Potentially‐Preventable Deaths Following Colorectal Surgery: A National Retrospective Cohort Study of Surgical Mortality Authors: Cameron I Wells, Scott J P McLaughlin, Matthew J McGuiness, Maximillian O Joret, Wasilwa Baraza, Greg O'Grady, Ian P Bissett Published: 2026 ABSTRACT ABSTRACTBackgroundMany deaths after colorectal resection may be preventable, but administrative datasets provide limited insights into the clinical processes leading to mortality. We used the Australian and New Zealand Audit of Surgical Mortality (ANZASM) to quantify potentially preventable deaths after colorectal surgery and describe modifiable clinical management issues (CMI) in contemporary Australasian practice.MethodsRetrospective cohort study of prospectively collected ANZASM data. Adults who died after elective or emergency colorectal resection between 2010 and 2023 were included. Potentially preventable mortality was defined as at least one CMI assessed as definitely or potentially preventable and contributing to death. CMIs were summarised by phase of care and thematically analysed.ResultsAmong 1058 deaths after colorectal resection, 321 (30.3%) were classified as potentially preventable. Potentially preventable deaths were more frequent after elective than emergency surgery (35.3 vs. 27.8%, p = 0.01). Leading causes of death were sepsis, multiorgan failure and aspiration/pneumonia. Anastomotic leak was the most common postoperative complication, implicated in 19.6% of deaths. Preventable mortality was independently associated with lower estimated operative risk, postoperative complications, unplanned ICU admission and rural hospital care. Common themes included suboptimal patient selection, physiological optimisation, anastomotic decision‐making and delayed recognition or escalation of deterioration.ConclusionsNearly one third of audited deaths after colorectal surgery were potentially preventable, most arising from failures of selection, optimisation or rescue rather than technical errors. Strengthening preoperative triage, multidisciplinary decision‐making and improving postoperative surveillance and escalation systems offer the greatest opportunities to reduce preventable mortality in colorectal surgery.Keywords: colorectal surgery, hospitals, mortality, patient safety, postoperative complications Introduction Despite improvements in surgical technique and perioperative care, colorectal resection remains associated with substantial rates of complications and mortality [1, 2]. Contemporary Australasian data report 90‐day mortality rates of up to 4.7% following colorectal cancer resection, increasing to over 10% for emergency cases [1, 3]. Given the high incidence of colorectal cancer and the frequency of emergency colorectal operations, these deaths represent a substantial public health burden. Hospital variation in mortality following colorectal surgery is driven by differences in ‘failure to rescue’ (FTR), the risk of death following a complication, rather than differences in the complication rate [1, 4, 5]. FTR reflects institutional capacity in the recognition and management of complications [6], with delays in the diagnosis and treatment of anastomotic leaks and other complications consistently identified as key contributors to FTR [3, 7, 8]. Such variation implies many deaths are potentially preventable, influenced by system, team and decision‐making factors [9]. Population‐based registries and administrative datasets can quantify variation in outcomes, but provide limited insight into the clinical processes underlying preventable mortality. In contrast, the Australian and New Zealand Audit of Surgical Mortality (ANZASM) captures detailed clinical peer‐review assessments of every in‐hospital surgical death in Australia [10]. Prior ANZASM studies have examined elective hepatopancreatic, oesophagogastric and renal surgery [11, 12, 13, 14], identifying systems‐level perioperative processes amendable to improvement and highlighting the importance of non‐technical skills [15]. In contrast, colorectal resection is frequently urgent or emergent, often performed by non‐colorectal surgeons and is not easily centralised. Colorectal surgery represents a high‐volume, high‐risk field that spans both elective and emergency practice, making it an ideal context to examine the clinical and system factors underlying preventable surgical deaths. Therefore, this study aimed to examine the causes of death following emergency and elective colorectal resection, quantify the proportion of deaths considered potentially preventable and identify modifiable clinical management issues (CMIs) underlying these deaths. Methods This national retrospective cohort study analysed prospectively collected data from the ANZASM database. All patients who died following colorectal resection for any indication, with or without formation of an ileostomy or