ARTICLE IN PRESS
Predictors of surgical complications: A systematic review Annelies Visser, MSc, Bart Geboers, BSc, Dirk J. Gouma, MD, PhD, J. Carel Goslings, MD, PhD, and Dirk T. Ubbink, MD, PhD, Amsterdam, The Netherlands
Background. Operative complications occur more frequently, often are more preventable, and their consequences can be more severe than other types of complications. Controversy exists regarding how best to identify and predict operative complications. Several studies on predictive factors for operative complications focused on a specific predictor for a specific outcome. To develop a reliable tool to identify patients with operative complications, insight in predictive factors for operative complications is required. Patients and Methods. We searched all publications addressing predictive factors for the development of operative complications in adult patients admitted to the gastrointestinal, vascular, or general surgery departments. Data were extracted regarding study design, patient characteristics, operative specialty, types of operative procedures, types of complications, possible predictors, and associated complication risk increase (expressed as an odds ratio; OR). Results. The final set of 30 articles yielded a total of 53 predictive factors studied in various settings, operative specialties, and disorders. To focus our analysis we selected the 25 most robust and clinically applicable factors (ie, appearing in 3 or more studies). These factors were then categorized into 4 different groups: Patient-related factors, Co-morbidities, Laboratory values, and Surgery-related factors. The most predictive factors for morbidity in these groups were body mass index (ORs from 1.80 to 6.30), age (1.02– 4.62 years), American Society of Anesthesiologists classification (1.77–7.10), dyspnea (1.23–1.30), serum creatinine (1.39–2.14), emergency surgery (1.50–2.54), and functional status (1.36–4.07). Conclusion. This review presents a set of factors predictive of operative complications for general surgery departments. These easily retrievable factors can and should be validated in the specific patient populations of each hospital. (Surgery 2015;j:j-j.) From the Department of Surgery, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
APPROXIMATELY 40% OF IN-HOSPITAL COMPLICATIONS are related to operative procedures,1 whereas the incidence of complications is 2–4.5 times greater in surgery than in general medicine.1-3 Operative complications occur more often than other types of complications, are more often preventable, and their consequences are more severe.4 This knowledge charges us with the professional, moral, and ethical duty to minimize the incidence of these operative complications.5 This task requires accurate information about the incidence of perioperative complications, based on a reliable and comprehensive documentation. Such a complication registry also might Accepted for publication January 14, 2015. Reprint requests: Annelies Visser, MSc, Academic Medical Center, Department of Surgery, H1-213, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.01.012
help determine which complications are predictable and preventable. Such registries have been introduced by various operative specialties around the world and are considered essential for an optimum quality of care.6-10 Controversy exists, however, how best to identify these complications.11 Conventional approaches to identify complications, such as incident reports or voluntary complication reporting, had limited success in the identification of complications,11 whereas the use of a verbal inventory of complications during handover meetings also was found to be imperfect, identifying only 86% of all operative complications during hospitalization.12 Besides, it takes costly time from clinicians. Hence, hospitals would benefit from a more effective way to identify complications and to complete their registration. For this purpose, a trigger tool, consisting of a set of predictors, or thresholds, might be useful and more efficient to indicate patients at risk of a complication. Patient at risk in the context of a trigger tool should be defined as patients who SURGERY 1
ARTICLE IN PRESS 2 Visser et al
may have experienced complications during hospitalization.13,14 Trigger tools, comprising a range of predictors, currently are used to check the charts of patients at risk for complications, to complete the complication registry database.15,16 As an example, the Institute for Healthcare Improvement developed a global trigger tool.17 This trigger-based chart audit aims to provide a practical tool to identify patients at risk of complications after discharge. Substantial variations of global trigger tool findings, however, appeared to be common.15 This may be the result of the interobserver variation among different reviewers when identifying complications11,13,18 or a low rate of detection.19,20 Several studies on predictive factors for operative complications focused on a specific predictor for a specific outcome, eg, emergency admission as a predictor for death,21 or on surgical parameters (eg, surgery time) as predictors of postoperative cardiac events.22 Academic hospitals are highly subspecialized but, at least in Europe, they represent only a small part of all operative care, most of which is conducted in nonacademic centers. To develop a reliable tool to identify patients at risk for operative complications after discharge, an overview is needed of all predictive factors for operative complications in general. Hence, the aim of this systematic review was to search the literature for factors that may identify patients who have sustained an in-hospital complication (postoperative morbidity or mortality) in operative (general, gastrointestinal and vascular) patients. METHODS This systematic review was conducted along the STROBE guidelines.23 Search strategy. The search strategy was validated by a clinical librarian. We searched for publications addressing predictive factors for the development of operative complications in adult patients admitted to the gastrointestinal, vascular, or general surgery departments (see full search strategy in Appendix 1). Operative complications were defined as any medical adverse outcome occurring between admission and 30 days after operation. Any retrospective or prospective study design was accepted that focused on prognostic factors for operative complications. The study results should be presented as ratios for each predictive factor based on multivariable regression analysis. Case reports, abstracts, conference proceedings, and editorials were excluded from the search. Because of the wide search strategy and sensitive search
Surgery j 2015
terms we limited the study population to the 3 operative specialties mentioned previously. This limitation also was useful to find more specific predictive factors for each type of surgery. Depending on the number of predictive factors we would find, we also planned to select only the most robust and clinically applicable factors for further analysis. The Medline and Embase databases were searched from 2000 through March 2013. We did not use the Cochrane database because studies on harm are unlikely to be available in the form of RCTs. Eligible studies were selected and duplicates were removed (Fig). Study selection and appreciation. Titles and abstracts of the studies were judged for eligibility by 2 reviewers independently. Suitable articles that matched the predefined selection criteria were then obtained in full. The methodologic validity of the selected articles was critically appraised by 2 reviewers independently via the relevant checklists from the Dutch Cochrane center (http://dcc. cochrane.org/). In case of disagreement, consensus was reached through discussion among the reviewers. Data extraction and analysis. Data were extracted regarding study design, patient characteristics, operative specialty, types of surgical procedures, types of complications, possible predictive factors, and associated complication risk increase (odds ratio; OR). The ORs from multivariate regression analyses in the different studies, i.e., corrected for other confounders, were used to make general statements per predictive factor. A meta-analysis will be conducted if the patient characteristics and outcomes are homogenous. RESULTS Study descriptions. Our search rendered 648 hits, of which 39 matched our inclusion criteria (Fig). Eighteen of these described different predictive factors for overall mortality or morbidity in general surgery patients. General surgery includes all surgical areas; gastrointestinal, trauma, endocrine, oncology, and vascular surgery. No specific subspecialty was described. The remaining 21 articles addressed specific subspecialty, operations, complications, or predictive factors. After reading the full-text articles, 9 of these were excluded because no multivariable analysis was performed, no predictive factors were given, or the article tested an existing trigger tool and did not describe the contribution of individual predictive factors.
ARTICLE IN PRESS Visser et al 3
Surgery Volume j, Number j
Fig. Flow diagram of study inclusion.
The final set included 30 articles (Fig), comprising 11 articles focusing on general surgery, 9 on gastrointestinal surgery, 5 on vascular surgery, and 5 combining general and vascular surgery. Of these, 25 were retrospective and 5 were prospective cohort studies (Table I). All articles were published between 2001 and 2013. Most studies were conducted in the United States (27 studies) and some in Europe (3 studies) or Canada (1 study). Study size varied widely between 226 and 964,263 patients. Most of these studies derived their data from a prospective national database (National Surgical Quality Improvement Program). These data contained preoperative patient characteristics, clinical risk factors, and postoperative outcomes.
The 30 articles yielded a total of 53 predictive factors studied in various settings, specialties, and disorders. To focus our analysis, we selected the 25 most robust and clinically applicable predictive factors, using 2 criteria: (1) cited in 3 or more articles and (2) the factor is recorded routinely and readily available in hospital databases. The 25 predictive factors were then categorized into 4 different groups: Patient-related factors, Comorbidities, Laboratory values, and Surgery (procedure)-related factors. Associated complications were described in terms of morbidity and mortality. Because of the expected heterogeneity of the patient characteristics and outcomes, we refrained from conducting a metaanalysis.
