Protocolized care pathways in emergency general surgery: a systematic review and meta-analysis

Abstract

Background

Emergency abdominal surgery is associated with significant postoperative morbidity and mortality. The delivery of standardized pathways in this setting may have the potential to transform clinical care and improve patient outcomes.

Methods

The OVID SP versions of MEDLINE, EMBASE and the Cochrane Central Register of Controlled Trials were searched between January 1950 and October 2022. All randomized and non-randomized cohort studies comparing protocolized care streams with standard care protocols in adult patients (>18 years old) undergoing major emergency abdominal surgery with 30-day follow-up data were included. Studies were excluded if they reported on standardized care protocols in the trauma or elective setting. Outcomes assessed included length of stay, 30-day postoperative morbidity, 30-day postoperative mortality and 30-day readmission and reoperations rates. Risk of bias was assessed using ROBINS-I for non-randomized studies and RoB-2 for randomized controlled trials. Meta-analysis was performed using random effects modelling.

Results

Seventeen studies including 20 927 patients were identified, with 12 359 patients undergoing protocolized care pathways and 8568 patients undergoing standard care pathways. Thirteen unique protocolized pathways were identified, with a median of eight components (range 6–15), with compliance of 24–100%. Protocolized care pathways were associated with a shorter hospital stay compared to standard care pathways (mean difference −2.47, 95% c.i. −4.01 to −0.93, P = 0.002). Protocolized care pathways had no impact on postoperative mortality (OR 0.87, 95% c.i. 0.41 to 1.87, P = 0.72). A reduction in specific postoperative complications was observed, including postoperative pneumonia (OR 0.42 95% c.i. 0.24 to 0.73, P = 0.002) and surgical site infection (OR 0.34, 95% c.i. 0.21 to 0.55, P < 0.001).

Discussion

Protocolized care pathways in the emergency setting currently lack standardization, with variable components and low compliance; however, despite this they are associated with short-term clinical benefits.

The delivery of postoperative care following emergency laparotomy is highly variable and significantly impacts postoperative outcomes. Protocolized postoperative pathways in emergency laparotomy are compared to standard care by meta-analysis to determine their impact on postoperative clinical outcomes. This meta-analysis identified 13 unique protocolized care pathways that were associated with a reduction in length of stay, morbidity and mortality.

Background

Major emergency abdominal surgery is often undertaken in a high-risk cohort of patients and is associated with significant 30-day postoperative morbidity and mortality, with reported rates of 14–47% and 10–20%^1–5 respectively. The Emergency Laparotomy Network Audit reported variation in postoperative mortality between 3.7% and 41% in 1853 patients across 37 UK hospitals⁶. Similar variation in postoperative mortality has been observed in the USA⁷. Underpinning these observed differences in emergency laparotomy outcomes are differences in processes and delivery of care^8–10. Consequently, there is increasing recognition for the standardization of care in this cohort of patients in a bid to improve the delivery of emergency care and patient outcomes. It is also essential to understand the current variation in clinical practice and care and its impact on outcomes, identifying care processes that are associated with superior outcomes, to drive standardization.

The delivery of standardized, protocolized streams of care in the emergency setting may have the potential to transform clinical care and improve patient outcomes. Protocolized care streams can include care bundles, standardized care pathways and enhanced recovery after surgery (ERAS) protocols. All of these aim to attenuate the perioperative stress response and accelerate postoperative functional recovery by employing a number of standardized component interventions across the perioperative pathway. Employing protocolized care pathways consisting of evidence-based interventions in the elective setting has been shown to be associated with reduced length of stay, reduced morbidity and an earlier return to activities of daily living¹¹. Consequently, there has been widespread adoption of ERAS protocols within elective surgery including in colorectal surgery¹¹, upper gastrointestinal surgery¹², gynaecology¹³, urology¹⁴ and orthopaedics¹⁵. There is now emerging interest in adopting protocolized care pathways in the emergency setting in an attempt to standardize the delivery of clinical care, reduce variation in hospital outcomes and to lead to patient benefit¹⁵. The aim of this systematic review and meta-analysis was to evaluate the impact of protocolized care pathways on postoperative outcomes in the emergency setting. A secondary aim was to understand the components of the pathways and report implementation.

