SURGERY IN LOW AND MIDDLE INCOME COUNTRIES Analysis of Surgical Mortalities Using the Fishbone Model for Quality Improvement in Surgical Disciplines M. S. Moeng1 • T. E. Luvhengo2 Accepted: 20 November 2021 / Published online: 4 February 2022 � The Author(s) under exclusive licence to Société Internationale de Chirurgie 2022 Abstract Background The healthcare industry is complex and prone to the occurrence of preventable patient safety incidents. Most serious patient safety events in surgery are preventable. Aim This study was conducted to determine the rate of occurrence of preventable mortalities and to use the fishbone model to establish the main contributing factors. Methods We reviewed the records of patients who died following admission to the surgical wards. Data regarding their demography, diagnosis, acuity, comorbidities, categorization of death and contributing factors were extracted from the Research Electronic Data Capture (REDCap) database. Factors which contributed to preventable and potentially preventable mortalities were collated. The fishbone model was used for root cause analysis. The study received prior ethical clearance (M190122). Results Records of 859 mortalities were found, of which 65.7% (564/859) were males. The median age of the patients who died was 49 years (IQR: 33–64 years). The median length of hospital stay before death was three days (IQR: 1–11 days). Twenty-four percent (24.1%) of the deaths were from gastrointestinal (GIT) emergencies, 18.4% followed head injury and 17.0% from GIT cancers. Overall, 5.4% of the mortalities were preventable, and 41.1% were considered potentially preventable. The error of judgment and training issues accounted for 46% of mortalities. Conclusion Most surgical mortalities involve males, and around 46% are either potentially preventable or pre- ventable. The majority of the mortality were associated with GIT emergencies, head injury and advanced malig- nancies of the GIT. The leading contributing factors to preventable and potentially preventable mortalities were the error of judgment, inadequate training and shortage of resources. Introduction The complexity of the healthcare industry predisposes it to the occurrence of patient safety incidents (PSI) [1]. Structural or process failures at multiple levels are involved in almost all serious patients’ safety incidents [2–5]. More than a third of serious patients’ safety incidents in surgery are either preventable or potentially preventable [6, 7]. The occurrence of preventable PSI reflects a poorly functioning healthcare system [6, 8]. Some organizations use the rate of occurrence of preventable mortalities for benchmarking and ranking of hospitals [9]. & M. S. Moeng maeyane.moeng@wits.ac.za 1 Charlotte Maxeke Johannesburg Academic Hospital (CMJAH), University of the Witwatersrand, Box 7053, Cresta, Johannesburg, Republic of South Africa 2 Clinical Head Department of Surgery, CMJAH, University of the Witwatersrand, Johannesburg, Republic of South Africa 123 World J Surg (2022) 46:1006–1014 https://doi.org/10.1007/s00268-021-06414-8 http://crossmark.crossref.org/dialog/?doi=10.1007/s00268-021-06414-8&domain=pdf https://doi.org/10.1007/s00268-021-06414-8 Although the recent advances in healthcare in high-in- come countries (HICs) have led to improved surgical out- comes, the same cannot be said for healthcare facilities in low- and middle-income countries (LMICs) [10]. Among the some of the barriers to improvement in surgical care is inability to collect valuable data, lack of standardization of care and/or care pathways, poor leadership and a prevailing culture of blame [1, 11–14]. Preferably, data collection, evaluation and thorough root cause analysis (RCA) should be done regularly to identify gaps to enable targeted intervention [15, 16]. Among the commonly used RCA strategies is the brainstorming method, the lean method, the use Shewhart control charts or Pareto charts, process mapping, interre- lationship diagrams [17], the five whys method, the current reality tree method, fault tree analysis [18], human factors classification framework [19] and the Ishikawa cause and effect diagram (fishbone model) [15, 17, 20]. The fishbone model is preferred because it is simple to learn and facil- itates identification and grouping of factors that contributed to unsatisfactory performance [15, 16, 21]. This study was conducted