 ICC (8). Then, starting from 25 les on average per hospital and model, the design effect is around 1.5, and 380 les must be studied per model intervention (190 before and 190 after) to obtain the same information amount.
The number of "control" interventions is halved.
All comparisons were made with Chi2 test or Fischer test for discrete variable and t-test or Wilcoxon ranksum test for continuous variable between 2015 and 2017 period for the whole population and each centre for control and ERP surgical models.
A linear mixed model was used to identi ed variables related to LOS among patient's characteristics, by taking into account the existing correlation between les from the same surgical department.
Missing data were analysed depending on the type of item and best available interpretation. Most of items that were not available in the patient's chart were considered as absent and non-performed.

Data collected
Seventeen hospitals to participate with a total of 29 surgical and 17 anaesthesia departments involved.
No data were collected in 2017 in 11 surgical departments. A total of 1321 patient's le were collected in 18 surgical departments (812 in 2015 and 509 in 2017; Table 1). Data were collected for "before" between November 2017 and June 2018 and for "after" between December 2018 and February 2020 Demographics and patients characteristics Data are listed in Table 2. The patients are younger in 2017 only after TKA. In most surgical models, patients are living alone more frequently in 2017.     The incidence of complications within the rst month was unchanged in all type of surgical models except in THA patients (11.1% in 2015 versus 1.5% in 2017, p = 0.029). The prolongation of hospital stay related to complications was reduced in frequency after TKA.

Patient management and ERP compliance
Some ERP items were detailed in the results but treated as one single item (i.e. oral or subcutaneous thromboprophylaxis; technique of regional anaesthesia for pain control; less invasive surgical technique). Therefore the total effective item number was 17 for TKA, 21 for colectomy and 16 for hysterectomy.
Orthopaedic surgery (Fig. 1) For TKA, the regional anaesthesia techniques (i.e. femoral block, in ltration and adductor canal block) evolved over time and the three techniques were considered as appropriate and cumulated in one item. The applied items of ERP (i.e: >70% application) increased from 5 to 7 out of 17 between 2015 and 2017. Five items of patient management had already reached optimal appropriation level in 2015 (i.e. > 70%) and did not regress (antibiotic prophylaxis, hypothermia, regional anaesthesia) or even progressed signi cantly in 2017 (thrombosis and PONV prevention). Signi cant improvement in ERP was observed in 8 additional items (preoperative information and fasting, strengthening of the quadriceps, intraoperative exacyl, no tourniquet use, early mobilization, intravenous line discontinuation, drainage removal). Further improvements are still required for day of surgery admission. A similar pro le was observed for THA with applied items of ERP (i.e: >70% application) increased from 5 to 7 out of 17 between 2015 and 2017 (data not shown).
Visceral surgery (Fig. 2) For colectomy, applied items of ERP (i.e: >70% application) were stable at 6 out of 21 in 2015 and 2017. Within the six items of patient management which had already reached optimal appropriation level in 2015 (i.e. > 70%) some were stable (antibiotic prophylaxis, hypothermia, PONV) when intraoperative use of dexamethasone or no colic preparation decline and thrombosis prevention signi cantly increases.
Signi cant improvement in patient management and ERP was observed for preoperative fasting, early feeding and mobilization and intravenous line discontinuation. Numerous improvements are still required on preoperative information, use of carbohydrate, same day admission, use of local anaesthetic, volemia monitoring, nasogastric tube, urinary drain and drain removal. For gastrectomy applied items of ERP (i.e: >70% application) decline from 5 out of 21 in 2015 to 4 out of 21 in 2017.
Gynaecological surgery (Fig. 3) For hysterectomy, applied items of ERP (i.e: >70% application) increased from 5 to 7 out of 16 between 2015 and 2017. Five items of patient management had already reached optimal appropriation level in 2015 (i.e. > 70%) and were stable (intraoperative use of dexamethasone, antibiotic prophylaxis, hypothermia, PONV and thrombosis prevention). Signi cant improvement in patient management and ERP was observed for compression stockings, modern fasting protocol, non-invasive surgery, early feeding and mobilization, early discontinuation for vaginal lock or bladder catheter. Numerous improvements are still required on preoperative information, same day admission, use of local anaesthetic, and early discontinuation of intravenous infusion. A similar pro le was observed for ovariectomy with applied items of ERP (i.e: >70% application) increases from 6 to 7 out of 21 between 2015 and 2017.
Individual trajectory of different structures (Fig. 4A, 4B, 4C) All centres with data collected at 2015 and 2017, progressed in ERP between 2015 and 2017 for TKA (7/7) and hysterectomy (6/6) with no signi cant difference between centres. Evolution was mixed for colectomy with some centres reducing their application (2/5) and some improving (3/5) with no signi cant difference between centres. For the 6 learning centres participating to the Regional Health Agency training program the trajectory was not signi cantly different from other centres (p > 0.99). The single expert centre in gynaecological surgery has the best performance on 2017; the 2 expert centres in orthopaedic surgery were the second and third best centres in 2017; in visceral surgery the only expert centre did not collect data in 2017.
Correlation between patient management and outcome (Fig. 5, Fig. 6) All individual data of both periods before and after were analysed for relation between individual LOS and number of ERP items applied. There is a signi cant correlation between LOS reduction with increased application of ERP for TKA, colectomy and hysterectomy (Spearman test; p value and correlation coe cient: TKA: 0.00003 and − 0.202; colectomy: < 0.001 and − 0.232; hysterectomy: 0.0018 and − 0.182 respectively). When grouping in three categories according to a previous publication describing a dose dose-response relationship between various level of ERP item application and LOS (6, 8) (i.e. ERP items application < 50%; 50-70%, > 70%), there is a signi cant relation for TKA (p = 0.002) and hysterectomy (p = 0.002); Kruskal Wallis test.