Methods

This systematic review was conducted according to a prespecified protocol based on guidance from the Centre for Reviews and Dissemination¹⁶ and the Cochrane Handbook¹⁷ and is reported in line with the PRISMA statement¹⁸. Our protocol was registered with the international prospective register of systematic reviews, PROSPERO (CRD42020218481).

Eligibility criteria

All randomized and non-randomized cohort studies comparing protocolized care streams with standard care protocols in adult patients (>18 years old) undergoing major emergency abdominal surgery were included. All studies reporting follow-up data at 30 days were included. Studies that reported longer-term follow-up data were included if 30-day data could be easily extracted. Protocolized care pathways were defined as multimodal perioperative care bundles, perioperative protocols, dedicated clinical pathways or ERAS protocols comprising of several components, which were delivered to all patients in a standardized fashion within the individual healthcare system. Standard care was defined as postoperative care that was delivered, which was not predetermined or protocolized across the healthcare system. Studies were excluded if they reported on standardized care protocols in the trauma or elective setting or did not report on treatment-related outcomes of interest.

Search strategy

The OVID SP versions of MEDLINE (1950 to 31 October 2022), EMBASE (1980 to 31 October 2022) and the Cochrane Central Register of Controlled Trials were searched using the following search terms ‘emergency surgery’, ‘laparotomy ‘enhanced recovery’, ‘fast track’, ‘multimodal’, ‘care bundles’, ‘perioperative protocols’, ‘care pathways’ separated by the Boolean operator ‘AND’. Reference lists of included articles were hand-searched to identify any additional studies. All citations were collated within EndNote X7.8®, USA and duplicates were removed. All relevant titles and abstracts were screened independently by two reviewers (D.H. and B.G.). The full-text versions of potentially eligible abstracts were retrieved in full. Only studies that fulfilled all eligibility criteria were included. Any conflicts were resolved through discussion with a third party (J.M.B.).

Data extraction

Extraction of data for selected studies was conducted by one reviewer, and the accuracy of extracted data was verified by a second reviewer. Summary data were extracted under the broad categories of study design, patient and clinical characteristics, and reported outcomes. Data extracted on study design includes study type, country, number of participating centres and patients and details of the protocolized care pathway including component interventions and compliance. Available summary data extracted on patient and clinical characteristics included age, gender, ASA grade and indication for surgery and operation performed. Treatment-related outcomes include 30-day morbidity, mortality, readmission and reoperation and return to gastrointestinal function were extracted when reported by individual studies.

Outcome assessment

Outcomes assessed included length of stay, 30-day postoperative morbidity, 30-day postoperative mortality and 30-day readmission and reoperation rates. Hospital stay was defined as total hospital length of stay. Postoperative morbidity was defined as any reported complication occurring within 30 days of the index operation. Postoperative mortality was defined as death due to any cause within 30 days of the index operation. Reoperations were defined as a return to theatre within 30 days of the index operation. Readmission was defined as readmission to hospital within 30 days of the index operation. These are widely accepted generic definitions. If a study-specific definition employed a different definition but used the same time point, this definition was accepted as being sufficiently similar.

Study quality

Methodological quality assessment of included studies was undertaken using the ‘Risk of Bias In Non-Randomized Studies of Intervention’ (ROBINS-I) assessment tool¹⁹ for non-randomized studies and the Cochrane risk-of-bias tool for randomized trials (RoB 2) (RCTs) by both reviewers²⁰.

Data analysis

For studies with a comparative arm an exploratory meta-analysis of study summary data was undertaken using the software package RevMan (version 5.3 Copenhagen, 2014). Meta-analysis was separately performed on randomized clinical trials and observational studies, with at least three studies included per outcome. The odds ratio using 95% confidence intervals were calculated for binary outcome data and the standardized mean difference (SMD) with a 95% c.i. were calculated for continuous outcomes. For binary outcomes (for example, morbidity) the Mantel–Haenszel method was used for calculation of OR using the random-effects model²¹. For continuous outcomes (for example, length of stay) the inverse-variance method was used for combination of SMD. Heterogeneity was examined using the I² and χ² statistics and graphical exploration of the funnel plots. Higher values of I² and χ² statistics signified increasing levels of inconsistency and heterogeneity between studies, with P < 0.05 and a maximum I² value of 30% identifying low heterogeneity. Where standard deviations were not reported, these were estimated using ranges or P in accordance with the Cochrane collaborative guidelines²². Forest plots were derived for graphical representation.