to determine the rate of occurrence of pre- ventable and potentially preventable mortalities in surgical patients. Furthermore, the fishbone model was used to establish the main contributing factors. Methods Records of consecutive surgical patients who were 18 years and older who died during admission to the sur- gical wards in the period starting from January 1, 2017, to December 31, 2018, were studied. Patients who died after they were sent out to other departments and those whose records were incomplete were excluded. Permission to conduct the study was received from the Research Review Board of Charlotte Maxeke Johannesburg Academic Hospital (CMJAH) and Human Research Ethics Commit- tee (Medical) of the University of the Witwatersrand (M190122). The Department of Surgery at CMJAH started running its morbidity and mortality meetings (M&M) on the REDCap platform in 2016. Patients’ demographics, admission diagnosis, comorbidities, surgical procedures, intra-operative adverse events, the grade of postoperative complications according to the Clavien–Dindo classifica- tion and overall outcome are captured during collection of data for M&M by individual divisions and units of the department. Preventability of the serious adverse events and their contributing factors are also recorded. The units marked the factors which they considered to have con- tributed to each preventable or potentially pre- ventable mortality from a drop-down list. The list of contributing factors include training, supervision, lack of resources, equipment issues, infrastructure problems and communication failure. There was also an option to add any other factor which was thought to have contributed to a mortality. The decisions that were made at unit-based M&M meetings are ratified at the meeting of the entire department. An additional verification was done by a sur- geon with more than 20 years’ experience. The senior surgeon was blinded to the allocated category of deaths and their contributing factors that were assigned during the weekly departmental morbidity and mortality meeting. A mortality was considered preventable if it was not expected to have occurred in any setting regardless the available resources, including the level of expertise. The potentially preventable mortalities meant the deaths that could not have occurred if the setting was ideal, and all the resources were available. The mortalities which were considered not have been preventable including deaths due to advanced cancer, severe traumatic brain injury, arrival in the extremes due to established sepsis, severe burn injury, advanced HIV disease or irreversible organ dysfunction. Categorical data were reported in numbers and per- centages. Mean with standard deviation was used for continuous data where it was appropriate. A Chi-square test was used to find the association between gender, pathology and how preventable mortality was. For the relationship between continuous variables, the nonparametric Kruskal– Wallis test was used, as the data were not normally dis- tributed. A p-value below 0.05 was regarded as being statistically significant. Finally, the factors which con- tributed to preventable and potentially preventable mortal- ities were identified and collated, and a fishbone diagram. We used the themes which are used in our M&M for grouping of the factors instead of the usual man, machine, materials, method, measurement and mother nature (Environment). Results A total of 862 mortalities were captured, which is 8.2% of the 10,529 patients who were admitted to the surgical wards during the study period (Table 1). Eight hundred and fifty-nine records were suitable for further analysis. Around 66.0% (564/859) of patients who died were males. The median age of male patients who died was 44.9 years (IQR: 33–64 years). The median length of stay of patients who died was three days (IQR: 1–11 days). Most mortalities involved patients who had GIT emergencies at 24.1% (207/859), head injury 18.4% (158/859) and GIT malignancies 17.0% (146/859). Overall, 5.4% of the mortalities were categorized during weekly M&M as preventable (Table 2). World J Surg (2022) 46:1006–1014 1007 123 Forty-six mortalities (5.4%: 46/859) were categorized as preventable during the weekly morbidity and mortality meetings, of which 65.2% (30/46) were males. The median age of patients whose deaths were considered not