ARTICLE IN PRESS 4 Visser et al
Surgery j 2015
Table I. Quality of included studies
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
Checked item
% of studies
Clear definition of study population? Exclusion of selection bias? Clear definition of exposure? Clear method to evaluate/assess exposure? Clear definition of outcome? Clear method to evaluate/assess outcome? Outcome determined blind from exposure? Affects this the evaluation of the outcome? Follow-up long enough? Selective lost to follow up excluded? Are confounders described?
100 60 77 67 90 63 10 0 93 90 93
Methodologic quality. Overall, the included studies were judged as valid and of reasonably good quality (Table I).24-53 In 10% of the studies, the outcomes were determined while researchers were unaware of the exposure. In 40% of all studies, selection bias could not be excluded because they excluded patient records with missing laboratory values or patient characteristics from their study cohort. Three of 4 studies described a clear method to evaluate exposure and outcome. Outcome variables. The 25 predictive factors we studied predicted a wide range of outcomes, eg, mortality or various morbidities (Appendix 2). Because of the heterogeneity of these results, we describe each predictive factor separately for the outcomes morbidity and mortality. The significant ORs found for each predictive factor are summarized in Table II. Table II also shows the number of studies showing significant ORs for each predictive factor compared with the total number of studies describing this factor. Patient-related factors. In this category, we found 9 significant factors (Table II). Most studies described age and body mass index (BMI) as predictive factors (15 and 9 studies, respectively). Both predictive factors showed ORs ranging from 1.03 to 5.32 and from 0.74 to 6.30, respectively. Age: Increasing age was found to be a predictive factor in 15 studies.25-37,40,42 Older age (with different cut-off values) was related to a greater risk of postoperative complications, resulting in greater morbidity and greater mortality. In 9 of the 15 studies, a risk increase was found using a cut-off value of greater than 65 years.26,28-30,33,34,36,37,42 BMI: BMI as a significant predictive factor for complications was found in 9 studies.26,29,32-36,38,39 Six of these studies found obesity (BMI >30 or
>35) predictive of morbidity.26,29,33,34,38,39 In contrast, in one study32 researchers found that underweight patients (BMI 1.5 mg/dL. This seems to predict a cardiac complication in particular, which was found in 4 of 5 studies investigating serum creatinine as predictive factor for morbidity.25,27,29,30 Three studies found that a high serum creatinine level also was predictive of mortality (using various cut-off values).27,33,42 Surgery-related factors. In this category we found 3 significant predictive factors (Table II). Most studies (n = 11) described ‘‘Emergency surgery’’ as a predictive factor. Emergency surgery: Emergency surgery was predictive of morbidity in seven general and vascular surgery studies compared with nonemergent operations, showing a range of ORs from 1.50 to 2.54.26-30,37,41 Patients who underwent emergency surgery also had a greater risk of mortality (OR from 1.90–3.33) than nonemergent surgery patients (5 studies).27,29,34,35,42
ARTICLE IN PRESS Visser et al 5
Surgery Volume j, Number j
Table II. Predictors per category
Predictor
Number of studies showing significant ORs/vs number of studies showing nonsignificant ORs
Odds ratio ranges (N)
Odds ratio ranges Morbidity (N)
15/16
1.03–5.32 (8)
1.02–4.62 (11)
7/7 3/3
0.74 (1) 1.40–1.52 (2)
1.80–6.30 (6) 1.90 (1)
6/6 3/3 7/7 6/6 4/4 5/6 4/4 4/6
1.19 2.70 1.43–3.78 1.30–1.98 1.55–3.09 0.75 — 1.65–2.40
(1) (1) (4) (3) (4) (1)
1.15–2.34 1.40–1.60 1.36–4.07 1.31–1.67 1.30 1.19–1.62 1.15–2.80 1.61–1.80
(6) (2) (4) (5) (1) (4) (4) (2)
12/13 7/8 6/7 5/5 5/5 4/4 3/3 4/6 3/3 3/3
1.54–11.60 1.22–6.25 1.37–2.87 2.10–2.97 1.31–1.97 1.80–3.54 — 1.37–10.98 2.30–5.70 —
(4) (5) (3) (3) (4) (3)
1.77–7.10 1.22–1.30 1.21–2.00 1.32–1.99 1.22–1.67 1.47–1.77 1.30–7.63 1.84–5.14 — 1.21–1.70
(8) (3) (5) (3) (3) (3) (3) (2)
6/6 5/5 4/4 2/3
1.70–1.71 1.33–2.20 1.82 1.33
(3) (2) (1) (1)
1.39–2.14 1.13–1.67 1.22–1.40 1.44–1.72
(5) (4) (3) (1)
Patient-related factors Age, y BMI Obesity Underweight Sex Male Female Functional status Long-term steroid use DNR Smoking Alcohol abuse Recent weight loss Comorbidities ASA classification Dyspnea Previous cardiac intervention/failure Preoperative sepsis COPD Ascites CVA Diabetes Dialysis Hypertension Laboratory values Increased creatinine Preoperative albumin Increased white blood cell count Hyponatremia Surgery-related factors Emergency operation Intraoperative transfusion Increase in operation time
10/11 3/4 3/3
(2)
(3) (3)
1.90–3.33 (5) 2.58 (1) —
(3)
1.50–2.54 (7) 1.09–2.60 (2) 1.00–2.20 (3)
Numbers of studies showing significant ORs and ranges mortality and morbidity (95% confidence interval). ASA, American Society of Anesthesiologists; BMI, body mass index; COPD, chronic obstructive pulmonary disease; CVA, cerebral vascular accident; DNR, do not resuscitate; OR, odds ratio.