Results

Seventeen studies were included in this systematic review, of which six were RCTs^23–28 and 11 observational studies: four retrospective cohort studies^29–32, one prospective cohort study³³, one propensity-matched cohort study³⁴, one case–control study³⁵ and four before-and-after studies^36–39 (Table 1, Fig. 1). A total of 20 927 patients undergoing emergency abdominal surgery were included, with 12 359 patients undergoing protocolized care pathways and 8568 patients undergoing standard care pathways. A total of 10 022 ASA I–II and 10 657 ASA III–V patients were included across all studies. Patient characteristics are outlined in Table 2.

Table 1.

Study characteristics

Author	Year	Years included	Country	Study type	Protocolized care pathway	Patient population	Total No. of patients
Shang	2018	2010–2017	China	Propensity-matched cohort study	ERAS	Obstructed colonic cancer	636
Mohsina	2018	2014–2016	India	Single-centre RCT	ERAS	Perforated Duodenal Ulcer	102
Gonenc	2014	2012–2013	Turkey	Single-centre RCT	ERAS	Laparoscopic Perforated Duodenal Ulcer	47
Liska	2019	2014–2017	USA	Retrospective cohort	ERAS	Non-elective colorectal surgery	404
Wisley	2015	2008–2012	Australia	Retrospective cohort	ERAS	Major emergency abdominal surgery	370
Vinas	2020	2011–2017	Spain	Prospective cohort	ERAS	Left colonic perforation	50
Saurabh	2020	2017–2018	India	Single-centre RCT	ERAS	Small bowel pathology	82
Lohsiriwat	2014	2011–2013	Thailand	Retrospective cohort	ERAS	Obstructed colorectal cancer	60
Tengburg	2017	–	Denmark	Case–control study	Perioperative protocol	Emergency abdominal surgery	1200
Vester-Andersen	2015	2010–2012	Denmark	Multicentre RCT	Intermediate Care	Emergency abdominal surgery	286
Aggarwal	2019	2016–2017	UK	Before-and-after study	Care bundle	Emergency abdominal surgery	14 809
Huddart	2014	2012–2013	UK	Before-and-after study	Care bundle	Emergency abdominal surgery	726
Jordan	2020	2014–2019	UK	Before-and-after study	Care bundle	Emergency abdominal surgery	930
Moller	2011	2008–2009	Denmark	Before-and-after study	Perioperative protocol	Perforated Peptic Ulcer	627
Trangbæk	2022	2015–2019	Denmark	Retrospective cohort	Care bundle	Emergency abdominal surgery	378
Pranavi	2022	2018–2020	India	Single-centre RCT	ERAS	Perforation of viscus with peritonitis	120
Sharma	2021	2019–2020	India	Single centre RCT	ERAS	Acute intestinal obstruction or intestinal perforation	100

Fig. 1.

Search strategy.

Table 2.

Patient characteristics

Author	PCP age	SCP age	Male PCP	Male SCP	Female PCP	Female SCP	ASA I–II PCP	ASA I–II SCP	ASA III–V PCP	ASA III–V SCP
Shang	Median 66	Median 65	192	201	126	117	216	220	102	98
Mohsina	Mean 45	Mean 44	44	44	6	5	50	49	0	0
Gonenc	Mean 35	Mean 37	5	6	16	20	21	26	0	0
Liska	Mean 49.2	Mean 51.3	66	122			31	55	104	214
Wisley	68	68	87	85			71	74	130	95
Vinas	Mean 54	Mean 58	21	17	8	4
Saurabh	Mean 45	Mean 43	27	22	8	13	35	35
Lohsiriwat	Mean 57	Mean 62	14	24	6	16
Tenburg	–	–	–	–	–	–	358	326	236	263
Vester-Andersen	Median 73	Median 73	66	75	78	67	78	79	66	63
Aggarwal	Mean 65	Mean 65.3	4336	2625	4911	2937	4375	2542	4872	3020
Huddart	Mean 65.8	Mean 65.6	202	138	225	161	186	122	235	177
Jordan	Mean 64	Mean 65	–	–	–	–	340	64	437	89
Moller	Median 75	Median 69	57	239	60	271	55	270	62	240
Trangbæk	Median 71	Median 69	51	124	69	134	72	195	48	63
Pranavi	Mean 45	Mean 47	53	51	8	8	40	37	21	22
Sharma	Median 33	Median 45	38	37	12	13

PCP, protocolised care pathway; SCP, standard care pathway.