pre- ventable was 46 years, for potentially preventable mortali- ties 51 years, and 45 years for preventable mortalities. The difference in the median ages of patients who had the three categories of mortalities was statistically significant (p- value = 0.0028). Similarly, the difference in median length of stay before death of patients in the three groups was also found to be statistically significant (p-value\0.0001). Most of the mortalities which were considered not pre- ventable were from severe head injury or advanced GIT malignancies (Fig. 1). In 5.9% (20/337) of the decision of the record regarding preventability of the mortality was not made during the M&M meeting. The senior surgeon clas- sified 75.6% (595/787) of the mortalities as not preventable (Table 3). Table 1 Breakdown of number of admissions, operations, morbidity and mortalities per specialty Parameter General surgery Trauma Vascular Transplant Total Admissions 6124 (58.16%) 2733 (25.96%) 1111 (10.56%) 561 (5.32%) 10,529 (100%) Elective operations 1632 (57.5%) 248 (8.7%) 713 (25.1%) 244 (8.6%) 2837 (100%) Emergency operations 1402 (42.8%) 1155 (35.3%) 450 (13.7%) 267 (8.2%) 3274 (100%) Morbidities 467 (51.1%) 257 (28.1%) 34 (3.7%) 156 (17.1%) 914 (100%) Mortalities 404 (46.9%) 352 (40.8%) 88 (10.2%) 18 (2.1%) 862 (100%) Table 2 Demography, diagnoses and classification of mortalities in patients who died (N = 859) Variable Class Number (%) Sex Male 564 (66.0%) Female 290 (34.0%) Age Median (IQR) 49 (33–64) years Length of stay before death Median (IQR) 3 (1–11) days Pathology classification Blunt trauma 74 (8.6%) Penetrating trauma 100 (11.7%) Head injury 158 (18.4%) Vascular elective 15 (1.8%) Vascular emergency 81 (9.4%) GIT malignancy 146 (17.0%) GIT non-malignancy 207 (24.1%) Endocrine/Breast malignancy 15 (1.8%) Endocrine/breast non-malignancy 4 (0.5%) Soft tissue malignancy 6 (0.7%) Soft tissue non-malignancy 21 (2.5%) Burns 20 (2.3%) Transplant 11 (1.3%) Mortality Not preventable 459 (53.5%) Potentially preventable 353 (41.1%) Preventable 46 (5.4%) 1008 World J Surg (2022) 46:1006–1014 123 The factors which contributed to most preventable and potentially preventable mortalities included issues related to the error of judgment (25.8%: 65/252), inadequate training (20.2%: 51/252) and shortage of resources (15.5%: 39/252) (Table 4). Examples of what were classified under the error of judgment include sending unstable patients for radiological Fig. 1 Pathology classification of mortality Table 3 Comparison of categorization of mortalities by M&M and a senior surgeon Classification of mortalities Departmental morbidity and mortality meeting (N = 859) Senior general surgeon (n = 787) Preventable 46 (5.1%) 28 (3.6%) Potentially preventable 352 (41.0%) 158 (20.1%) Not preventable 456 (53.1%) 595 (75.6%) Not known 5 (0.6%) 6 (0.8%) Table 4 Categorization and weighting of factors which contributed to preventable and potentially preventable mortalities Category Number Actual percentage (%) Cumulative percentage (%) Error of judgment 65 25.8 25.8 Training 51 20.2 46.0 Resources 39 15.5 61.5 Systems failure 35 13.9 75.4 Diagnostic dilemma 25 9.9 85.3 Patient factors 16 6.3 91.6 Communication gap 8 3.2 94.8 Lack of supervision 6 2.4 97.20 Nosocomial infection 6 2.4 99.6 Industrial action 1 0.4 100 Total 252 100 100 World J Surg (2022) 46:1006–1014 1009 123 investigations, not performing abbreviated surgery, under- estimating the significance of intra-operative bleeding, doing a primary anastomosis instead of bringing out a stoma, not calling for help, delay in performing a relook, admitting a patient to a low-level care area, not tracing the results of ordered investigations and premature discharge from ICU or the hospital. Additionally, the error of judg- ment also included underestimation the severity of a dis- ease or a postoperative complication leading to delayed intervention and overestimation of the capability of junior staff members. Included among the issues related to problems of training were diagnostic difficulties, delayed referral, delayed intervention and poor surgical techniques. On the other hand, the shortage of resources included ICU beds, timeous availability of operating theater and unavailability of appropriately qualified staff (Fig. 2). Discussion Patient safety incidents including SAEs are intimately linked to the delivery of healthcare services. It