DISCUSSION This systematic review rendered 53 predictive factors predictive of the development of operative complications (i.e., postoperative morbidity or mortality) in patients undergoing gastrointestinal, general, and vascular surgery, in the period up to 30 days after discharge. Of these, 25 were highly associated with operative complications and readily available from hospital databases. An additional advantage is that most of them indicate conditions already known before admission. This finding is in contrast with predictive factors associated with demand for care, which also are related to the
occurrence of complications but only become manifest during hospitalization.54 Some of these 25 patient-related and surgery-related comorbidities and laboratory tests may seem obvious potential predictors, but the value of a systematic review is to summarize available evidence. Here, we provide general surgery departments with a complete overview of possible predictors from the literature. Other, nonpatient-related factors also may influence the risk of operative complications, apart from patient-related factors. Studer and Inderbitzin55 reviewed surgery-related, nontechnical risk factors influencing the outcome of operative
ARTICLE IN PRESS 6 Visser et al
care. They found working hours, surgeon skills, handoffs, and checklists to be important factors predicting operative complications. These factors, however, are less easy to monitor and incorporate in a trigger tool. Some limitations of our review deserve mentioning. Many of the factors found here were studied for their association with a specific complication rather than any complication. Some studies focused on factors predicting complications for certain specific disorders and may not be indicative for complications in other circumstances. Moreover, the studies in this review applied different cut-off points or slightly different definitions of the exposure or outcome. Hence, we attempted to categorize the different predictive factors to summarize our findings. The great clinical heterogeneity among the included studies precludes a meta-analysis; however, our systematic review generated and supported the hypothesis that a set of these factors may predict the occurrence of a complication in operative patients. We selected factors based on the number of publications in which they were used. This does not necessarily mean these are most relevant, but reflects their importance in clinical practice. It should also be noted that 27 of the 30 studies we found were American, whereas 25 of these studies retrieved their data from the American College of Surgeons’ National Surgical Quality Improvement Program. This is an outcomebased, risk-adjusted program developed to improve the quality of surgical care in the United States. This validated database provides patientlevel information for operative procedures, is available for research, and provides a source of reliable clinical information.43 In the United States, researchers have easy access to this source of information, which may explain why most of the articles we found were of American origin. This might be a potential source of bias. Other countries may have different definitions of health and sickness and preferences regarding diagnoses and treatment differ among both surgeons and patients. Also, the facilities within the health care systems are likely to differ among countries. This is, however, no reason to assume their findings cannot be extrapolated to other Western countries. The set of factors as found in this systematic review may be useful for hospital care administration managers of general surgery departments to optimize their registration of complications. Despite the current era of analysing subspecialty-
Surgery j 2015
driven outcomes and predictive factors, the Dutch Government requires all general departments of surgery to use a general complication registry. The goal of complication registration consists of at least 3 aspects; first, to inform the patient about the outcome of a specific operative intervention in that particular center, second to review and analyze the data to evaluate the current process of care and therefore the quality of care, and third to use this data to develop and implement initiatives to improve the quality of care. Complication registration is an outcome-driven registration. It enables us to review trends in complication frequencies like increasing postoperative infections. These trends should be reviewed and analysed on the higher level of general surgery because the process or actions for improvement can transcend subspecialties. The risk of missing trends by analyzing smaller subgroups makes complication registration relevant to the general surgery department. Traditional efforts to scrutinize for complication development comprise voluntary reporting or incident reports of all patients. These methods often have been poorly successful in the detection of complications. Various other risk scoring systems have been introduced to identify surgical complications, like ASA classification,56 APACHE (Acute Physiology and Chronic Health Evaluation),57 and POSSUM (Psychological and Operative Severity Score for the Enumeration of mortality and morbidity)58; however, these systems also have their limitations,59 including interobserver variation (ASA), complexity (APACHE), and overestimation of mortality in lower risk groups (POSSUM). In this review, we did not take into account the severity of the complications detected, which also may be useful when weighing the impact of predictive factors.60 In this review the odds provided for each factor were only used to select the set of 25 most predictive factors as these may vary depending on the patient mix in which they were assessed. Hence, this set of factors should be validated in a specific patient population of each hospital for significance and cut-off values. The most predictive factors can then be used for the development of a customized trigger tool applicable to the specific patient mix admitted and treated in that hospital. We believe that a trigger tool based on these predictive factors, this patient mix, and these types of surgery might be most effective in helping hospitals to identify patients at risk of operative complications and to improve the current complication registry data. Further research should clarify
ARTICLE IN PRESS Surgery Volume j, Number j
the usefulness and external validity of trigger tool as compared to other means of identifying surgical complications in hospitalized patients. This review extracted 25 predictive factors from the literature that are all independently associated with greater risks of operative complications. Although many of the described predictors may be rather obvious, the results give a valuable overview of the literature and could be useful in guiding general surgery departments to focus on the right aspects in their search for clinical quality improvement. Most of these factors are easily retrievable from hospital data and can be validated for the case mix of a specific hospital. SUPPLEMENTARY DATA Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.surg.2015.01.012
REFERENCES 1. Vries de EN, Ramrattan MA, Smorenburg SM, Gouma DJ, Boermeester MA. The incidence and nature of in-hospital adverse events: a systematic review. Qual Saf Health Care 2008;17:216-23. 2. Brennan TA, Leape LL, Laird NM, Hebert L, Localio AR, Lawthers AG, et al. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. Qual Saf Health Care 2004;13:145-52. 3. Baines RJ, Langelaan M, de Bruijne MC, Asscheman H, Spreeuwenberg P, van de Steeg L, et al. Changes in adverse event rates in hospitals over time: a longitudinal retrospective patient record review study. BMJ Qual Saf 2013;22:290-8. 4. Zegers M, Bruijne de MC, Keizer de B, Merten H, Groenewegen PP, van der Wal G, et al. The incidence, root-causes, and outcomes of adverse events in surgical units: implication for potential prevention strategies. Patient Saf Surg 2011;5:13. 5. Visser A, Ubbink DT, van Wijngaarden AK, Gouma DJ, Goslings JC. Quality of care and analysis of surgical complications. Dig Surg 2012;29:391-9. 6. Bosma E, Veen EJ, Jongh de MA, Roukema JA. Variable impact of complications in general surgery: a prospective cohort study. Can J Surg 2012;55:163-70. 7. Lawson EH, Louie R, Zingmond DS, Brook RH, Hall BL, Han L, et al. A comparison of clinical registry versus administrative claims data for reporting of 30-day surgical complications. Ann Surg 2012;256:973-81. 8. Hlaing T, Hollister L, Aaland M. Trauma registry data validation: essential for quality trauma care. J Trauma 2006; 61:1400-7. 9. Gunnarsson U, Seligsohn E, Jestin P, P ahlman L. Registration and validity of surgical complications in colorectal cancer surgery. Br J Surg 2003;90:454-9. 10. Bruce JL, Russell EM, Mollison J, Krukowski ZH. The measurement and monitoring of surgical adverse events. Health Technol Assess 2001;5:1-194. 11. Resar RK, Rozich JD, Simmonds T, Haraden CRA. Trigger tool to identify adverse events in the intensive care unit. Jt Comm J Qual Patient Saf 2006;32:585-90.
Visser et al 7
12. Ubbink DT, Visser A, Gouma DJ, Goslings JC. Registration of surgical adverse outcomes: a reliability study in a university hospital. BMJ Open 2012;2. 13. Griffin FA, Classen DC. Detection of adverse events in surgical patients using the Trigger Tool approach. Qual Saf Health Care 2008;17:253-8. 14. Boer de M, Kiewiet JJ, Boeker EB, Ramrattan MA, Dijkgraaf MG, Lie-A-Huen L, et al. SUREPILL Study Group. A targeted method for standardized assessment of adverse drug events in surgical patients. J Eval Clin Pract 2013;19: 1073-82. 15. Plessen von C, Kodal AM, Anhoj J. Experiences with global trigger tool reviews in five Danish hospitals: an implementation study. BMJ Open 2012;2:e001324. 16. Moore C, Li J, Hung C, Downs J, Nebeker JR. Predictive value of alert triggers for identification. J Patient Saf 2009; 5:223-8. 17. Griffin FA, Resar RK. IHI Global Trigger Tool for Measuring Adverse Events. IHI Innovation Series white paper. 2nd ed. Cambridge, MA: Institute for Healthcare Improvement; 2009. 18. O’Leary KJ, Devisetty VK, Patel AR, Malkenson D, Sama P, Thompson WK, et al. Comparison of traditional trigger tool to data warehouse based screening for identifying hospital adverse events. BMJ Qual Saf 2013;22:130-8. 19. Forster AJ, Andrade J, Van Walraven C. Validation of a discharge summary term search method to detect adverse events. J Am Med Inform Assoc 2005;12:200-6. 20. Melton GB, Hripcsak G. Automated detection of adverse events using natural language processing of discharge summaries. J Am Med Inform Assoc 2005;12:448-57. 21. Boulain T, Runge I, Delorme N, Bouju A, Valery A. Patients hospitalized in general wards via the emergency department early identification of predisposing factors of death or unexpected intensive care unit admission---a historical prospective. Emerg Med Int 2014;2014:203747. 22. Bae MH, Lee JH, Yang DH, Park HS, Chae SC. Usefulness of surgical parameters as predictors of postoperative cardiac events in patients undergoing non-cardiac surgery. Circ J 2014;78:718-23. 23. Elm von E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vanderbroucke JP, the STROBE initiative. The strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007;370:1453-7. 24. Gupta H, Gupta PK, Fang X, Miller WJ, Cemaj S, Forse RA, et al. Development and validation of a risk calculator predicting postoperative respiratory failure. Chest 2011;140: 1207-15. 25. Gupta PK, Gupta H, Sundaram A, Kaushik M, Fang X, Miller WJ, et al. Development and validation of a risk calculator for prediction of cardiac risk after surgery. Circulation 2011;124:381-7. 26. Kheterpal S, O’Reilly M, Englesbe MJ, Rosenberg AL, Shanks AM, Zhang L, et al. Preoperative and intraoperative predictors of cardiac adverse events after general, vascular, and urological surgery. Anesthesiology 2009;110:58-66. 27. O’Brien MM, Gonzales R, Shroyer AL, Grunwald GK, Daley J, Henderson WG, et al. Modest serum creatinine elevation affects adverse outcome after general surgery. Kidney Int 2002;62:585-92. 28. Veltkamp SC, Kemmeren JM, van der Graaf Y, Edlinger M, van der Werken C. Prediction of serious complications in patients admitted to a surgical ward. Br J Surg 2002;89: 94-102.