Study bias

The overall risk of bias was high in one of six RCTs, with three RCTs considered as having some concerns and two considered low risk (Supplementary material). Areas of concern for bias were outcome measurement, deviation from intended interventions and randomization allocation. For observational studies, 2 of 11 studies were identified as being seriously biased due to issues with participant selection and classification of interventions, with the remaining 9 studies classified as being moderately biased.

Protocolized care pathways

Protocolized care pathways were defined in different ways, with 10 studies reporting on emergency ERAS protocols, 4 reporting on care bundles, 2 reporting on the implementation of a perioperative protocol and 1 study defined its care pathway as intermediate care. We identified 13 unique pathways. Two studies had similar ERAS protocols; both these RCTs were performed in the same unit investigating emergency ERAS protocols in perforated duodenal ulcer²⁴ and small bowel obstruction²³. Three studies employing perioperative care bundles consisted of the same key components^36–38.

Fifteen studies reported on the components of their protocolized care pathway (Supplementary material). Protocols consisted of a median of 8 components, with protocols including between 6 and 15 components. The most common preoperative component was antibiotic prophylaxis, with 14 (93.3%) studies incorporating this. The delivery of goal-directed fluid therapy was the most commonly delivered intraoperative intervention, with nine (60.0%) studies incorporating this component. The most common postoperative component was early oral intake, with 11 (73.3%) studies incorporating this component.

Although there was consistency in the classification of components delivered across individual ERAS programmes, there was significant variation in the delivery of the individual components. Ten studies reported early mobilization as part of their protocolized care pathways (Supplementary material); however, they specified different interventions. Similarly, eight studies reported a variety of different timings for the resumption of early oral intake. The use of nutritional supplements, chewing gum to stimulate gastrointestinal motility, incentive spirometry and chest physiotherapy were not common components of emergency protocolized care pathways.

Compliance to emergency protocolized care pathways were reported by nine studies (Supplementary material). These studies reported compliance rates of 24–100% to individual components of emergency ERAS programmes. None of the studies reported overall compliance to the care pathways. Preoperative components of risk stratification, senior clinical input, antibiotic prophylaxis, thromboembolism prophylaxis, postoperative nausea and vomiting (PONV) prophylaxis, no premedication and glycaemic control had high adherence. Goal-directed fluid therapy had a compliance rate of 38–84.6% across the studies as an intraoperative intervention. There was poor compliance across the majority of postoperative components, with Vinas et al. reporting compliance rates of 42.3–79.3% across all six of the postoperative interventions making up the care pathway employed in this study³³. Liska et al. reported compliance to individual components of their non-opioid multimodal analgesic strategy, reporting a low compliance rate of 17% with the use of patient-controlled analgesia and a high compliance rate of 90.4% with postoperative acetaminophen use²⁹.

Outcomes

Length of stay

Four RCTs reported mean length of hospital stay across 316 (1.51%) patients. Protocolized care pathways were associated with a significantly shorter length of stay compared to standard care pathways (SMD −2.47, 95% c.i. −4.01 to −0.93, P = 0.002; Fig. 2). There was significant heterogeneity between the studies included (χ² = 86.66, df = 4 (P = 0.0001); I² = 97%).

Fig. 2.

Pooled meta-analysis of RCTs forest plot length of stay.

Mortality

Four RCTs reported 30-day postoperative mortality across 535 (2.5%) patients. There was no significance difference in postoperative mortality compared to standard care pathways (OR 0.87, 95% c.i. 0.41 to 1.87, P = 0.72; Fig. 3a). There was low heterogeneity observed between the studies included (τ² = 0.00; χ² = 0.25, df = 2 (P = 0.88); I² = 0%). On assessing 90-day mortality data, through the inclusion of data reported by the Enhanced Peri-Operative Care for High-risk patients (EPOCH) trial, five RCTs, including 16 391 patients, did not demonstrate any difference in postoperative mortality (OR 0.97, 95% c.i. 0.90 to 1.05, P = 0.42).