colostomy, between 1 January 2010 and 31 December 2023 were included. Patients were excluded if they did not undergo resectional surgery (i.e., minor colorectal procedures or bypass), were aged < 18 years at the time of surgery, were not primarily under the care of a General or Colorectal Surgeon, lacked assessor commentary or had > 5% missing data. The study was approved by the ANZASM Committee and Auckland Health Research Ethics Committee (reference AH29043), and is reported according to the STROBE guidelines (Supporting Information Appendix 1) [16]. ANZASM Process ANZASM is a national, independent, peer‐reviewed audit of all inpatient surgical deaths in Australia, directed by the Royal Australasian College of Surgeons (RACS), to support quality improvement rather than assign blame or determine negligence [10]. Since 2010, all public hospitals and most private hospitals have participated in ANZASM, with the exception of New South Wales, which report separately to the Collaborating Hospitals Audit of Surgical Mortality. For each death, the treating surgeon completes a standardised case form including demographics, comorbidities, perioperative risk assessment and details of peri‐operative management. The cause of death is assigned by the surgeon, though a coroner's evaluation with post‐mortem examination may occur when uncertain. An independent first‐line peer reviewer (from the same specialty, but different hospital) evaluates each case and categorises CMIs by severity using a structured matrix for impact and preventability. CMIs are graded as: (1) ‘Area for consideration’: an aspect of care could have been better (2) ‘Area of concern’: an aspect of care should have been better or (3) ‘Adverse event’: an aspect of care contributed to or caused death. For all CMIs, both their impact on death and potential preventability are graded by the reviewer and assigned using a consistent framework. Where required, a second‐line full case‐note review is undertaken using a similar process. Study Endpoints and Definitions The primary endpoint was the proportion of potentially preventable deaths, as judged by the ANZASM review process. Following the audit framework, a death was classified as potentially preventable when at least one CMI was judged ‘definitely’ or ‘probably’ preventable and assessed as ‘causative’ or ‘may have contributed to’ the death [12]. Deaths were considered non‐preventable if no CMI was identified, if CMIs were ‘probably not’ or ‘definitely not’ preventable or if CMIs made ‘no difference’ to mortality. When both assessors identified the same CMI, it was counted once; discrepant gradings were resolved by assigning the higher severity category. Secondary endpoints included the distribution of causes of death, the spectrum and frequency of complications and cofactors preceding death and the burden and phase of care (preoperative, intraoperative, postoperative) for CMIs. Thematic Analysis CMIs were categorised by the phase of care, and an inductive, data‐driven thematic analysis using the methods of Braun and Clarke was used to classify CMIs within overarching themes for each phase [17]. Four researchers (C.I.W., S.J.P.M., M.J.M. and M.O.J.), all of whom are experienced general surgery trainees and academic researchers, independently coded CMIs and resolved differences by discussion with a senior author when necessary. Coding agreement was reached by consensus. Statistical Analysis Analyses were conducted using R (version 4.4.3, R Foundation for Statistical Computing, Vienna, Austria). Categorical variables were summarised as counts (percentages) and continuous variables as mean (standard deviation [SD]) or median (interquartile range [IQR]) as appropriate. Group comparisons used chi squared tests for categorical data, and t‐tests and Mann–Whitney U tests depending on distribution. To examine clinical factors associated with potentially preventable mortality, factors with p < 0.15 on univariate analysis were input into a logistic regression model. A two‐tailed p < 0.05 was considered statistically significant. Missing data were excluded by complete‐case analysis. Given the major differences in presentation and care for emergency and elective colorectal surgery, we undertook subgroup analyses for these two groups. Results From 2010 to 2023, ANZASM recorded 1493 surgical deaths following colorectal surgery. After exclusions for minor procedures (n = 64), stoma formation without concurrent resection (n = 183), palliative bypass (n = 130) and non‐General/Colorectal surgeon care (n = 58), 1058 cases were included in the analysis. Demographics are reported in Table 1. Right‐sided resections (32.0%) and anterior resection (24.0%) were the commonest procedures performed. Emergency operations (64.9%) were more frequent than elective surgery (35.1%). Patients undergoing emergency surgery were slightly younger; however, they had