is not possible to eliminate all SAEs in the healthcare industry, irrespective of the income status of a community. The healthcare industry’s complexity makes it vulnerable to the occurrence of SAEs at multiple points [19]. Some of the PSIs end as just near misses, while others lead to a sig- nificant prolongation of the length of hospital stay, dis- ability, or death. Overview of mortality This study has shown that more than 45% of surgical mortalities are either potentially preventable or pre- ventable; this is close to the 48.3% which was reported by Chen et al. in 2017 [7]. Following their analysis in the Victorian Audit of Surgical Mortality, different evaluators Fig. 2 Fishbone diagram depicted factors which contributed to preventable mortalities in 2018 (n = 141) 1010 World J Surg (2022) 46:1006–1014 123 in the same study reported a preventable mortality rate of 14.6% [7]. However, Moon et al. in 2015 reported a pre- ventable mortality rate of 25.2%, a figure which most likely have included deaths that would have been considered potentially preventable in other settings [21]. On average, 4.3% of mortalities in the current study were preventable, which is close to the 3.4% reported in a publication by Bindroo et al. in 2015 [22]. Conducting this study was an opportunity to evaluate the regular M&M meetings and to identify factors to be tar- geted for quality improvement. The overall mortality rate of 8.2% in the study period was within the background of management of patients in a LMIC environment [8, 10]. The preventable mortality was low at 5.1% during the departmental deliberations but even lower when indepen- dently assessed by a senior surgeon. Around 65.7% of the patients who died were males and their average age was 44.9 years, compared to 57.5 years for females. The top 3 causes of death were non-malignant conditions of the GIT, head injury and advanced GIT malignancies. The higher mortality in males is not sur- prising as complicated non-malignant conditions of the GIT such as ruptured appendix [23] and perforated peptic ulcer [24], trauma [4, 21, 25, 26] are seen in males. It emphasizes the reality that some common GIT malignan- cies are more prevalent in adult males [27, 28]. Factors associated with mortality The outcome following surgery is influenced by the age of a patient, the severity of the underlying conditions, asso- ciated comorbidities, American Society of Anaesthesiolo- gists class and availability of resources [9, 29, 30]. Fifty- four percent of the patients who died in the current study had comorbidities, with hypertension being the most associated illness at 28.4%. Similarly, Lees et al. also found that hypertension was the most common comorbidity in the elderly patients who had poor outcomes following emergency surgery [30]. The association of HIV was recorded as a contributory factor in only 7.8%. This con- tribution was far below what was expected as the preva- lence of HIV in the adult population in South Africa was around 15% [31]. The profile of surgical conditions is influenced by the geographical location and economic status of a community, which would also influence the causes of death. Whereas non-communicable diseases are more prevalent in high-income countries, infectious disease and trauma are the leading causes of death in LMICs [8, 22]. The fact that non-communicable diseases were the leading causes of deaths in the current study may be reflective of a country in economic transition. Most of the victims of trauma and violence were males, which is also reflected in a high mortality rate in men [4, 21, 25]. It is estimated that globally up to 48.3% of trauma mortalities are either preventable or potentially preventable [32]. The factors which influence the occur- rence of mortality in trauma include the mechanism of injury, site of injury, severity of injury, time of injury, time it took to initiation of therapy and the level of care of a receiving healthcare facility [33]. Management of trauma patients requires coordination across the care pathways which has many layers starting with prehospital care, care during transfer to or between hospitals, emergency departments, intra-operative care, and postoperative man- agement [34]. The multiplicity of the layers involved in trauma care makes it to be vulnerable to deviations and at risk of occurrence of serious PSI. The common non-trauma