Fig. 3.

a. Pooled meta-analysis of RCTs forest plot overall mortality. b Pooled meta-analysis of cohort studies forest plot mortality

M-H, Mantel–Haenszel.

Ten cohort studies reported 30-day postoperative mortality across 20 132 (96.2%) patients. These cohort studies reported lower 30-day postoperative mortality associated with protocolized care pathways compared to standard care pathways (OR 0.78, 95% c.i. 0.71–0.86, P < 0.001; Fig. 3b). Low heterogeneity was observed (τ² = 0.00; χ² = 7.39, df = 9 (P = 0.60); I² = 0%).

Morbidity

Overall, 10 studies reported 30-day postoperative morbidity across 3513 (16.7%) patients, including 3 RCTs in 415 (1.9%) patients and 7 cohort studies in 3098 (14.8%) patients. On meta-analysis of the RCTs, protocolized pathways provided no benefit over standardized care pathways in reducing postoperative morbidity (OR 1.28, 95% c.i. 0.84 to 1.95, P = 0.260; Supplementary material). There was low heterogeneity observed between the studies included (τ² = 0.00; χ² = 0.42, df = 2 (P = 0.81); I² = 0%). Meta-analysis of the seven cohort studies identified lower postoperative morbidity associated with protocolized care pathways compared to standard care pathways (OR 0.76, 95% c.i. 0.65 to 0.88, P = 0.0003). There was low heterogeneity observed between the studies included (τ² = 0.00; χ² = 4.01, df = 6 (P = 0.68); I² = 0%).

Five RCTs reported postoperative pneumonia across 451 patients (2.1%). Protocolized care pathways were associated with a lower rate of postoperative pneumonia (OR 0.42 95% c.i. 0.24 to 0.73, P = 0.002; Supplementary material). There was low heterogeneity observed between the studies included (τ² = 0.00; χ² = 2.03, df = 4 (P = 0.73); I² = 0%).

Five RCTs reported surgical site infection (SSI) across 451 (2.1%) patients. Patients participating in protocolized care pathways had significantly fewer SSIs compared to those undergoing standard care (OR 0.34, 95% c.i. 0.21 to 0.55, P < 0.001; Supplementary material). The observed heterogeneity between studies was considered to be low (τ² = 0.05; χ² = 4.76, df = 6 (P = 0.31); I² = 16%).

Reoperation

Eight studies, including four RCTs, reported 30-day reoperation rates across 2367 (11.31%) patients. On meta-analysis of the RCTs, there were no observed differences between protocolized care pathways and standard care pathways (OR 1.45, 95% c.i. 0.31 to 6.71, P = 0.640; Supplementary material). There was moderate heterogeneity observed between studies (τ² = 0.66; χ² = 3.09, df = 2 (P = 0.21); I² = 35%). Similarly, the meta-analysis of the cohort studies did not observe any differences in reoperation rates between protocolized pathways and standard care pathways (OR 1.18, 95% c.i. 0.64 to 2.17, P = 0.60). There was moderate heterogeneity observed between studies (τ² = 0.20; χ² = 6.56, df = 3 (P = 0.09); I² = 54%).

Readmission

Ten studies, including five RCTs, reported 30-day readmission rates across 1971 (9.41%) patients. Five RCTs, including 451 patients (2.10%), demonstrated no observed differences between protocolized care pathways and standard care pathways in readmission rates (OR 2.09, 95% c.i. 0.58 to 7.57, P = 0.260; Supplementary material). There was low heterogeneity observed between the studies included (τ² = 0.00; χ² = 0.22, df = 1 (P = 0.64); I² = 0%). Five cohort studies including 1520 (7.31%) patients had similar results, with no observed differences between the two pathways with regards to readmission rates (OR 1.10, 95%, c.i. 0.79–1.55, P = 0.570). There was heterogeneity observed between the studies included (τ² = 0.00; χ² = 0.95, df = 3 (P = 0.57); I² = 0%).

Gastrointestinal function

Time to flatus

Four RCTs reported time to passage of flatus across 389 (1.85%) patients. Time to passage of flatus was shorter in patients undergoing the protocolized care pathways compared to patients undergoing standard care pathways, with an SMD of −2.45 days (95% c.i. −3.77 to −1.14, P < 0.001; Supplementary material). There was high heterogeneity observed between the studies included (τ² = 1.68, χ² = 82.01, df = 5 (P < 0.00001); I² = 96%).