higher ASA grades, were more likely to have surgery in public hospitals, at rural centres, with benign indications for surgery (Table 1). Emergency cases also had a much higher likelihood of having a preoperative judgement of an ‘expected’ or ‘considerable’ risk of death. TABLE 1Demographic and clinical factors of the included elective and emergency cohorts of mortality following colorectal surgery.VariableElective (n = 363)Emergency (n = 695)Total (n = 1058) p Age (years)Mean (SD)76.7 (10.3)74.4 (13.2)75.2 (12.3)0.004GenderFemale122 (33.6)325 (46.8)447 (42.2)< 0.001Male241 (66.4)370 (53.2)611 (57.8)Time to death (days)Mean (SD)19.4 (18.9)19.8 (33.1)19.7 (29.0)0.85Hospital statusPrivate133 (36.6)103 (14.8)236 (22.3)< 0.001Public230 (63.4)592 (85.2)822 (77.7)Hospital locationMetropolitan293 (80.7)529 (76.1)822 (77.7)0.10Rural70 (19.3)166 (23.9)236 (22.3)ASA grade1–253 (15.2)31 (4.6)84 (8.2)< 0.0013225 (64.7)212 (31.4)437 (42.7)468 (19.5)334 (49.5)402 (39.3)52 (0.6)98 (14.5)100 (9.8)Preoperative death riskExpected/considerable100 (27.5)483 (79.5)583 (55.1)< 0.001Moderate155 (42.7)162 (23.3)317 (30.0)Small/minimal103 (28.4)43 (6.2)146 (13.8)(Missing) 5 (1.4) 7 (1.0) 12 (1.1)Patient transferred8 (2.2)175 (25.6)183 (17.6)< 0.001DiagnosisBenign84 (23.1)529 (76.1)613 (57.9)< 0.001Cancer279 (76.9)166 (23.9)445 (42.1)OperationAbdominoperineal resection42 (11.6)7 (1.0)49 (4.6)< 0.001Anterior resection125 (34.4)134 (19.3)259 (24.5)Hartmann's procedure37 (10.2)119 (17.1)156 (14.7)Left/sigmoid colectomy38 (10.5)71 (10.2)109 (10.3)Ileocolic/right/extended right hemicolectomy87 (24.0)254 (36.5)341 (32.2)Total colectomy13 (3.6)51 (7.3)64 (6.0)Other colectomy21 (5.8)59 (8.5)80 (7.6)Consultant surgeon roleOperating297 (81.8)526 (75.7)823 (77.8)< 0.001Assist31 (8.5)29 (4.2)60 (5.7)In theatre7 (1.9)20 (2.9)27 (2.6)No consultant28 (7.7)120 (17.3)148 (14.0)Consultant anaesthetist present357 (99.4)668 (98.5)1025 (98.8)0.31Operation duration (h)Mean (SD)3.2 (1.8)2.5 (1.3)2.7 (1.6)< 0.001Unplanned ICU admission194 (54.5)191 (27.9)385 (37.0)< 0.001Unplanned reoperation165 (46.1)175 (25.4)340 (32.4)< 0.001Postoperative complication277 (77.2)331 (47.8)608 (57.8)< 0.001Number of complicationsMean (SD)1.1 (1.0)0.7 (0.9)0.8 (0.9)< 0.001Abbreviations: ICU, intensive care unit; SD, standard deviation. Demographic and clinical factors of the included elective and emergency cohorts of mortality following colorectal surgery. Abbreviations: ICU, intensive care unit; SD, standard deviation. Causes of Death and Complications The most frequent causes of death were multiorgan failure (24.5%) and sepsis (22.8%), followed by pneumonia (16.2%), respiratory failure (11.4%) and bowel ischemia (7.5%). Causes of death were broadly similar between elective and emergency surgery (Figure 1). Causes of inpatient postoperative death following colorectal resection, stratified by elective/emergency surgery. Complications were reported for 608 patients (57.8%). Among elective patients the proportion with identified complications prior to death was 77.2%, versus 47.8% in emergency surgery. The median number of complications reported per patient was 1 (IQR 1–2). The most common complications prior to mortality were anastomotic leak (19.6%), pneumonia/aspiration (10.8%), generalised sepsis (8.7%), bowel ischemia (7.4%) and bleeding (7.0%). The five most common complications were similar between elective and emergency surgery, with anastomotic leak being the most common in both cohorts (Table 2). TABLE 2Rates of the 15 most common postoperative complications among patients who died following elective and emergency colorectal surgery.ComplicationElective (n = 363)Emergency (n = 695)Anastomotic leak99 (26.4%)111 (16.0%)Pneumonia/aspiration58 (16.0%)56 (8.1%)Bowel ischemia36 (9.9%)42 (6.0%)Bleeding36 (9.9%)38 (5.5%)Sepsis34 (9.4%)58 (8.3%)Ileus17 (4.7%)9 (1.3%)Myocardial infarction13 (3.6%)12 (1.7%)Bowel obstruction9 (2.5%)3 (0.4%)Bowel perforation8 (2.2%)18 (2.6%)Respiratory failure8 (2.2%)7 (1.0%)Stroke8 (2.2%)6 (0.9%)Pulmonary embolism6 (1.7%)6 (0.9%)Renal failure6 (1.7%)11 (1.6%)Atrial fibrillation4 (1.1%)4 (0.6%) Rates of the 15 most common postoperative complications among patients who died following elective and emergency colorectal surgery. The most reported contributing cofactors were cardiovascular disease (60.9%), age (55.9%), respiratory disease (39.0%), advanced malignancy (24.9%) and renal disease (23.1%). These were also similar between elective and emergency surgery (Table S1). CMIs and Potentially Preventable Mortality There were 1134 CMIs identified for 540 patients (51.0%) in this cohort (689 in 328 emergency patients and 445 in 212 elective patients). The median number of CMIs per patient among this group was 2 (IQR 1–3, range 1–9). Postoperative CMIs predominated in elective surgery, whereas preoperative CMIs were more common in emergency surgery (Figure 