ARTICLE IN PRESS 8 Visser et al
29. Crawford RS, Cambria RP, Abularrage CJ, Conrad MF, Lancaster RT, Watkins MT, et al. Preoperative functional status predicts perioperative outcomes after infrainguinal bypass surgery. J Vasc Surg 2010;51:351-8. 30. Davenport DL, Ferraris VA, Hosokawa P, Henderson WG, Khuri SF, Mentzer RM Jr. Multivariable predictors of postoperative cardiac adverse events after general and vascular surgery: results from the patient safety in surgery study. J Am Coll Surg 2007;204:1199-210. 31. Dhungel B, Diggs BS, Hunter JG, Sheppard BC, Vetto JT, Dolan JP. Patient and peri-operative predictors of morbidity and mortality after esophagectomy: American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), 2005-2008. J Gastrointest Surg 2010;14: 1492-501. 32. Giles KA, Wyers MC, Pomposelli FB, Hamdan AD, Ching YA, Schermerhorn ML. The impact of body mass index on perioperative outcomes of open and endovascular abdominal aortic aneurysm repair from the National Surgical Quality Improvement Program, 2005-2007. J Vasc Surg 2010;52:1471-7. 33. Greenblatt DY, Kelly KJ, Rajamanickam V, Wan Y, Hanson T, Rettammel R, et al. Preoperative factors predict perioperative morbidity and mortality after pancreaticoduodenectomy. Ann Surg Oncol 2011;18:2126-35. 34. Kneuertz PJ, Pitt HA, Bilimoria KY, Smiley JP, Cohen ME, Ko CY, et al. Risk of morbidity and mortality following hepato-pancreato-biliary surgery. J Gastrointest Surg 2012; 16:1727-35. 35. Nafiu OO, Kheterpal S, Moulding R, Picton P, Tremper KK, Campbell DA Jr, et al. The association of body mass index to postoperative outcomes in elderly vascular surgery patients: a reverse J-curve phenomenon. Anesth Analg 2011;112:23-9. 36. Nelson MT, Greenblatt DY, Soma G, Rajimanickam V, Greenberg CC, Kent KC. Preoperative factors predict mortality after major lower-extremity amputation. Surgery 2012;152:685-94. 37. Neumayer L, Hosokawa P, Itani K, El-Tamer M, Henderson WG, Khuri SF. Multivariable predictors of postoperative surgical site infection after general and vascular surgery: results from the patient safety in surgery study. J Am Coll Surg 2007;204:1178-87. 38. Johnson ON III, Sidawy AN, Scanlon JM, Walcott R, Arora S, Macsata RA, et al. Impact of obesity on outcomes after open surgical and endovascular abdominal aortic aneurysm repair. J Am Coll Surg 2010;210:166-77. 39. Mathur AK, Ghaferi AA, Osborne NH, Pawlik TM, Campbell DA, Englesbe MJ, et al. Body mass index and adverse perioperative outcomes following hepatic resection. J Gastrointest Surg 2010;14:1285-91. 40. Farhat J, Velanovich V, Falvo J, Horst MH, Swartz A, Patton JH, et al. Are the frail destined to fail? Frailty index as predictor of surgical morbidity and mortality in the elderly. J Trauma Acute Care Surg 2012;72:1526-31. 41. Glance LG, Lustik SJ, Hannan EL, Osler TM, Mukamel DB, Qian F, et al. The Surgical Mortality Probability Model: derivation and validation of a simple risk prediction rule for noncardiac surgery. Ann Surg 2012;255:696-702. 42. Kazaure H, Roman S, Sosa JA. High mortality in surgical patients with do-not-resuscitate orders: analysis of 8256 patients. Arch Surg 2011;146:922-8. 43. Breitenstein S, DeOliveira ML, Raptis DA, Slankamenac K, Kambakamba P, Nerl J, et al. Novel and simple preoperative
Surgery j 2015
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55. 56. 57.
58. 59.
60.
score predicting complications after liver resection in noncirrhotic patients. Ann Surg 2010;252:726-34. Obeid NM, Azuh O, Reddy S, Webb S, Reickert C, Velanovich V, et al. Predictors of critical care-related complications in colectomy patients using the National Surgical Quality Improvement Program: exploring frailty and aggressive laparoscopic approaches. J Trauma Acute Care Surg 2012; 72:878-83. Ata A, Lee J, Bestle SL, Desemone J, Stain SC. Postoperative hyperglycemia and surgical site infection in general surgery patients. Arch Surg 2010;145:858-64. Halfon P, Egglo Y, Matter M, Kallay C, Melle van G, Burnand B. Risk-adjusted rates for potentially avoidable reoperations were computed from routine hospital data. J Clin Epidemiol 2007;60:56-67. Leung AA, McAlister FA, Rogers SO Jr, Pazo V, Wright A, Bates DW. Preoperative hyponatremia and perioperative complications. Arch Intern Med 2012;172:1474-81. Nath B, Li Y, Carroll JE, Szabo G, Tseng JF, Shah SA. Alcohol exposure as a risk factor for adverse outcomes in elective surgery. J Gastrointest Surg 2010;14:1732-41. Saxton A, Velanovich V. Preoperative Frailty and Quality of Life as Predictors of Postoperative Complications. Ann Surg 2011;253:1223-9. Leichtle SW, Mouawad NJ, Lampman R, Singal B, Cleary RK. Does Preoperative Anemia Adversely Affect Colon and Rectal Surgery Outcomes? J Am Coll Surg 2011;212: 187-94. Regenbogen SE, Bordeianou L, Hutter M, Gawande AA. The intraoperative Surgical Apgar Score predicts postdischarge complications after colon and rectal resection. Surgery 2010;148:559-66. Scarborough JE, Pappas TN, Cox MW, Bennett KM, Shortell CK. Surgical trainee participation during infrainguinal bypass grafting procedures is associated with increased early postoperative graft failure. J Vasc Surg 2012;55:715-20. Trencheva K, Morrissey KP, Wells M, Mancuso CA, Lee SW, Sonoda T, et al. Identifying important predictors for anastomotic leak after colon and rectal resection: prospective study on 616 patients. Ann Surg 2013;257:108-13. Oostveen van CJ, Ubbink DT, Huis In Het Veld JG, Bakker PJ, Vermeulen H. Factors and models associated with the amount of hospital care services as demanded by hospitalized patients: a systematic review. PLoS One 2014;9:e98102. Studer P, Inderbitzin D. Surgery-related risk factors. Curr Opin Crit Care 2009;15:328-32. Saklad M. Grading of patients for surgical procedures. Anesthesiology 1941;2:281-4. Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. APACHE– Acute Physiology and Chronic Health Evaluation: a physiologically based classification system. Crit Care Med 1981;9:591-7. Copeland GH, Jones D, Walters M. POSSUM: a scoring system for surgical audit. Br J Surg 1991;78:355-60. Shah N, Hamilton M. Clinical review: Can we predict which patients are at risk of complications following surgery? Crit Care 2013;17:226. Vollmer CM, Lewis RS, Hall BL, Allendorf JD, Beane JD, Behrman SW, et al. Establishing a Quantitative Benchmark for Morbidity in Pancreatoduodenectomy using ACSNSQIP, the Accordion Severity Grading System, and Postopererative Morbidity Index. Ann Surg 2015;261:527-36.