surgical emergencies are acute appendicitis, complicated peptic ulcer disease, acute pan- creatitis, bowel obstruction, skin and soft tissue infection, complicated hernia and diabetic foot sepsis. What is common across the non-trauma surgical emergencies is the higher mortality in males, older adults and delayed pre- sentation [24]. The average age of patients who died from other causes in the current study was at least 12 years older than those following trauma which is in keeping with the demography of the victims of trauma [32]. Cancer was the third most common cause of surgical mortalities in our study, 5.3% of which was potentially preventable. The common malignancies in surgical patients include cancer of the breast, skin, colon, liver, esophagus, stomach, pancreas and soft tissue. The above cancers are more prevalent after the age of 50 years [28] and are considered irresectable when they are either locally advanced or metastatic. However, local complications such as bleeding, perforation and obstruction may necessitate emergency surgery even for tumors which are irresectable. These complications may further accelerate death if these complications are secondary to iatrogenic causes. Quality control factors associated with outcomes Dogget argues that there is a cause for each problem [17]. It is, therefore, reasonable to expect that there would be an underlying cause for every preventable or potentially pre- ventable mortality. Among the factors commonly associ- ated with preventable mortalities are medication errors, delay in diagnosis, delay in transfer and violation of management guidelines [35]. The fishbone method is among the tools used for root cause analysis and is pre- ferred because of the combination of its simplicity and that it can facilitate the grouping of causal factors [17]. From our study the error of judgment, training and supervision had more impact on preventable or potentially pre- ventable mortalities than shortage of resources and systems issues. Medication-related errors did not contribute World J Surg (2022) 46:1006–1014 1011 123 significantly to preventable and potentially pre- ventable mortalities in surgical patients in the current study. Some of the weaknesses of the fishbone method are that it is subjective, leads to fragmentation of contributing factors that may be interrelated and may hide systems issues [15, 17]. The complexity of the healthcare industry is erroneously compared with that of the nuclear, petrochemical, aviation and airline industry, which is unfortunate. The healthcare industry is much more complex and not easy to coordinate as every step during the delivery of healthcare relies on a combination of trust and independence among healthcare workers themselves and their patients. Furthermore, it is unlikely that a mistake made by a lone healthcare worker would cause a preventable PSI [2, 11] as the factors that contribute to poor healthcare outcomes rarely occur in isolation [36]. Things such as timeous availability of resources, appropriate training, professionalism, service delivery processes and systems issues significantly influ- ence surgical outcomes [4]. While the practice in all other industries emphasizes avoidance of risks, the same cannot be said for the healthcare industry. Flights are not allowed to take off even when the risk is minimal, whereas patients are given a chance in the healthcare industry even in situ- ations when the prognosis is guarded [36]. An essential step for quality improvement programs in any industry is knowing the baseline and to continuously collect data. Unfortunately, there are very few studies emanating from the African continent reporting on quality and patients’ safety programs [37]. Some of the advanced reasons for limited reports of PSI from Africa include inability to collect useful data, fear of shame or economic implications, and the potential to attract medicolegal liti- gations [14, 38, 39]. However, the benefits of honest reporting of clinical outcomes far outweigh the risks [11, 35, 40–43]. Limitations The study was a retrospective review of records, and some records were incomplete, which could have influenced the results. Additionally, medication-related errors were most probably underreported. The reliability of categorization and weighting of factors that contributed to the outcome of interest using the fishbone method