2). Preventability of preoperative, intraoperative and postoperative clinical management issues across elective and emergency surgery. Of all CMIs, 63.2% were classified as an area of ‘consideration’, 24.5% as a ‘concern’ and 12.3% as an overt ‘adverse event’. Most issues were either ‘definitely’ (18.6%) or ‘probably’ (47.9%) preventable, while a minority were ‘definitely not’ (3.4%) or ‘probably not’ (30.1%) preventable. CMIs were considered to have directly caused the death of a patient who would have survived for 12.2% of issues, while 67.0% ‘may have contributed’, and 20.7% ‘made no difference’. Overall, 321 deaths in this cohort (30.3%) were classified as potentially preventable, with a greater proportion of preventable deaths among elective compared with emergency surgery (35.3 vs. 27.8%, p = 0.01). Factors Associated With Potentially Preventable Mortality Cases with potentially preventable deaths were more likely to have lower ASA scores, lower preoperative risks of death and to be undergoing elective surgery (Table 3). Patients with postoperative complications, unplanned reoperations and unplanned ICU admissions were all more likely to have potentially preventable mortality. Patients with preventable deaths had a higher number of postoperative complications and a higher number of CMIs. Patients who died following an anastomotic leak were more likely to be classified as preventable mortality (53.1 vs. 24.8%, p < 0.001). TABLE 3Clinical factors associated with potentially preventable mortality following colorectal surgery.VariableNot preventable (n = 737, 69.7%)Preventable (n = 321, 30.3%) p Age (years)Mean (SD)75.2 (12.7)75.3 (11.3)0.86GenderFemale313 (70.0)134 (30.0)0.88Male424 (69.4)187 (30.6)Time to death (days)Mean (SD)18.5 (24.1)22.2 (37.9)0.06Hospital statusPrivate174 (73.7)62 (26.3)0.14Public563 (68.5)259 (31.5)Hospital locationMetropolitan583 (70.9)239 (29.1)0.11Rural154 (65.3)82 (34.7)Admission typeElective235 (64.7)128 (35.3)0.01Emergency502 (72.2)193 (27.8)ASA grade1–247 (56.0)37 (44.0)0.0013287 (65.7)150 (34.3)4296 (73.6)106 (26.4)578 (78.0)22 (22.0)Preoperative death riskExpected/considerable449 (77.0)134 (23.0)< 0.001Moderate197 (62.1)120 (37.9)Small/minimal83 (56.8)63 (43.2)(Missing) 8 (66.7) 4 (33.3)Patient transferred127 (69.4)56 (30.6)0.99DiagnosisBenign437 (71.3)176 (28.7)0.20Cancer300 (67.4)145 (32.6)OperationAbdominoperineal resection36 (73.5)13 (26.5)0.09Anterior resection173 (66.8)86 (33.2)Hartmann's procedure122 (78.2)34 (21.8)Left/sigmoid colectomy77 (70.6)32 (29.4)Ileocolic/right/extended right hemicolectomy227 (66.6)114 (33.4)Total colectomy41 (64.1)23 (35.9)Other colectomy61 (76.2)19 (23.8)Consultant surgeon roleOperating568 (69.0)255 (31.0)0.001Assist35 (58.3)25 (41.7)In theatre14 (51.9)13 (48.1)No consultant120 (81.1)28 (18.9)Consultant anesthetist present716 (69.9)309 (30.1)0.24Operation duration (hours)Mean (SD)2.7 (1.5)2.8 (1.6)0.15Unplanned ICU admission230 (59.7)155 (40.3)< 0.001Unplanned reoperationYes199 (58.5)141 (41.5)< 0.001Postoperative complicationYes376 (61.8)232 (38.2)< 0.001Number of complicationsMedian (IQR)1 (0–1)1 (0–1)< 0.001No640 (75.2)211 (24.8)< 0.001 –Anastomotic leak 97 (46.9)110 (53.1)< 0.001 –Pneumonia 64 (56.1)50 (43.9)0.001 –Sepsis 55 (59.8)37 (40.2)0.042 –Bowel Ischemia 49 (62.8)29 (37.2)0.216 –Bleeding 46 (62.2)28 (37.8)0.186Number of clinical management issuesMedian (IQR)0 (0–1)2 (1–3)< 0.001 –Preoperative 0.2 (0.5)0.8 (1.0)< 0.001 –Intraoperative 0.1 (0.3)0.6 (0.8)< 0.001 –Postoperative 0.1 (0.4)0.9 (1.1)< 0.001Abbreviations: ICU, intensive care unit, IQR, interquartile range; SD, standard deviation. Clinical factors associated with potentially preventable mortality following colorectal surgery. –Anastomotic leak Anastomotic leak –Pneumonia Pneumonia –Sepsis Sepsis –Bowel Ischemia Bowel Ischemia –Bleeding Bleeding –Preoperative Preoperative –Intraoperative Intraoperative –Postoperative Postoperative Abbreviations: ICU, intensive care unit, IQR, interquartile range; SD, standard deviation. Multivariable analysis identified that potentially preventable mortality was independently associated with lower preoperative risks of death (aOR 1.73, 95% CI 1.21–2.49, p = 0.003), occurrence of postoperative complications (aOR 1.90, 95% CI 1.36–2.65, p < 0.001), unplanned ICU admission (aOR 1.52, 95% CI 1.09–2.11, p = 0.01) and rural hospital care (aOR 1.48, 95% CI 1.04–2.11, p = 0.03) (Table S2). Subgroup analyses of elective and emergency surgery revealed no major differences compared to the overall analysis (Tables S3 and S4). Thematic Analysis of CMIs Preoperative Preoperative issues most frequently related to the appropriateness of surgery and the decision to operate (Table 4). In many cases, reviewers questioned whether surgery was indicated or likely to benefit the patient, particularly in those with advanced disease or significant comorbidities. Delays