Patient's characteristics impacting LOS
In univariate analysis for all data collected in the three surgical models gender, age and comorbidities were associated with LOS. Men had a longer average length of stay than women (p = 0.0062); the length of stay increases by an average of 0.1 days for an increase of one year of age; the length of stay was higher for high Charlson categories. In multivariate analysis on all the data, sex and age were associated with the length of stay. Regarding the results of analyses by surgical model, only age was associated with LOS after visceral surgery (colectomy and gastrectomy) gynaecological surgery (hysterectomy and ovariectomy). Life style and ASA score were not associated with LOS.

Structural and organizational evaluation with interview of professionals
In each centre, one anaesthetist and one surgeon were interviewed regarding their perception of ERP development in the speci c surgery analysed by the survey. The results in the three surgical models were similar with declaration of increased implication of health care providers in ERP between 2015 and 2017. However, compared to observed data in ERP application, the declarative information generally overestimates the level of performance. As an example the rate of speci c patient information on ERP was declared as 100% when maximal observed rate was 32% for TKA. The health care providers declare that they use speci c indicators to monitor ERP evaluation in 50% of cases in 2015 and 100% in 2017.
Despite institutional availability in all surgical and anaesthesiology departments, rate of GRACE survey tool was declared as persistently low (0% in 2015 and 14% in 2017). Frequently declared encountered di culties are lack of adherence to the ER concept and lack of multidisciplinary communication, limited manpower, speci cally ERP nurse coordinator. Organization of ERP was also limited by the di culty to organize admission on same day, lack of available time to start ERP and patients' reluctance.

Discussion
This survey in twenty-nine surgical departments located in seventeen different university hospitals of Assistance Publique Hôpitaux de Paris (APHP) describes an improvement in the development on ERP in orthopaedic and gynaecological surgery between 2015 and 2017 with correlation between ERP items application and reduction of LOS and a signi cant reduction of LOS for TKA surgery.

LOS and postoperative complications
We observe a signi cant reduction of LOS after TKA and a slight increase after THA and no modi cation in the other surgical models. The mean LOS in 2017 in our survey for TKA, colectomy and hysterectomy is much longer than the LOS reported presently in the literature (respectively 4 days for TKA and 2 days for colectomy) and underlines mainly the important room for improvement still existing in our institution (9,10). Interestingly in all individual data collected in our survey in 2015 and 2017, a signi cant correlation exists between the number of ERP items applied and LOS. After TKA and hysterectomy, LOS showed a negative relationship when more than 70% of ERP are applied. These results are in line with previous publication supporting that increased application of ERP protocols impact favourably LOS (11). Therefore the LOS reduction observed after TKA may partially re ect the bene cial impact of ERP items increased application. In multivariate analysis, other patient's related factors such as younger age and female gender seem associated with reduced LOS. Younger age in the 2017 population was observed only after TKA and it certainly participates in 2017 LOS reduction after TKA.
The general principle of ERP is to enhance recovery, reduce complications with potential related LOS reduction (12). In our survey the incidence of postoperative complications after TKA, colectomy and hysterectomy is comparable to what is described in the literature (9,(13)(14). We observe a reduction of complications only after THA with surprisingly a slight prolongation of LOS. Although the reduction of complications incidence might correlate with the positive evolution in ERP application observed after THA, as described in the literature (11), the disconnection with LOS rather suggest impact of other nonmedical factors such as unavailability of recovery institution for patients after surgery or patients unwilling to be discharged from hospital. After TKA the incidence of complications was similar but prolongation of LOS related to complications was reduced suggesting a reduced severity of post-surgical complications impacting positively LOS.