is highly dependent on the number of raters and their expertise. Other disadvan- tages of the fishbone method include its high subjectivity and inability to assist in identifying a specific root cause. The fishbone method is not good for identification of the possible interconnectedness of the causal factors. Conclusion The majority of surgical mortalities involve males, and around 46% are either preventable or potentially pre- ventable. The majority of mortalities are related to GIT emergencies, head injury and advanced malignancies of the GIT. The two most common contributing factors to pre- ventable and potentially preventable surgical mortalities are the error of judgment, training and limited availability of resources. Identification of factors which contribute to preventable or potentially preventable mortalities present an opportunity for improvement the quality of surgical care. Recommendations A follow-up prospective study involving a larger data set should be conducted to include a detailed analysis of additional factors contributing to preventable and poten- tially preventable mortalities is necessary. Acknowledgements We would like to acknowledge Dr. Khan F, Derwin M, Sischy E, Elfick M, Gaynor M, Jagjivan K, Aboo F, Kassel T, Ferguson T, Nanabhay A, Jivan K, Mabaso C, Khusi S, Solundwana B, Samuels G and Williams R for their participation in the development of the research proposal and application of regula- tory proposal as part of the mandatory requirement for the 4th Year Medical Students at University of the Witwatersrand. Funding No grant funding for the study. Declarations Conflict of interest The authors have no conflict to disclose. References 1. Leape LL (2015) Patient safety in the era of healthcare reform. Clin Orthop Relat Res 473:1568–1573. https://doi.org/10.1007/ s11999-014-3598-6 2. Leape LL (2014) The checklist Conundrum. N Engl J Med 370:1063–1064. https://doi.org/10.1056/NEJMe1315851 3. Reilly J, Myers J, Salvardor D, Trowbridge R (2014) Use of a novel, modified fishbone diagram to analyze diagnostic errors. Diagnosis 1(2):167–171. https://doi.org/10.1515/dx-2013-0040 4. Steenkamp C, Kong VY, Clarke DL et al (2017) The effect of systematic factors on the outcome of trauma laparotomy at a major trauma centre in South Africa. Ann R Coll Surg Engl 99:540–544. https://doi.org/10.1308/rcsann.2017.0079 5. Seshia S, Young B, Makhinson M et al (2018) Gating the holes in the Swiss cheese (part1): expanding professor reason’s model of patient safety. J Eval Clin Pract 24:187–197. https://doi.org/10. 1111/jep.12847 6. Anderson O, Davis CA, Hanna GB, Vincent CA (2013) Surgical adverse events: a systematic review. Am J Surg 206(2):253–262. https://doi.org/10.1016/j.amjsurg.2012.11.009 1012 World J Surg (2022) 46:1006–1014 123 https://doi.org/10.1007/s11999-014-3598-6 https://doi.org/10.1007/s11999-014-3598-6 https://doi.org/10.1056/NEJMe1315851 https://doi.org/10.1515/dx-2013-0040 https://doi.org/10.1308/rcsann.2017.0079 https://doi.org/10.1111/jep.12847 https://doi.org/10.1111/jep.12847 https://doi.org/10.1016/j.amjsurg.2012.11.009 7. Chen A, Retegan C, Vinluan J, Beiles CB (2017) Potentially preventable deaths in the victorian audit of surgical mortality. ANZ J Surg. https://doi.org/10.1111/ans.13804 8. Abejew AA, Tamir AS, Kerie MW (2014) Retrospective analysis of mortalities in a tertiary care hospital in Northeast Ethiopia. BMC Res Notes 7:46. https://doi.org/10.1186/1756-0500-7-46 9. Shidara Y, Fujita Y, Fukunaga S, Ikeda K, Uemura M (2016) In- hospital mortality after surgery: a retrospective cohort study in a Japanese university hospital. Springer Plus 5:680. https://doi.org/ 10.1186/s40064-016-2279-1 10. Roy A, van der Weijden T, de Vries N (2017) Predictors and consequences of rural clients’ satisfaction level in the district public-private mixed system health system of Bangladesh. Global Health Res Policy 2:31. https://doi.org/10.1186/s41256-017- 0052-9 11. Leape LL (2000) Reporting of medical errors: time for a reality check. Qual Health Care 9:144–145. https://doi.org/10.1136/qhc. 9.3.144 12. Leape LL (2006) When good doctors go bad: a systems problem. Ann Surg 244(5):649–652. https://doi.org/10.1097/01.sla. 0000243602.60353.b1 13. Chung KC, Kotsis SV (2012) Complications in surgery: root cause analysis and preventive measures. Plast Reconst Surg 129(6):1421–1427. https://doi.org/10.1097/PRS.0b013e3182ec da0 14. Oyebode F (2013) Clinical errors and medical negligence. Med Princ Pract 22:323–333. https://doi.org/10.1159/000346296 15. Kwok YTA, Mah APY, Pang KMC (2020) Our first review: an evaluation of effectiveness of root cause analysis recommenda- tions in Hong Kong public hospitals. BMC Health Serv Res 20:507. https://doi.org/10.1186/s12913-020-05356-6 16. Drehobl P, Stover BH, Koo D (2014) On the road to a stronger public health workforce: visual tools to address complex chal- lenges. Am J Prev Med 47(5):S280–S285. https://doi.org/10. 