in reaching the decision to operate were also observed. TABLE 4Thematic analysis of preoperative, intraoperative and postoperative clinical management issues identified following colorectal surgery.Preoperative (n = 372)Intraoperative (n = 253)Postoperative (n = 365)Decision to operate and appropriateness of surgeryIntraoperative decision‐making and judgementDelayed recognition and response to deterioration –Decision to operate at all 21.5% –Choice of operation performed 30.0% –Delay in diagnosis of complications 9.0% –Delay in decision to operate 1.1% –Decision to form an anastomosis 19.4% –Delay to reoperation 8.8% –No formation of defunctioning stoma 7.9% –Delay in diagnosis of anastomotic leak 6.8%Delays in diagnosis and decision‐making –Decision making at reoperation 4.3% –Delayed escalation of care 4.4% –Delay to index operation 19.6% –Delay to diagnosis 18.3%Technical and procedural errorsSuboptimal management of complications –Delay in hospital transfer 4.0% –Iatrogenic injuries 15.0% –Lack of nasogastric tube insertion 9.0% –Delay to surgical referral 3.5% –Technical anastomotic issue 4.3% –Management of complications 8.5% –Inadequate hemostasis 2.0% –Management of anastomotic leak 2.7%Inadequate assessment and optimisation –Technical stoma issue 1.2% –Management of respiratory failure 1.4% –Inadequate assessment of risk/fitness 5.6% –Lack of optimisation 2.4%Supervision and experience gapsDeficiencies in postoperative monitoring and support –Inadequate investigation of pathology 2.4% –Surgeon inexperience 5.9% –Lack of ICU involvement 6.0% –Inadequate documentation 1.3% –Prolonged operative duration 1.6% –Early discharge from ICU/HDU 1.6% –Inadequate fluid resuscitation 1.3% –Missed intraoperative pathology 1.6% –Inadequate postoperative monitoring 1.4%Communication and coordination failuresCommunication and team factors –Clinician communication failures 2.2% –Clinician communication failures 4.9% –Poor inter‐hospital communication 1.3% –Futile postoperative management 2.5% –Lack of consultant input 1.3% –Poor documentation of care 1.4%Fluid, nutrition and physiological management –Postoperative fluid management 4.7% –Perioperative anticoagulation management 4.4% –Lack of venous thromboembolism prophylaxis 2.5% –Use of enteral feeding 2.2% –Management of medical comorbidities 1.6% –Electrolyte management 1.4% –Lack of nutritional support 1.1%Abbreviations: HDU, high dependency unit; ICU, intensive care unit. Thematic analysis of preoperative, intraoperative and postoperative clinical management issues identified following colorectal surgery. –Decision to operate at all Decision to operate at all –Choice of operation performed Choice of operation performed –Delay in diagnosis of complications Delay in diagnosis of complications –Delay in decision to operate Delay in decision to operate –Decision to form an anastomosis Decision to form an anastomosis –Delay to reoperation Delay to reoperation –No formation of defunctioning stoma No formation of defunctioning stoma –Delay in diagnosis of anastomotic leak Delay in diagnosis of anastomotic leak –Decision making at reoperation Decision making at reoperation –Delayed escalation of care Delayed escalation of care –Delay to index operation Delay to index operation –Delay to diagnosis Delay to diagnosis –Delay in hospital transfer Delay in hospital transfer –Iatrogenic injuries Iatrogenic injuries –Lack of nasogastric tube insertion Lack of nasogastric tube insertion –Delay to surgical referral Delay to surgical referral –Technical anastomotic issue Technical anastomotic issue –Management of complications Management of complications –Inadequate hemostasis Inadequate hemostasis –Management of anastomotic leak Management of anastomotic leak –Technical stoma issue Technical stoma issue –Management of respiratory failure Management of respiratory failure –Inadequate assessment of risk/fitness Inadequate assessment of risk/fitness –Lack of optimisation Lack of optimisation –Inadequate investigation of pathology Inadequate investigation of pathology –Surgeon inexperience Surgeon inexperience –Lack of ICU involvement Lack of ICU involvement –Inadequate documentation Inadequate documentation –Prolonged operative duration Prolonged operative duration –Early discharge from ICU/HDU Early discharge from ICU/HDU –Inadequate fluid resuscitation Inadequate fluid resuscitation –Missed intraoperative pathology Missed intraoperative pathology –Inadequate postoperative monitoring Inadequate postoperative monitoring –Clinician communication failures Clinician communication failures –Clinician communication failures Clinician communication failures –Poor inter‐hospital communication Poor inter‐hospital communication –Futile postoperative management Futile postoperative management –Lack of consultant input Lack of consultant input –Poor documentation of care Poor documentation of care –Postoperative fluid management Postoperative fluid management –Perioperative anticoagulation management Perioperative anticoagulation management –Lack of venous thromboembolism prophylaxis Lack of venous thromboembolism prophylaxis –Use of enteral feeding Use of enteral feeding –Management of medical comorbidities Management of medical comorbidities –Electrolyte management Electrolyte management –Lack of nutritional support Lack of nutritional support Abbreviations: HDU, high dependency unit; ICU, intensive care unit. A second major theme involved delays in diagnosis and progression to definitive management, including late recognition of surgical pathology, delayed transfer between hospitals and postponed referral to surgical teams. Problems with preoperative assessment and optimisation were also noted, such as incomplete evaluation of operative risk, limited optimisation of comorbidities or insufficient preoperative documentation. Less commonly, clinician communication and coordination failures were described, particularly around inter‐hospital handovers and consultant surgeon involvement in care. Intraoperative Intraoperative issues were dominated by concerns regarding surgical judgement and decision‐making, including the selection of procedure, whether to construct an anastomosis and the decision to omit a defunctioning stoma (Table 4). Technical and procedural errors formed the next most frequent group, encompassing iatrogenic injuries and anastomotic or haemostatic problems. In a smaller number of cases, concerns were raised regarding supervision and experience, including operations performed by junior surgeons with limited oversight, prolonged operative duration or missed intraoperative findings. Postoperative The postoperative phase was characterised by a high proportion of issues involving delayed recognition and response to deterioration (Table 4). These included missed or late diagnoses of postoperative complications, particularly anastomotic leaks and delays in returning to theatre once deterioration occurred. Problems with complication management were also frequent, such as incomplete or delayed interventions, and omission of nasogastric decompression in the setting of ileus. Related to this were deficiencies in postoperative monitoring and support, including limited access to or premature discharge from high‐dependency care and inadequate clinical observation on the ward. Other recurring themes included communication failures and suboptimal physiological management (notably in fluid, anticoagulation, nutrition and pre‐existing medical comorbidities). Discussion In this national peer‐review of surgical mortality, we examined over 1000 deaths following colorectal resection across Australia and found that nearly one‐third were potentially preventable. Although emergency surgery accounted for the majority of deaths, preventable mortality was proportionally higher among elective cases. Preventable deaths were driven predominantly by non‐technical factors, including suboptimal patient selection, anastomotic decision making and deficiencies in postoperative monitoring and escalation [18, 19]. Preventable deaths after colorectal surgery are less driven by intraoperative technical errors, and more from modifiable processes of decision‐making, optimisation and ‘rescue’ from complications, as well as health systems structural factors. This highlights the importance of strengthening preoperative assessment, multidisciplinary planning and postoperative surveillance to reduce mortality and improve outcomes. Interestingly, preventable mortality was proportionally higher among elective cases, despite these patients having lower ASA grades and lower estimated preoperative risk. This may reflect the perceived safety of ‘low‐risk’ elective settings, where postoperative deterioration is under‐recognised or escalation delayed. In contrast, clinicians managing frail or emergency patients may maintain a lower threshold for concern, engage in more intensive monitoring and escalate earlier, thereby reducing FTR despite greater physiological risk. This pattern may also partly reflect the audit structure itself, with elective deaths more likely to be perceived as ‘unexpected,’ while deaths in physiologically extreme or emergency contexts may be judged unavoidable. Although these associations may not directly reflect causal relationships, they underscore how preventable deaths may occur when perceived risk is low and responses to deterioration are delayed. Emergency colorectal resections accounted for approximately two‐thirds of deaths in this cohort, highlighting the morbidity associated with acute presentations of colorectal pathology. Notably, consultant surgeons were not present in theatre for 17.3% of emergency resections, contrasting with near‐universal consultant anesthetist presence and highlighting a potential gap in