Weekend effect in emergency laparotomy: a propensity score‐matched analysis

The ‘weekend effect’ is the term given to the observed discrepancy regarding patient care and outcomes on weekends compared to weekdays. This study aimed to determine whether the weekend effect exists within Aotearoa New Zealand (AoNZ) for patients undergoing emergency laparotomy (EL), given recent advances in management of EL patients.


Introduction
Emergency laparotomy (EL) refers to any emergent open abdominal surgery, and commonly occurs in the context of a patient presenting to the emergency department with an acute abdomen.
Emergency laparotomy is associated with 30-day morality rate of nearly 15% for all patients, which increases to almost 25% for those aged over 80. 1,2 In a bid to improve EL outcomes, it is important to identify significant factors related to high morbidity and mortality rates.
One potential factor influencing EL-outcomes is the so-called 'weekend effect'. The weekend effect describes a rise in the mortality rates and complication rates of patients admitted to hospital over the weekend, compared to patients who present during weekdays. Many studies have observed significantly poorer patient outcomes over the weekend for a range of different medical conditions, including myocardial infarction, pulmonary embolism, acute abdomen, and emergency surgery. [3][4][5][6][7][8] Possible explanations for the weekend effect can be classified as either patient factors (such as delay in presentation, illness severity, age, ethnicity and underlying pathology) 9,10 or hospital characteristics (in particular, staffing deficits and more junior clinicians). 3,[11][12][13][14][15] A 2017 meta-analysis 16 proved inconclusive in delineating a specific aetiology. Understanding the mechanisms underlying the weekend effect is important for improving healthcare systems and patient outcomes. Of particular concern is the impact of the weekend effect on emergent surgery such as EL. So far, there is equivocal evidence regarding the existence of the weekend effect for EL. 17,18 In a bid to improve patient outcomes post-EL within Aotearoa New Zealand (AoNZ) specifically, it is important to determine whether a weekend effect exists and if so, the factors which may contribute to this phenomenon. The aim of this study was to investigate the relationship between mortality and post-operative complications, and the day of the week during which the operation was performed. We hypothesise that the present model of acute surgical care at weekends in AoNZ may obviate the weekend effect.

Participants
We studied adults (aged >18) who underwent emergency laparotomy in AoNZ between November 2017 and November 2021.

Weekend study arms
We retrospectively defined the study cohorts according to the day of the week during which the acute laparotomy was performed. The 'weekend group' was defined as those who underwent surgery on either a Saturday or Sunday. The 'weekday group' underwent surgery between Monday to Friday inclusive.

Ethics approval
Approval for this study was granted by The Health and Disability Ethics Committee (HDEC) of AoNZ on 18/08/2017. In addition, ethics approval was sought from local institutions where data was collected. This study has been conducted in accordance with the standards set by the Declaration of Helsinki. 19 Surgical procedure All EL were performed by consultant surgeons or senior registrars under supervision.

Data collection
Clinical and surgical data was collated from a large multi-centre AoNZ database-'Risk Estimation for Acute Laparotomy (REAL), for which recruitment occurred at five hospitals in AoNZ (Middlemore, Auckland City, North Shore, Waikato, and Christchurch). For the purposes of data collection, emergency laparotomies were defined as any type of open abdominal incision considered a general surgical procedure. Procedures were excluded if classified as elective (NCEPOD classification Code 4); exclusively diagnostic; vascular surgery; obstetric or gynaecological laparotomies; caesarean sections; hernia repairs; appendicectomies; cholecystectomies; organ transplants; colostomy or ileostomy procedures; or were performed in the context of specific underlying pathological diseases (ruptured ectopic pregnancies, pelvic abscesses due to pelvic inflammatory disease, blunt or penetrating trauma and sclerosing peritonitis). The definitive list of included vs. excluded procedures aligns with the methodology of the National Emergency Laparotomy Audit, which is a database for perioperative management within the National Health Service (England and Wales). 20

Definition of outcomes
The primary outcome of interest was 30-day mortality rates. Secondary outcomes included post-operative complications, and postoperative requirements for critical care support.

Propensity scores matched analysis
A propensity score matched analysis was conducted using R software to remove confounding factors for the incidence of peri and post-operative events.
The propensity score of each patient was assigned by a logistic regression model based on morbid obesity, defined as BMI (>35 or less than 35 kg/m 2 ), anticoagulation status, ECOG and the presence of ischaemic heart disease, congestive heart failure and chronic obstructive pulmonary disease. Patients in the weekday and weekend groups were matched using a 1:1 ratio using the nearest propensity score on the logit scale. All data were compared between the matched and unmatched groups.
Age and gender were not included in the propensity score analysis, as they were already well matched at baseline (Table 1).

Statistical testing
Statistical analyses were performed by using the Chi squared or Fisher's exact T test. Death was defined as any mortality occurring during the first 30 days post-operatively. All P values <0.05 were recognized as statistically significant. All statistical analyses were conducted using R software.