1016/j.amepre.2014.07.013 17. Dogget AM (2005) Root cause analysis: a framework for tool selection. Qual Manag J 12(4):34–45. https://doi.org/10.1080/ 10686967.2005.11919269 18. Bowie P, Skinner J, de Wet C (2013) Training health care pro- fessionals in root cause analysis: a cross-sectional study of posttraining experiences, benefits and attitudes. BMC Health Serv Res 13:50 19. Mitchell RJ, Williamson A, Molesworth B (2016) Application of a human factors classification framework for patient safety to identify precursor and contributing factors to adverse clinical incidents in hospital. Appl Ergon 52:185–198. https://doi.org/10. 1016/j.apergo.2015.07.018 20. Abecassis ZA, McElroy LM, Patel RM et al (2015) Applying fault tree analysis to the prevention of wrong site surgery. J Surg Res 193(1):88–94. https://doi.org/10.1016/j.jss.2014.08.062 21. Moon S, Lee SH, Ryoo HW et al (2015) Preventable trauma death rate in Daegu, South Korea. Clin Exp Emerg Med 2:236–243. https://doi.org/10.15441/ceem.15.085 22. Bindroo S, Saraf R (2015) Surgical mortality audit-lessons learned in a developing nation. Int Surg 100(6):1026–1032. https://doi.org/10.9738/intsurg-d-14-00212.1 23. Nsuti R, Kruger D, Luvhengo TE (2014) Clinical presenation of acute appendicitis in adults at the Chris Hani Baragwanath aca- demic hospital. Int J Emerg Med 7:12. https://doi.org/10.1186/ 1865-1380-7-12 24. Moller MH, Engebjerg MC, Adamsen S et al (2012) The peptic ulcer perforation (PULP) score: a predictor of mortality following peptic ulcer perforation A cohort study. Acta Anaesthesiol Scand 56(5):655–662 25. DiMaggio C, Ayoung-Chee P, Shinseki M et al (2016) Traumatic injury in the United States: in-patient epidemiology 2000–2011. Injury 47(7):1393–1403. https://doi.org/10.1016/j.injury.2016.04. 002 26. Bradshaw D, Nannan N, Pillay-van WV et al (2019) Burden of disease in South Africa: protracted transitions driven by social pathologies. SAMJ 2019 109(11b):69–76. https://doi.org/10. 7196/SAMJ.2019.v109i11b.14273 27. White A, Ironmonger L, Steele R et al (2018) A review of sex- related differences in colorectal cancer incidence, screening uptake, routes to diagnosis, cancer stage and the survival in the UK. BMC Cancer 18:906. https://doi.org/10.1186/s12885-018- 4786-7 28. Kim SE, Paik HY, Yoon H et al (2015) Sex-and gender specific disparities in colorectal cancer risk. World J Gastroentero 21(7):5167–5175. https://doi.org/10.3748/wjg.v21.i17.5167 29. Symons NRA, Moorthy K, Almoudaris AM et al (2013) Mortality in high-risk emergency general surgical admissions. Br J Surg. https://doi.org/10.1002/bjs.9208 30. Lees MC, Merani S, Tauh K et al (2015) Perioperative factors predicting poor outcome in elderly patients following emergency general surgery: a multivariate regression analysis. J Can Chir 58(5):312–317. https://doi.org/10.1503/cjs.011614 31. Parker E, Judge M, Macete E et al (2021) HIV infection in Eastern and Southern Africa: highest burden, largest challenges, great potential. Southern Afr J HIV Med 22:1. https://doi.org/10. 4102/sajhivmed.v22i1.1237 32. Chiara O, Cimbanassi S, Pitidis A et al (2006) Preventable Trauma deaths; from panel review to population- based studies. World J Emerg Surg 1:2. https://doi.org/10.1186/ 1749-7922-1-12 33. Boschini LP, Lu-Myers Y, Msiska N et al (2016) Effect of direct and indirect transfer status on trauma mortality in sub-Saharan Africa. Injury 47(5):1118–1122. https://doi.org/10.1016/j.injury. 2016.01.015 34. Kim SC, Song KJ, Shin SD et al (2015) Preventable deaths in patients with traumatic brain injury. Clin Exp Emerg Med 2(1):51–58. https://doi.org/10.15441/ceem.14.023 35. Donaldson LJ, Panesar SS, Darzi A (2014) Patient-safety related hospital deaths in England: thematic96 analysis of incidents reported to a national database, 2010–2012. PLOS Med 11(6):e1001667. https://doi.org/10.1371/journal.pmed.1001667 36. Su CF, Chu CM, Yuan YJ et al (2017) Use of a modified fishbone diagram to analyze diagnostic errors in emergency physicians. J Acute Med 7(4):149–157. https://doi.org/10.6705/j.jacme.2017. 