senior surgical oversight for high‐risk patients that warrants attention in the context of contemporary standards in emergency laparotomy care. Although deaths following emergency resection were less likely to be deemed preventable based on review of CMIs, it is likely that upstream factors such as availability of screening colonoscopy and overall access to healthcare influence these presentations, emphasising the importance of social determinants of health and population‐based screening programmes [20]. A major theme from this analysis was the decision to operate in patients that were unlikely to derive meaningful benefit, potentially reflecting futile surgical management [21]. Such decisions are shaped by optimism bias, time pressures and the cultural expectation to ‘do something’ [22]. Strengthening preoperative decision‐making requires structured senior and multidisciplinary review, explicit futility discussions and shared decision‐making that balances surgical feasibility against realistic postoperative outcomes [23]. Incorporating prediction tools may aid in decision making; refined models and clinical pathways using local data may also provide more precise insights [24]. Intraoperative decision‐making, particularly regarding anastomotic formation, was a recurrent theme among preventable deaths. Decisions to construct an anastomosis in high‐risk patients with marginal physiology, intraoperative contamination or poor perfusion were frequently questioned by reviewers, while in contrast, technical anastomotic issues were uncommonly raised as concerns. These cases illustrate that the threshold for performing or deferring an anastomosis remains inconsistent and may be influenced by individual surgeon demographics, personality and previous experiences [25]. While it is often easy to retrospectively criticise decision making, implementation of structured frameworks for intraoperative decision‐making may support multidisciplinary communication and safer, reproducible operative judgement [26]. Due to data limitations, we could not examine the association of surgical approach (e.g., open versus minimally invasive) with preventable mortality following colorectal surgery. The importance of postoperative vigilance and timely response to complications is well established, with consistent evidence identifying FTR as the major driver of variation in postoperative mortality between hospitals [1, 4, 5]. Among preventable deaths in this cohort, postoperative CMIs most often reflected delayed recognition of deterioration, inadequate escalation or insufficient senior involvement. Common scenarios included missed or late diagnosis of anastomotic leak, unrelieved ileus leading to aspiration [27], unrecognised sepsis and failure to reoperate or involve critical care in a timely manner. Anastomotic leak remains one of the most serious complications after both elective and emergency colorectal surgery [3], and diagnosis is frequently delayed despite biochemical monitoring of serum C‐reactive protein, procalcitonin or other emerging biomarker assays [28]. Delays to diagnosis and subsequent reoperation continue to contribute to preventable deaths [3, 29]. Improving outcomes therefore depends on early recognition of deterioration, rapid senior review and institutional investment in structured surveillance pathways, early warning systems and standardised escalation protocols [30]. These findings are likely translatable to other high‐income countries, with implications for surgeons, hospitals and health systems targeting preventable mortality following colorectal surgery. Patients treated at rural hospitals in this study had a higher risk of preventable mortality, highlighting the importance of strengthening perioperative systems across an entire healthcare system, particularly for emergency surgery. At a system level, structured preoperative triage and multidisciplinary review should be routine for high‐risk patients, ensuring that physiological optimisation and the potential benefits and risks of surgery are fully considered. Postoperatively, standardisation of care and escalation processes is essential. Although enhanced recovery programmes (ERPs) have improved recovery and reduced complications, their focus is largely on convalescence rather than mortality, and compliance remains highly variable [31]. The investigation and management of postoperative deterioration also require improvement [28, 32]. Tools such as the Dutch LeaK (DULK) score have been proposed to aid in the early identification of anastomotic leak, but have not been validated in emergency settings, and evidence regarding their clinical utility is mixed [33]. Hospitals must implement clear escalation pathways and adequate staffing to support timely recognition and response to deterioration. For surgical education, these findings highlight that decision‐making, communication