Time to defaecation

Three RCTs reported time to defaecation across 289 (1.38%) patients. Time to defaecation was shorter in patients in the protocolized care pathways arm compared to patients in the standard pathway arm, with an SMD of −4.01 (95% c.i. −6.36 to −1.66, P = 0.001; Supplementary material). There was significant heterogeneity observed between the studies included (τ² = 4.06, χ² = 87.81, df = 2 (P = 0.001); I² = 98%).

Time to diet

Three RCTs reported time to liquids across 289 (1.38%) patients. Time to liquids was shorter in patients undergoing protocolized care pathways compared to the standard pathway arm, with an SMD of −4.18 (95% c.i. −6.58 to −1.77, P < 0.0007; Supplementary material). There was significant heterogeneity observed between the studies included (τ² = 4.17, χ² = 102.73, df = 3 (P < 0.001); I² = 98%).

Four RCTs reported time to solid diet across 336 (1.60%) patients. Time to oral diet was shorter in patients undergoing protocolized care pathways compared to standard pathway arm, with an SMD of −4.20 (95% c.i. −6.21 to −2.19, P < 0.001; Supplementary material). There was significant heterogeneity observed between the studies included (τ² = 3.74, χ² = 128.83, df = 4 (P = 0.001); I² = 98%).

Postoperative nausea and vomiting

Three RCTs reported outcomes on PONV in 269 (1.28%) patients. There were fewer symptoms of PONV for patients undergoing protocolized care pathways compared to standard care (OR 0.24, 95% c.i. 0.09 to 0.65, P = 0.005; Supplementary material). Significant heterogeneity was observed between the studies included (τ² = 0.34, χ² = 3.69, df = 2 (P = 0.16); I² = 46%).

Postoperative ileus

Three RCTs reported the incidence of postoperative ileus across 216 (1.03%) patients. No significant differences were observed in the rate of postoperative ileus between protocolized care pathways and standard care pathways (OR 0.89, 95% c.i. 0.37 to 2.14, P = 0.790; Supplementary material). There was low heterogeneity observed between the studies included (τ² = 0.00; χ² = 1.38, df = 2 (P = 0.50); I² = 0%).

Discussion

Protocolized care pathways in the emergency setting appear to be associated with better postoperative recovery, including shorter length of stay, a reduction in specific complications of pneumonia and SSI, and improved gastrointestinal function compared to standard pathways. Reoperation and readmission rates and rates of postoperative ileus appear to be similar. Outcomes were examined across 10 different types of protocolized care pathways consisting of between 6 and 15 component interventions, with several key common components across multiple pathways, albeit delivered in a variety of ways.

Previous works have reported improved clinical outcomes associated with ERAS programmes in the emergency setting^40–42. In a meta-analysis of six studies consisting of 1334 patients undergoing emergency abdominal surgery, Hajibandeh et al. reported reduced length of stay and postoperative complications and improved gastrointestinal function associated with emergency ERAS programmes⁴⁰. However, there has been no documentation of pathway design. The present systematic review assesses both pathway design and compliance. The design of the majority of protocolized pathways included in this systematic review shared several key components, including senior clinical input, antibiotic prophylaxis, goal-directed fluid therapy, early removal of nasogastric tube, early oral intake, early mobilization and early removal of urinary catheter. The compliance to these key components is variable, ranging from 17% to 100%. Most of these key component interventions were adapted from existing elective pathways, with no interventions specifically developed for the emergency setting. Contextually relevant pathway design coupled with high intervention fidelity is central to ensuring a successful and effective protocolized pathway, as demonstrated by the EPOCH trial⁴³. The failure of the EPOCH trial to demonstrate clinical effectiveness of protocolized care pathways was largely due to implementation failure of individual component interventions. Contextually relevant and realistic pathway design of high-fidelity, easily implementable interventions are likely to make the most clinical impact in the emergency setting⁴⁴.