0704.003 37. Atanasov AG, Yeung AWK, Klager E et al (2020) First, do no harm (gone wrong): total-scale analysis of medical errors scien- tific literature. Front Public Health. https://doi.org/10.3389/fpubh. 2020.558913 38. Aziz O, Fink D, Hobbs L et al (2011) Hospital mortality under surgical care. Ann R Coll Surg Engl 93:193–200. https://doi.org/ 10.1308/003588411x563411 39. Manaseki-Holland S, Lilford RJ, Te AP et al (2019) Ranking hospitals based on preventable hospital death rates: a systematic review with implications for both direct measurement and indi- rect measurement through standardized mortality rates. Milbank Q 97(1):228–284 40. Stelfox HT, Palmisani S, Scurlock C et al (2006) The ‘‘To Err is Human’’ report and the patient safety literature. Qual Saf Health Care 2006(15):174–178. https://doi.org/10.1136/qshc.2006. 017947 41. Chou E, Abboundi H, Khan MS et al (2015) Should surgical outcomes be published? J R Soc Med 108(4):127–135. https:// doi.org/10.1177/0141076815578652 42. Gandhi TK, Kaplan GS, Leape L et al (2018) Transforming concepts in patient safety: a progress report. BMJ Qual Saf 27:1019–1026. https://doi.org/10.1136/bmjqs-2017-007756 World J Surg (2022) 46:1006–1014 1013 123 https://doi.org/10.1111/ans.13804 https://doi.org/10.1186/1756-0500-7-46 https://doi.org/10.1186/s40064-016-2279-1 https://doi.org/10.1186/s40064-016-2279-1 https://doi.org/10.1186/s41256-017-0052-9 https://doi.org/10.1186/s41256-017-0052-9 https://doi.org/10.1136/qhc.9.3.144 https://doi.org/10.1136/qhc.9.3.144 https://doi.org/10.1097/01.sla.0000243602.60353.b1 https://doi.org/10.1097/01.sla.0000243602.60353.b1 https://doi.org/10.1097/PRS.0b013e3182ecda0 https://doi.org/10.1097/PRS.0b013e3182ecda0 https://doi.org/10.1159/000346296 https://doi.org/10.1186/s12913-020-05356-6 https://doi.org/10.1016/j.amepre.2014.07.013 https://doi.org/10.1016/j.amepre.2014.07.013 https://doi.org/10.1080/10686967.2005.11919269 https://doi.org/10.1080/10686967.2005.11919269 https://doi.org/10.1016/j.apergo.2015.07.018 https://doi.org/10.1016/j.apergo.2015.07.018 https://doi.org/10.1016/j.jss.2014.08.062 https://doi.org/10.15441/ceem.15.085 https://doi.org/10.9738/intsurg-d-14-00212.1 https://doi.org/10.1186/1865-1380-7-12 https://doi.org/10.1186/1865-1380-7-12 https://doi.org/10.1016/j.injury.2016.04.002 https://doi.org/10.1016/j.injury.2016.04.002 https://doi.org/10.7196/SAMJ.2019.v109i11b.14273 https://doi.org/10.7196/SAMJ.2019.v109i11b.14273 https://doi.org/10.1186/s12885-018-4786-7 https://doi.org/10.1186/s12885-018-4786-7 https://doi.org/10.3748/wjg.v21.i17.5167 https://doi.org/10.1002/bjs.9208 https://doi.org/10.1503/cjs.011614 https://doi.org/10.4102/sajhivmed.v22i1.1237 https://doi.org/10.4102/sajhivmed.v22i1.1237 https://doi.org/10.1186/1749-7922-1-12 https://doi.org/10.1186/1749-7922-1-12 https://doi.org/10.1016/j.injury.2016.01.015 https://doi.org/10.1016/j.injury.2016.01.015 https://doi.org/10.15441/ceem.14.023 https://doi.org/10.1371/journal.pmed.1001667 https://doi.org/10.6705/j.jacme.2017.0704.003 https://doi.org/10.6705/j.jacme.2017.0704.003 https://doi.org/10.3389/fpubh.2020.558913 https://doi.org/10.3389/fpubh.2020.558913 https://doi.org/10.1308/003588411x563411 https://doi.org/10.1308/003588411x563411 https://doi.org/10.1136/qshc.2006.017947 https://doi.org/10.1136/qshc.2006.017947 https://doi.org/10.1177/0141076815578652 https://doi.org/10.1177/0141076815578652 https://doi.org/10.1136/bmjqs-2017-007756 43. Ock M, Lim SY, Jo MW et al (2017) Frequency, expected effects, obstacles, and facilitators of disclosure of patient safety incidents: a systematic review. J Prev Med Public Health 2017(50):68–82. https://doi.org/10.3961/jpmph.16.105 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1014 World J Surg (2022) 46:1006–1014 123 https://doi.org/10.3961/jpmph.16.105 Analysis of Surgical Mortalities Using the Fishbone Model for Quality Improvement in Surgical Disciplines Abstract Background Aim Methods Results Conclusion Introduction Methods Results Discussion Overview of mortality Factors associated with mortality Quality control factors associated with outcomes Limitations Conclusion Recommendations Funding References