Results
The study included 487 patients. The 'weekend group' included a total of 132 participants; the 'weekday group' included 355 participants. Patient characteristics are summarized in Table 1.
Primary and secondary outcomes are shown in Table 2. There were no significant differences between weekend or weekday mortality rates both before and after matching (before matching: P = 0.430, after matching: P = 464) The only significant difference (P = 0.003) was between planned re-operation rates following weekend vs. weekday procedures.

Discussion
This study collated data from five public hospitals in AoNZ over 4 years. It found that patients undergoing emergency laparotomy had similar postoperative mortality whether operated on during the weekday or weekend. We propose that the lack of a weekend effect in AoNZ is a testament to the recent national efforts to provide consistent care across the country at all hours.
In 2010, General Surgeons Australia published a 12-point plan to guide safer and more consistent emergency laparotomy care within the acute surgical unit (ASU). Based on these recommendations, a model of care was then adopted within a high-volume AoNZ institution, with favourable clinical outcomes. 21,22 Since becoming widely implemented across AoNZ healthcare centres, there has been ongoing evidence that it has helped shorten emergency wait-times, has allowed earlier recognition and management of sepsis, and has promoted protocol-driven care for many pathologies that ultimately require acute emergency laparotomy. 23 Our results are therefore in keeping with the overall positive outcomes of this new framework of care. Unfortunately, due to the lack of data predating the implementation of the 12-step plan, it is difficult to directly conclude that it has negated the weekend effect. However, given the evidence supporting its outcomes, we propose that the lack of a weekend effect within AoNZ healthcare Age and gender were not included in the propensity matched score.
Abbreviations: BMI, body mass index; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; F, female; IHD, ischemic heart disease; M, male; N, no; Y, yes. system is at least partially attributable to the successful roll-out of the 12-step plan.
In considering alternative reasons for our results, we note that the absence of a weekend effect for mortality post-EL in this AoNZ cohort study is concordant with several UK-based studies including the National Emergency Laparotomy Audit (NELA) that have been published within the last 7 years. 18,24,25 Nageswaran et al. 18 suggested that a lack of the weekend effect in their cohort may be a reflection of the nature of EL as a more critical procedure that is prioritized by medical staff. Furthermore, advances in medical technology over the years may have independently improved surgical outcomes and reduced the gap between weekends and weekdays.
Our comparative study demonstrated that reoperation rates are significantly higher following weekend AL compared to weekday surgery, both planned and unplanned. We posit that planned reoperations following weekend surgery may be in part due to limited subspeciality surgeon availability on the weekends. Therefore, general surgeons may find themselves in instances where damagecontrol emergency laparotomies 26 are conducted on the weekend with a view toward definitive surgery occurring the next available weekday. Rostering of a consultant surgeon at all times is protective for emergency general surgical mortality rates, 27 as subspecialists are more equipped to tackle complex acute abdominal presentations within their field. Data from overseas has indicated that specific general surgical presentations may benefit from specialized surgeons performing the procedure. 28,29 If there are a limited number of subspecialists available on-call over the weekend, then patients presenting over the weekend who require specific procedures may be more likely to receive planned reoperations during a weekday. Higher rates of unplanned reoperation rates may be due to more junior staffing and thus lesser experience by the operator over the weekend.
Our study also aimed to investigate whether there were significant differences in the indications for EL on the weekend. We sought to investigate this by comparing the comorbidities of weekday and weekend EL patients; thus, investigating whether there were discrepancies in surgeons' propensity to operate based upon the day of the week. The present study also found that there were no significant differences between baseline patient characteristics. Previous papers have suggested that there is a patient selection bias due to weekend presentations more likely to be complicated by more extensive medical comorbidities or be considered sicker, [30][31][32][33] thus resulting in more complex procedures and potentially worse outcomes over the weekend. However, there were no statistically significant differences between patient cohorts. Thus, we conclude that in our study there were no differences between weekday and weekend in surgeons' indications to proceed to EL based upon patients' comorbidities.
The REAL database forms one of the most comprehensive datasets available on the quality of perioperative care for emergency laparotomy in AoNZ. We were able to successfully approximate randomisation through our use of propensity score matching, however our study's main limitation is derived from its use of observational data. The nature of acute emergency laparotomy makes randomized controlled trial (RCT) difficult. Our study generates hypotheses regarding resource allocation in relation to the working week but further studies are needed to corroborate these findings. A further limitation is that our study only included those patients who underwent surgery. Details of patients who did not undergo an operation were not added to the database and therefore outcomes for this cohort of patients remains unaccounted for. Regarding generalisability of our results, this study utilizes data from five large metropolitan hospitals. There are likely differences between outcomes at rural versus metropolitan centres, owing to staffing and resource inequities. Without clear data from rural and regional centres within AoNZ, the results of our study may be less relevant to these regions and thus further investigation is required to determine whether the weekend effect exists in rural AoNZ.
In summary, our study demonstrates that there is no weekend effect with regards to mortality rates within metropolitan AoNZ. This is likely successful reflection of the present model of acute surgical care at weekends in AoNZ. this article. They were excluded from editorial decision-making related to the acceptance and publication of this article. Editorial decision-making was handled independently by to minimize bias.