Evolution of ERP item application
Some ERP items were already largely applied in 2015 and were comforted in 2017 such as non-invasive surgery for colectomy and hysterectomy, no colic preparation for colectomy, prevention of PONV and hypothermia, antibioprophylaxis, thromboprophylaxis for the three surgical models and regional anaesthesia for pain control for TKA. Among these seven items, 2 are under the responsibility of surgeon (less invasive surgery and no colic preparation) and developed for a long time before ERP concept. The ve others are accessible to anaesthetist's prescription and relate to older speci c recommendations such as those of the French Society of Anaesthesia and Intensive Care on antibioprophylaxis (15), PONV (16), postoperative pain control (17), thrombosis (18). This anteriority may explain why some of those items were already largely applied in 2015.
We observe a trend to improvement in ERP application in all orthopaedic and gynaecological departments. Similar changes were not observed in visceral surgery where ERP application evolution appears heterogeneous although this particular surgical model has been targeted for ERP for a long time (7). The de nition of applied item of ERP (action traced in the patients' chart and percentage of application > 70%) can be considered as a high standard explaining potentially the nal low application rate (30-45%) of EPR items across the three surgeries in 2017. This level of appropriation has been suggested in single centre cohort study and con rmed in large national survey (7,11). This high ERP application is a guarantee of outcome improvement on adverse event and LOS (11). Our results are in line with the frequency of high ERP application observed in 21% of hospital in a large survey in USA (11). In addition, a larger number of ERP items progressed signi cantly between 2015 and 2017 re ecting extensive signi cant on-going progress in ERP application. This positive evolution probably explains the signi cant correlation in all individual data collected between ERP application and reduction of LOS.
When compared with TKA, colectomy and hysterectomy groups, no signi cant differences appear in the evolution of ERP application in control surgical groups. All together these results support the hypothesis of a common evolution in all involved centres of APHP institution potentially not related to the sensitization period in 2016.
Between 2015 and 2017 some items improved slightly or not at all and deserve further dissemination across all surgical models. Emerging pre-habilitation and preoperative information for TKA are still limited in 2017 (respectively 14.6% and 32.1%) and totally absent for colectomy and hysterectomy.
Although the bene t of pre-habilitation is still debated (19,20), preoperative patient information is crucial to prepare and follow up an ERP (21). The absence of dedicated personnel described in surgeon and anaesthesiologist interview may participate to this lack of preoperative information strategy (21). Modern fasting rules are slightly improving (20-40% in 2017 vs 0% in 2015), but carbohydrates which can signi cantly improve patient outcome after visceral surgery is still completely absent before colectomy (22). Although the absence of tourniquet for TKA surgery improved, it is still present for half of the patients (50% vs 17%). Early feeding on the day of surgery has progressed in 2017 (i.e. 98.4% for TKA, 86.4% for hysterectomy) but still half of patient do not bene t of this approach after colectomy. Some e cient techniques for pain control using local anaesthetic in 2017 are widespread in orthopaedic surgery but still rare for hysterectomy and slightly declining for colectomy. These analgesic techniques using regional anaesthesia have a positive impact of on ERP outcomes after colorectal surgery and should be used (23). Finally early mobilization which is widely mentioned in guidelines for ERP although its direct bene t on recovery and LOS is still debated (24), has developed but is still limited to half of patient on the day of surgery in 2017.
The reasons for these low implementation rates remain speculative and certainly multifactorial (5).
Firstly, knowledge and understanding of recommendations on ERP suppose rather French guidelines and during the period of our institutional survey French recommendations on ERP were proposed only by HAS in 2016 (4) and French Societies for colorectal surgery in 2014 (6)