This review demonstrates the potential benefits of a protocolized care pathway across 20 927 patients in the emergency setting across a range of pathway designs with regards to key short-term clinical outcomes including length of stay, reduction in complications such as pneumonia and SSI, and improved gastrointestinal function. It is likely these improvements have been observed in the protocolized pathway, due to the specific interventions delivered, for example, incentive spirometry and physiotherapy. However, it is important to note that although there is an overall improvement in gastrointestinal function, due to interventions such as early removal of nasogastric tube and early oral intake, there is no improvement in the incidence of postoperative ileus between the two groups. This is likely to reflect interventions that are targeted at resumption of diet and nutrition, as opposed to prevention of postoperative ileus. There are few high-quality data to suggest early resumption of oral intake is associated with reduced postoperative ileus⁴⁵. Furthermore, there are no interventions incorporated into these pathways that are targeted specifically for the prevention of postoperative ileus. It is important to ensure high-quality, evidence-based interventions are incorporated into protocolized pathways, which help mitigate for the significant complication profile of emergency laparotomy, which includes postoperative ileus.

However, a key limitation of this work is the differential number of patients across different clinical outcomes, with mortality data on 20 622 patients compared to data available on gastrointestinal function in 389 patients alone. The differential patient populations within RCTs and observational studies are also likely to contribute to the conflicting results regarding impact on key outcomes such as mortality and morbidity. Four RCTs consisting of 535 patients failed to demonstrate any impact of protocolized pathways on 30-day postoperative mortality (OR 0.87, P = 0.720), compared to 10 cohort studies consisting of 20 132 patients that demonstrated lower postoperative mortality in patients undergoing protocolized care (OR 0.78, P < 0.001). Similar, conflicting results were demonstrated, with regard to morbidity, with three RCTs consisting of 415 patients demonstrating no impact (OR 1.28, P = 0.260) and seven cohort studies consisting of 3098 patients demonstrating a reduction in postoperative morbidity (OR 0.76, P = 0.0003). It is likely this is due to the underpowering of these key clinical outcomes.

There is a lack of available data on outcomes prioritized by patients such as quality of life and quality of recovery⁴⁶, and by broader stakeholders and decision-makers, such as cost-effectiveness. There is also a lack of data on compliance to protocolized care pathways, which provides limited insights into the successful and unsuccessful implementation of overall pathways and individual component interventions. This systematic review focuses on short-term outcomes, with the consolidating of clinical outcomes at the same postoperative time point, and therefore excluded the EPOCH trial, which assessed longer-term outcomes at 90 and 180 days post-operation. The inclusion of EPOCH into our systematic review would have increased our sample size significantly, given that 15 873 participated in this trial; however, it would have contaminated our results based on the timing of its primary and secondary outcomes. Overall, the current evidence base underpinning protocolized pathways in the emergency laparotomy setting is limited by the variable study quality, size and outcome reporting.

Our systematic review demonstrates the potential short-term benefits of protocolized care pathways in the emergency setting, with reported improved postoperative mortality, reduced postoperative complications and improvements in postoperative recovery, with shorter length of stay and improved gastrointestinal function. However, these results should be interpreted carefully within the context of several methodological limitations. To enhance the benefits and improve implementation of these pathways requires consistent and contextually relevant and standardized pathway design and the development of interventions specific to the emergency setting. Emergency laparotomy pathways need to have high fidelity and need to be easy to implement—these two aspects must be appropriately evaluated to ensure success. Coupled with this, there needs to be consistent outcome reporting, including patient-centred outcomes, to truly elucidate the value and benefit of protocolized care in the emergency laparotomy setting.

Author contributions

Deena Harji (Conceptualization, Data curation, Formal analysis, Methodology), Ben Griffiths (Data curation, Formal analysis, Methodology, Writing—review & editing), Deborah Stocken (Conceptualization, Formal analysis, Supervision, Writing—review & editing), Rupert Pearse (Conceptualization, Supervision, Visualization, Writing—review & editing), Jane Blazeby (Conceptualization, Methodology, Supervision, Writing—review & editing), and Julia Brown (Conceptualization, Supervision, Visualization, Writing—review & editing)

Funding

This work was funded through a National Institute of Health and Care Research.

Conflict of interest

There no conflicts of interest or financial disclosures to disclose by any of the authors.

Supplementary material

Supplementary material is available at BJS online.

Data access, responsibility and analysis

D.H., B.G., D.S. and J.M.B. had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.