and situational awareness are as vital to outcomes as technical skill [18]. National audit programmes such as ANZASM provide a unique mechanism for embedding these lessons into practice. Regular feedback of peer‐reviewed cases within departmental morbidity and mortality meetings can promote a culture of reflection, accountability and continuous improvement. This analysis is limited by the lack of denominator data for patients who survived postoperatively. ANZASM only records inpatient mortality, previously shown to capture approximately 50% of 90‐day deaths following colorectal surgery, and this may introduce selection bias [34]. Additional limitations include the potential subjectivity of assessor comments, and variability in how CMIs were identified and classified as preventable, especially with hindsight and confirmation biases from retrospective review of case notes. The long study period overlaps with the implementation of several major changes in perioperative care including early warning scores and ERPs, which may limit the applicability of our findings to contemporary practice. Finally, we were unable to account for missing data, and for relevant data points not recorded in ANZASM such as surgical approach, surgeon sub‐specialisation or volume or use of ERPs. We were also unable to further subcategorise benign indications for surgery due to limitations in the granularity of diagnostic data. We were also unable to examine inter‐state or territory variation, which may be relevant given differences in service configuration and access to specialist care across Australasia, and represents an important area for future investigation. In summary, nearly one‐third of deaths following colorectal resection in this national audit were considered potentially preventable. Most arose from deficiencies in patient selection, physiological optimisation and postoperative recognition or management of complications, rather than technical errors. These findings highlight that prevention of mortality after colorectal surgery depends as much on sound judgement, preparation and vigilance as on operative skill. Strengthening multidisciplinary preoperative assessment, ensuring timely escalation of postoperative deterioration and embedding structured review and feedback processes within surgical audit programmes represent practical opportunities to reduce preventable deaths. Sustained focus on system and decision‐making factors offers the greatest potential to improve survival after colorectal surgery. Author Contributions Cameron I. Wells: methodology, software, formal analysis, writing – original draft, funding acquisition. Scott J. P. McLaughlin: data curation, validation, writing – review and editing. Greg O'Grady: supervision, writing – review and editing. Matthew J. McGuiness: data curation, validation, writing – review and editing. Ian P. Bissett: supervision, writing – review and editing. Wasilwa Baraza: writing – review and editing, supervision. Maximillian O. Joret: data curation, validation, writing – review and editing. Funding Cameron I. Wells is supported by a Health Research Council of New Zealand Clinical Research Training Fellowship (22/45). Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supporting Information: Appendix 1: STROBE checklist. Table S1: Most common identified cofactors among patients who died following elective and emergency colorectal surgery. Table S2: Results of multivariable analysis predicting potentially preventable mortality following colorectal surgery. Table S3: Results of multivariable analysis predicting potentially preventable mortality following elective colorectal surgery. Acknowledgements We would like to thank Dr. Nathan Procter and Amy Carter for their assistance with extracting ANZASM data for this analysis. Open access publishing facilitated by The University of Auckland, as part of the Wiley ‐ The University of Auckland agreement via the Council of Australasian University Librarians. Data Availability Statement The data analysed in this paper are available upon request through the ANZASM process. The code used for the analyses in this manuscript is available from the authors upon reasonable request. References Associated Data Supplementary Materials Supporting Information: Appendix 1: STROBE checklist. Table S1: Most common identified cofactors among patients who died following elective and emergency colorectal surgery. Table S2: Results of multivariable analysis predicting potentially preventable mortality following colorectal surgery. Table S3: Results of multivariable analysis predicting potentially preventable mortality following elective colorectal surgery. Data Availability Statement The data analysed in this paper are available upon request through the ANZASM process. The code used for the analyses in this manuscript is available from the authors upon reasonable request.