Predictors of surgical complications: A systematic review Annelies Visser, MSc, Bart Geboers, BSc, Dirk J. Gouma, MD, PhD, J. Carel Goslings, MD, PhD, and Dirk T. Ubbink, MD, PhD, Amsterdam, The Netherlands
Background. Operative complications occur more frequently, often are more preventable, and their consequences can be more severe than other types of complications. Controversy exists regarding how best to identify and predict operative complications. Several studies on predictive factors for operative complications focused on a specific predictor for a specific outcome. To develop a reliable tool to identify patients with operative complications, insight in predictive factors for operative complications is required. Patients and Methods. We searched all publications addressing predictive factors for the development of operative complications in adult patients admitted to the gastrointestinal, vascular, or general surgery departments. Data were extracted regarding study design, patient characteristics, operative specialty, types of operative procedures, types of complications, possible predictors, and associated complication risk increase (expressed as an odds ratio; OR). Results. The final set of 30 articles yielded a total of 53 predictive factors studied in various settings, operative specialties, and disorders. To focus our analysis we selected the 25 most robust and clinically applicable factors (ie, appearing in 3 or more studies). These factors were then categorized into 4 different groups: Patient-related factors, Co-morbidities, Laboratory values, and Surgery-related factors. The most predictive factors for morbidity in these groups were body mass index (ORs from 1.80 to 6.30), age (1.02– 4.62 years), American Society of Anesthesiologists classification (1.77–7.10), dyspnea (1.23–1.30), serum creatinine (1.39–2.14), emergency surgery (1.50–2.54), and functional status (1.36–4.07). Conclusion. This review presents a set of factors predictive of operative complications for general surgery departments. These easily retrievable factors can and should be validated in the specific patient populations of each hospital. (Surgery 2015;j:j-j.) From the Department of Surgery, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands
APPROXIMATELY 40% OF IN-HOSPITAL COMPLICATIONS are related to operative procedures,1 whereas the incidence of complications is 2–4.5 times greater in surgery than in general medicine.1-3 Operative complications occur more often than other types of complications, are more often preventable, and their consequences are more severe.4 This knowledge charges us with the professional, moral, and ethical duty to minimize the incidence of these operative complications.5 This task requires accurate information about the incidence of perioperative complications, based on a reliable and comprehensive documentation. Such a complication registry also might Accepted for publication January 14, 2015. Reprint requests: Annelies Visser, MSc, Academic Medical Center, Department of Surgery, H1-213, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.01.012
help determine which complications are predictable and preventable. Such registries have been introduced by various operative specialties around the world and are considered essential for an optimum quality of care.6-10 Controversy exists, however, how best to identify these complications.11 Conventional approaches to identify complications, such as incident reports or voluntary complication reporting, had limited success in the identification of complications,11 whereas the use of a verbal inventory of complications during handover meetings also was found to be imperfect, identifying only 86% of all operative complications during hospitalization.12 Besides, it takes costly time from clinicians. Hence, hospitals would benefit from a more effective way to identify complications and to complete their registration. For this purpose, a trigger tool, consisting of a set of predictors, or thresholds, might be useful and more efficient to indicate patients at risk of a complication. Patient at risk in the context of a trigger tool should be defined as patients who SURGERY 1
ARTICLE IN PRESS 2 Visser et al
may have experienced complications during hospitalization.13,14 Trigger tools, comprising a range of predictors, currently are used to check the charts of patients at risk for complications, to complete the complication registry database.15,16 As an example, the Institute for Healthcare Improvement developed a global trigger tool.17 This trigger-based chart audit aims to provide a practical tool to identify patients at risk of complications after discharge. Substantial variations of global trigger tool findings, however, appeared to be common.15 This may be the result of the interobserver variation among different reviewers when identifying complications11,13,18 or a low rate of detection.19,20 Several studies on predictive factors for operative complications focused on a specific predictor for a specific outcome, eg, emergency admission as a predictor for death,21 or on surgical parameters (eg, surgery time) as predictors of postoperative cardiac events.22 Academic hospitals are highly subspecialized but, at least in Europe, they represent only a small part of all operative care, most of which is conducted in nonacademic centers. To develop a reliable tool to identify patients at risk for operative complications after discharge, an overview is needed of all predictive factors for operative complications in general. Hence, the aim of this systematic review was to search the literature for factors that may identify patients who have sustained an in-hospital complication (postoperative morbidity or mortality) in operative (general, gastrointestinal and vascular) patients. METHODS This systematic review was conducted along the STROBE guidelines.23 Search strategy. The search strategy was validated by a clinical librarian. We searched for publications addressing predictive factors for the development of operative complications in adult patients admitted to the gastrointestinal, vascular, or general surgery departments (see full search strategy in Appendix 1). Operative complications were defined as any medical adverse outcome occurring between admission and 30 days after operation. Any retrospective or prospective study design was accepted that focused on prognostic factors for operative complications. The study results should be presented as ratios for each predictive factor based on multivariable regression analysis. Case reports, abstracts, conference proceedings, and editorials were excluded from the search. Because of the wide search strategy and sensitive search
Surgery j 2015
terms we limited the study population to the 3 operative specialties mentioned previously. This limitation also was useful to find more specific predictive factors for each type of surgery. Depending on the number of predictive factors we would find, we also planned to select only the most robust and clinically applicable factors for further analysis. The Medline and Embase databases were searched from 2000 through March 2013. We did not use the Cochrane database because studies on harm are unlikely to be available in the form of RCTs. Eligible studies were selected and duplicates were removed (Fig). Study selection and appreciation. Titles and abstracts of the studies were judged for eligibility by 2 reviewers independently. Suitable articles that matched the predefined selection criteria were then obtained in full. The methodologic validity of the selected articles was critically appraised by 2 reviewers independently via the relevant checklists from the Dutch Cochrane center (http://dcc. cochrane.org/). In case of disagreement, consensus was reached through discussion among the reviewers. Data extraction and analysis. Data were extracted regarding study design, patient characteristics, operative specialty, types of surgical procedures, types of complications, possible predictive factors, and associated complication risk increase (odds ratio; OR). The ORs from multivariate regression analyses in the different studies, i.e., corrected for other confounders, were used to make general statements per predictive factor. A meta-analysis will be conducted if the patient characteristics and outcomes are homogenous. RESULTS Study descriptions. Our search rendered 648 hits, of which 39 matched our inclusion criteria (Fig). Eighteen of these described different predictive factors for overall mortality or morbidity in general surgery patients. General surgery includes all surgical areas; gastrointestinal, trauma, endocrine, oncology, and vascular surgery. No specific subspecialty was described. The remaining 21 articles addressed specific subspecialty, operations, complications, or predictive factors. After reading the full-text articles, 9 of these were excluded because no multivariable analysis was performed, no predictive factors were given, or the article tested an existing trigger tool and did not describe the contribution of individual predictive factors.
ARTICLE IN PRESS Visser et al 3
Surgery Volume j, Number j
Fig. Flow diagram of study inclusion.
The final set included 30 articles (Fig), comprising 11 articles focusing on general surgery, 9 on gastrointestinal surgery, 5 on vascular surgery, and 5 combining general and vascular surgery. Of these, 25 were retrospective and 5 were prospective cohort studies (Table I). All articles were published between 2001 and 2013. Most studies were conducted in the United States (27 studies) and some in Europe (3 studies) or Canada (1 study). Study size varied widely between 226 and 964,263 patients. Most of these studies derived their data from a prospective national database (National Surgical Quality Improvement Program). These data contained preoperative patient characteristics, clinical risk factors, and postoperative outcomes.
The 30 articles yielded a total of 53 predictive factors studied in various settings, specialties, and disorders. To focus our analysis, we selected the 25 most robust and clinically applicable predictive factors, using 2 criteria: (1) cited in 3 or more articles and (2) the factor is recorded routinely and readily available in hospital databases. The 25 predictive factors were then categorized into 4 different groups: Patient-related factors, Comorbidities, Laboratory values, and Surgery (procedure)-related factors. Associated complications were described in terms of morbidity and mortality. Because of the expected heterogeneity of the patient characteristics and outcomes, we refrained from conducting a metaanalysis.
ARTICLE IN PRESS 4 Visser et al
Surgery j 2015
Table I. Quality of included studies
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
Checked item
% of studies
Clear definition of study population? Exclusion of selection bias? Clear definition of exposure? Clear method to evaluate/assess exposure? Clear definition of outcome? Clear method to evaluate/assess outcome? Outcome determined blind from exposure? Affects this the evaluation of the outcome? Follow-up long enough? Selective lost to follow up excluded? Are confounders described?
100 60 77 67 90 63 10 0 93 90 93
Methodologic quality. Overall, the included studies were judged as valid and of reasonably good quality (Table I).24-53 In 10% of the studies, the outcomes were determined while researchers were unaware of the exposure. In 40% of all studies, selection bias could not be excluded because they excluded patient records with missing laboratory values or patient characteristics from their study cohort. Three of 4 studies described a clear method to evaluate exposure and outcome. Outcome variables. The 25 predictive factors we studied predicted a wide range of outcomes, eg, mortality or various morbidities (Appendix 2). Because of the heterogeneity of these results, we describe each predictive factor separately for the outcomes morbidity and mortality. The significant ORs found for each predictive factor are summarized in Table II. Table II also shows the number of studies showing significant ORs for each predictive factor compared with the total number of studies describing this factor. Patient-related factors. In this category, we found 9 significant factors (Table II). Most studies described age and body mass index (BMI) as predictive factors (15 and 9 studies, respectively). Both predictive factors showed ORs ranging from 1.03 to 5.32 and from 0.74 to 6.30, respectively. Age: Increasing age was found to be a predictive factor in 15 studies.25-37,40,42 Older age (with different cut-off values) was related to a greater risk of postoperative complications, resulting in greater morbidity and greater mortality. In 9 of the 15 studies, a risk increase was found using a cut-off value of greater than 65 years.26,28-30,33,34,36,37,42 BMI: BMI as a significant predictive factor for complications was found in 9 studies.26,29,32-36,38,39 Six of these studies found obesity (BMI >30 or
>35) predictive of morbidity.26,29,33,34,38,39 In contrast, in one study32 researchers found that underweight patients (BMI 1.5 mg/dL. This seems to predict a cardiac complication in particular, which was found in 4 of 5 studies investigating serum creatinine as predictive factor for morbidity.25,27,29,30 Three studies found that a high serum creatinine level also was predictive of mortality (using various cut-off values).27,33,42 Surgery-related factors. In this category we found 3 significant predictive factors (Table II). Most studies (n = 11) described ‘‘Emergency surgery’’ as a predictive factor. Emergency surgery: Emergency surgery was predictive of morbidity in seven general and vascular surgery studies compared with nonemergent operations, showing a range of ORs from 1.50 to 2.54.26-30,37,41 Patients who underwent emergency surgery also had a greater risk of mortality (OR from 1.90–3.33) than nonemergent surgery patients (5 studies).27,29,34,35,42
ARTICLE IN PRESS Visser et al 5
Surgery Volume j, Number j
Table II. Predictors per category
Predictor
Number of studies showing significant ORs/vs number of studies showing nonsignificant ORs
Odds ratio ranges (N)
Odds ratio ranges Morbidity (N)
15/16
1.03–5.32 (8)
1.02–4.62 (11)
7/7 3/3
0.74 (1) 1.40–1.52 (2)
1.80–6.30 (6) 1.90 (1)
6/6 3/3 7/7 6/6 4/4 5/6 4/4 4/6
1.19 2.70 1.43–3.78 1.30–1.98 1.55–3.09 0.75 — 1.65–2.40
(1) (1) (4) (3) (4) (1)
1.15–2.34 1.40–1.60 1.36–4.07 1.31–1.67 1.30 1.19–1.62 1.15–2.80 1.61–1.80
(6) (2) (4) (5) (1) (4) (4) (2)
12/13 7/8 6/7 5/5 5/5 4/4 3/3 4/6 3/3 3/3
1.54–11.60 1.22–6.25 1.37–2.87 2.10–2.97 1.31–1.97 1.80–3.54 — 1.37–10.98 2.30–5.70 —
(4) (5) (3) (3) (4) (3)
1.77–7.10 1.22–1.30 1.21–2.00 1.32–1.99 1.22–1.67 1.47–1.77 1.30–7.63 1.84–5.14 — 1.21–1.70
(8) (3) (5) (3) (3) (3) (3) (2)
6/6 5/5 4/4 2/3
1.70–1.71 1.33–2.20 1.82 1.33
(3) (2) (1) (1)
1.39–2.14 1.13–1.67 1.22–1.40 1.44–1.72
(5) (4) (3) (1)
Patient-related factors Age, y BMI Obesity Underweight Sex Male Female Functional status Long-term steroid use DNR Smoking Alcohol abuse Recent weight loss Comorbidities ASA classification Dyspnea Previous cardiac intervention/failure Preoperative sepsis COPD Ascites CVA Diabetes Dialysis Hypertension Laboratory values Increased creatinine Preoperative albumin Increased white blood cell count Hyponatremia Surgery-related factors Emergency operation Intraoperative transfusion Increase in operation time
10/11 3/4 3/3
(2)
(3) (3)
1.90–3.33 (5) 2.58 (1) —
(3)
1.50–2.54 (7) 1.09–2.60 (2) 1.00–2.20 (3)
Numbers of studies showing significant ORs and ranges mortality and morbidity (95% confidence interval). ASA, American Society of Anesthesiologists; BMI, body mass index; COPD, chronic obstructive pulmonary disease; CVA, cerebral vascular accident; DNR, do not resuscitate; OR, odds ratio.
DISCUSSION This systematic review rendered 53 predictive factors predictive of the development of operative complications (i.e., postoperative morbidity or mortality) in patients undergoing gastrointestinal, general, and vascular surgery, in the period up to 30 days after discharge. Of these, 25 were highly associated with operative complications and readily available from hospital databases. An additional advantage is that most of them indicate conditions already known before admission. This finding is in contrast with predictive factors associated with demand for care, which also are related to the
occurrence of complications but only become manifest during hospitalization.54 Some of these 25 patient-related and surgery-related comorbidities and laboratory tests may seem obvious potential predictors, but the value of a systematic review is to summarize available evidence. Here, we provide general surgery departments with a complete overview of possible predictors from the literature. Other, nonpatient-related factors also may influence the risk of operative complications, apart from patient-related factors. Studer and Inderbitzin55 reviewed surgery-related, nontechnical risk factors influencing the outcome of operative
ARTICLE IN PRESS 6 Visser et al
care. They found working hours, surgeon skills, handoffs, and checklists to be important factors predicting operative complications. These factors, however, are less easy to monitor and incorporate in a trigger tool. Some limitations of our review deserve mentioning. Many of the factors found here were studied for their association with a specific complication rather than any complication. Some studies focused on factors predicting complications for certain specific disorders and may not be indicative for complications in other circumstances. Moreover, the studies in this review applied different cut-off points or slightly different definitions of the exposure or outcome. Hence, we attempted to categorize the different predictive factors to summarize our findings. The great clinical heterogeneity among the included studies precludes a meta-analysis; however, our systematic review generated and supported the hypothesis that a set of these factors may predict the occurrence of a complication in operative patients. We selected factors based on the number of publications in which they were used. This does not necessarily mean these are most relevant, but reflects their importance in clinical practice. It should also be noted that 27 of the 30 studies we found were American, whereas 25 of these studies retrieved their data from the American College of Surgeons’ National Surgical Quality Improvement Program. This is an outcomebased, risk-adjusted program developed to improve the quality of surgical care in the United States. This validated database provides patientlevel information for operative procedures, is available for research, and provides a source of reliable clinical information.43 In the United States, researchers have easy access to this source of information, which may explain why most of the articles we found were of American origin. This might be a potential source of bias. Other countries may have different definitions of health and sickness and preferences regarding diagnoses and treatment differ among both surgeons and patients. Also, the facilities within the health care systems are likely to differ among countries. This is, however, no reason to assume their findings cannot be extrapolated to other Western countries. The set of factors as found in this systematic review may be useful for hospital care administration managers of general surgery departments to optimize their registration of complications. Despite the current era of analysing subspecialty-
Surgery j 2015
driven outcomes and predictive factors, the Dutch Government requires all general departments of surgery to use a general complication registry. The goal of complication registration consists of at least 3 aspects; first, to inform the patient about the outcome of a specific operative intervention in that particular center, second to review and analyze the data to evaluate the current process of care and therefore the quality of care, and third to use this data to develop and implement initiatives to improve the quality of care. Complication registration is an outcome-driven registration. It enables us to review trends in complication frequencies like increasing postoperative infections. These trends should be reviewed and analysed on the higher level of general surgery because the process or actions for improvement can transcend subspecialties. The risk of missing trends by analyzing smaller subgroups makes complication registration relevant to the general surgery department. Traditional efforts to scrutinize for complication development comprise voluntary reporting or incident reports of all patients. These methods often have been poorly successful in the detection of complications. Various other risk scoring systems have been introduced to identify surgical complications, like ASA classification,56 APACHE (Acute Physiology and Chronic Health Evaluation),57 and POSSUM (Psychological and Operative Severity Score for the Enumeration of mortality and morbidity)58; however, these systems also have their limitations,59 including interobserver variation (ASA), complexity (APACHE), and overestimation of mortality in lower risk groups (POSSUM). In this review, we did not take into account the severity of the complications detected, which also may be useful when weighing the impact of predictive factors.60 In this review the odds provided for each factor were only used to select the set of 25 most predictive factors as these may vary depending on the patient mix in which they were assessed. Hence, this set of factors should be validated in a specific patient population of each hospital for significance and cut-off values. The most predictive factors can then be used for the development of a customized trigger tool applicable to the specific patient mix admitted and treated in that hospital. We believe that a trigger tool based on these predictive factors, this patient mix, and these types of surgery might be most effective in helping hospitals to identify patients at risk of operative complications and to improve the current complication registry data. Further research should clarify
ARTICLE IN PRESS Surgery Volume j, Number j
the usefulness and external validity of trigger tool as compared to other means of identifying surgical complications in hospitalized patients. This review extracted 25 predictive factors from the literature that are all independently associated with greater risks of operative complications. Although many of the described predictors may be rather obvious, the results give a valuable overview of the literature and could be useful in guiding general surgery departments to focus on the right aspects in their search for clinical quality improvement. Most of these factors are easily retrievable from hospital data and can be validated for the case mix of a specific hospital. SUPPLEMENTARY DATA Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.surg.2015.01.012
REFERENCES 1. Vries de EN, Ramrattan MA, Smorenburg SM, Gouma DJ, Boermeester MA. The incidence and nature of in-hospital adverse events: a systematic review. Qual Saf Health Care 2008;17:216-23. 2. Brennan TA, Leape LL, Laird NM, Hebert L, Localio AR, Lawthers AG, et al. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. Qual Saf Health Care 2004;13:145-52. 3. Baines RJ, Langelaan M, de Bruijne MC, Asscheman H, Spreeuwenberg P, van de Steeg L, et al. Changes in adverse event rates in hospitals over time: a longitudinal retrospective patient record review study. BMJ Qual Saf 2013;22:290-8. 4. Zegers M, Bruijne de MC, Keizer de B, Merten H, Groenewegen PP, van der Wal G, et al. The incidence, root-causes, and outcomes of adverse events in surgical units: implication for potential prevention strategies. Patient Saf Surg 2011;5:13. 5. Visser A, Ubbink DT, van Wijngaarden AK, Gouma DJ, Goslings JC. Quality of care and analysis of surgical complications. Dig Surg 2012;29:391-9. 6. Bosma E, Veen EJ, Jongh de MA, Roukema JA. Variable impact of complications in general surgery: a prospective cohort study. Can J Surg 2012;55:163-70. 7. Lawson EH, Louie R, Zingmond DS, Brook RH, Hall BL, Han L, et al. A comparison of clinical registry versus administrative claims data for reporting of 30-day surgical complications. Ann Surg 2012;256:973-81. 8. Hlaing T, Hollister L, Aaland M. Trauma registry data validation: essential for quality trauma care. J Trauma 2006; 61:1400-7. 9. Gunnarsson U, Seligsohn E, Jestin P, P ahlman L. Registration and validity of surgical complications in colorectal cancer surgery. Br J Surg 2003;90:454-9. 10. Bruce JL, Russell EM, Mollison J, Krukowski ZH. The measurement and monitoring of surgical adverse events. Health Technol Assess 2001;5:1-194. 11. Resar RK, Rozich JD, Simmonds T, Haraden CRA. Trigger tool to identify adverse events in the intensive care unit. Jt Comm J Qual Patient Saf 2006;32:585-90.
Visser et al 7
12. Ubbink DT, Visser A, Gouma DJ, Goslings JC. Registration of surgical adverse outcomes: a reliability study in a university hospital. BMJ Open 2012;2. 13. Griffin FA, Classen DC. Detection of adverse events in surgical patients using the Trigger Tool approach. Qual Saf Health Care 2008;17:253-8. 14. Boer de M, Kiewiet JJ, Boeker EB, Ramrattan MA, Dijkgraaf MG, Lie-A-Huen L, et al. SUREPILL Study Group. A targeted method for standardized assessment of adverse drug events in surgical patients. J Eval Clin Pract 2013;19: 1073-82. 15. Plessen von C, Kodal AM, Anhoj J. Experiences with global trigger tool reviews in five Danish hospitals: an implementation study. BMJ Open 2012;2:e001324. 16. Moore C, Li J, Hung C, Downs J, Nebeker JR. Predictive value of alert triggers for identification. J Patient Saf 2009; 5:223-8. 17. Griffin FA, Resar RK. IHI Global Trigger Tool for Measuring Adverse Events. IHI Innovation Series white paper. 2nd ed. Cambridge, MA: Institute for Healthcare Improvement; 2009. 18. O’Leary KJ, Devisetty VK, Patel AR, Malkenson D, Sama P, Thompson WK, et al. Comparison of traditional trigger tool to data warehouse based screening for identifying hospital adverse events. BMJ Qual Saf 2013;22:130-8. 19. Forster AJ, Andrade J, Van Walraven C. Validation of a discharge summary term search method to detect adverse events. J Am Med Inform Assoc 2005;12:200-6. 20. Melton GB, Hripcsak G. Automated detection of adverse events using natural language processing of discharge summaries. J Am Med Inform Assoc 2005;12:448-57. 21. Boulain T, Runge I, Delorme N, Bouju A, Valery A. Patients hospitalized in general wards via the emergency department early identification of predisposing factors of death or unexpected intensive care unit admission---a historical prospective. Emerg Med Int 2014;2014:203747. 22. Bae MH, Lee JH, Yang DH, Park HS, Chae SC. Usefulness of surgical parameters as predictors of postoperative cardiac events in patients undergoing non-cardiac surgery. Circ J 2014;78:718-23. 23. Elm von E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vanderbroucke JP, the STROBE initiative. The strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007;370:1453-7. 24. Gupta H, Gupta PK, Fang X, Miller WJ, Cemaj S, Forse RA, et al. Development and validation of a risk calculator predicting postoperative respiratory failure. Chest 2011;140: 1207-15. 25. Gupta PK, Gupta H, Sundaram A, Kaushik M, Fang X, Miller WJ, et al. Development and validation of a risk calculator for prediction of cardiac risk after surgery. Circulation 2011;124:381-7. 26. Kheterpal S, O’Reilly M, Englesbe MJ, Rosenberg AL, Shanks AM, Zhang L, et al. Preoperative and intraoperative predictors of cardiac adverse events after general, vascular, and urological surgery. Anesthesiology 2009;110:58-66. 27. O’Brien MM, Gonzales R, Shroyer AL, Grunwald GK, Daley J, Henderson WG, et al. Modest serum creatinine elevation affects adverse outcome after general surgery. Kidney Int 2002;62:585-92. 28. Veltkamp SC, Kemmeren JM, van der Graaf Y, Edlinger M, van der Werken C. Prediction of serious complications in patients admitted to a surgical ward. Br J Surg 2002;89: 94-102.
ARTICLE IN PRESS 8 Visser et al
29. Crawford RS, Cambria RP, Abularrage CJ, Conrad MF, Lancaster RT, Watkins MT, et al. Preoperative functional status predicts perioperative outcomes after infrainguinal bypass surgery. J Vasc Surg 2010;51:351-8. 30. Davenport DL, Ferraris VA, Hosokawa P, Henderson WG, Khuri SF, Mentzer RM Jr. Multivariable predictors of postoperative cardiac adverse events after general and vascular surgery: results from the patient safety in surgery study. J Am Coll Surg 2007;204:1199-210. 31. Dhungel B, Diggs BS, Hunter JG, Sheppard BC, Vetto JT, Dolan JP. Patient and peri-operative predictors of morbidity and mortality after esophagectomy: American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), 2005-2008. J Gastrointest Surg 2010;14: 1492-501. 32. Giles KA, Wyers MC, Pomposelli FB, Hamdan AD, Ching YA, Schermerhorn ML. The impact of body mass index on perioperative outcomes of open and endovascular abdominal aortic aneurysm repair from the National Surgical Quality Improvement Program, 2005-2007. J Vasc Surg 2010;52:1471-7. 33. Greenblatt DY, Kelly KJ, Rajamanickam V, Wan Y, Hanson T, Rettammel R, et al. Preoperative factors predict perioperative morbidity and mortality after pancreaticoduodenectomy. Ann Surg Oncol 2011;18:2126-35. 34. Kneuertz PJ, Pitt HA, Bilimoria KY, Smiley JP, Cohen ME, Ko CY, et al. Risk of morbidity and mortality following hepato-pancreato-biliary surgery. J Gastrointest Surg 2012; 16:1727-35. 35. Nafiu OO, Kheterpal S, Moulding R, Picton P, Tremper KK, Campbell DA Jr, et al. The association of body mass index to postoperative outcomes in elderly vascular surgery patients: a reverse J-curve phenomenon. Anesth Analg 2011;112:23-9. 36. Nelson MT, Greenblatt DY, Soma G, Rajimanickam V, Greenberg CC, Kent KC. Preoperative factors predict mortality after major lower-extremity amputation. Surgery 2012;152:685-94. 37. Neumayer L, Hosokawa P, Itani K, El-Tamer M, Henderson WG, Khuri SF. Multivariable predictors of postoperative surgical site infection after general and vascular surgery: results from the patient safety in surgery study. J Am Coll Surg 2007;204:1178-87. 38. Johnson ON III, Sidawy AN, Scanlon JM, Walcott R, Arora S, Macsata RA, et al. Impact of obesity on outcomes after open surgical and endovascular abdominal aortic aneurysm repair. J Am Coll Surg 2010;210:166-77. 39. Mathur AK, Ghaferi AA, Osborne NH, Pawlik TM, Campbell DA, Englesbe MJ, et al. Body mass index and adverse perioperative outcomes following hepatic resection. J Gastrointest Surg 2010;14:1285-91. 40. Farhat J, Velanovich V, Falvo J, Horst MH, Swartz A, Patton JH, et al. Are the frail destined to fail? Frailty index as predictor of surgical morbidity and mortality in the elderly. J Trauma Acute Care Surg 2012;72:1526-31. 41. Glance LG, Lustik SJ, Hannan EL, Osler TM, Mukamel DB, Qian F, et al. The Surgical Mortality Probability Model: derivation and validation of a simple risk prediction rule for noncardiac surgery. Ann Surg 2012;255:696-702. 42. Kazaure H, Roman S, Sosa JA. High mortality in surgical patients with do-not-resuscitate orders: analysis of 8256 patients. Arch Surg 2011;146:922-8. 43. Breitenstein S, DeOliveira ML, Raptis DA, Slankamenac K, Kambakamba P, Nerl J, et al. Novel and simple preoperative
Surgery j 2015
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55. 56. 57.
58. 59.
60.
score predicting complications after liver resection in noncirrhotic patients. Ann Surg 2010;252:726-34. Obeid NM, Azuh O, Reddy S, Webb S, Reickert C, Velanovich V, et al. Predictors of critical care-related complications in colectomy patients using the National Surgical Quality Improvement Program: exploring frailty and aggressive laparoscopic approaches. J Trauma Acute Care Surg 2012; 72:878-83. Ata A, Lee J, Bestle SL, Desemone J, Stain SC. Postoperative hyperglycemia and surgical site infection in general surgery patients. Arch Surg 2010;145:858-64. Halfon P, Egglo Y, Matter M, Kallay C, Melle van G, Burnand B. Risk-adjusted rates for potentially avoidable reoperations were computed from routine hospital data. J Clin Epidemiol 2007;60:56-67. Leung AA, McAlister FA, Rogers SO Jr, Pazo V, Wright A, Bates DW. Preoperative hyponatremia and perioperative complications. Arch Intern Med 2012;172:1474-81. Nath B, Li Y, Carroll JE, Szabo G, Tseng JF, Shah SA. Alcohol exposure as a risk factor for adverse outcomes in elective surgery. J Gastrointest Surg 2010;14:1732-41. Saxton A, Velanovich V. Preoperative Frailty and Quality of Life as Predictors of Postoperative Complications. Ann Surg 2011;253:1223-9. Leichtle SW, Mouawad NJ, Lampman R, Singal B, Cleary RK. Does Preoperative Anemia Adversely Affect Colon and Rectal Surgery Outcomes? J Am Coll Surg 2011;212: 187-94. Regenbogen SE, Bordeianou L, Hutter M, Gawande AA. The intraoperative Surgical Apgar Score predicts postdischarge complications after colon and rectal resection. Surgery 2010;148:559-66. Scarborough JE, Pappas TN, Cox MW, Bennett KM, Shortell CK. Surgical trainee participation during infrainguinal bypass grafting procedures is associated with increased early postoperative graft failure. J Vasc Surg 2012;55:715-20. Trencheva K, Morrissey KP, Wells M, Mancuso CA, Lee SW, Sonoda T, et al. Identifying important predictors for anastomotic leak after colon and rectal resection: prospective study on 616 patients. Ann Surg 2013;257:108-13. Oostveen van CJ, Ubbink DT, Huis In Het Veld JG, Bakker PJ, Vermeulen H. Factors and models associated with the amount of hospital care services as demanded by hospitalized patients: a systematic review. PLoS One 2014;9:e98102. Studer P, Inderbitzin D. Surgery-related risk factors. Curr Opin Crit Care 2009;15:328-32. Saklad M. Grading of patients for surgical procedures. Anesthesiology 1941;2:281-4. Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. APACHE– Acute Physiology and Chronic Health Evaluation: a physiologically based classification system. Crit Care Med 1981;9:591-7. Copeland GH, Jones D, Walters M. POSSUM: a scoring system for surgical audit. Br J Surg 1991;78:355-60. Shah N, Hamilton M. Clinical review: Can we predict which patients are at risk of complications following surgery? Crit Care 2013;17:226. Vollmer CM, Lewis RS, Hall BL, Allendorf JD, Beane JD, Behrman SW, et al. Establishing a Quantitative Benchmark for Morbidity in Pancreatoduodenectomy using ACSNSQIP, the Accordion Severity Grading System, and Postopererative Morbidity Index. Ann Surg 2015;261:527-36.
