Abstract

Objective

The purpose of this study was to determine risk factors of adverse events in five surgical procedures.

Design

Retrospective record review was used to determine adverse events and risk factors of 1177 surgical admissions. Procedures included in this study were transurethral resection of prostate, hysterectomy, hip and knee arthroplasty, cholecystectomy and herniorrhaphy. Risk factors included comorbidity, lifestyle factors and medications. Stepwise multiple logistic regression was used to determine predictors of adverse events.

Setting

Two teaching hospitals in regional New South Wales, Australia.

Participants

1177 surgical admissions for five high volume procedures.

Main outcome measures

Identified predictors of adverse events in surgical admissions.

Results

The adverse event rate was 23.1% for all procedures (range 17.5–33.7% for the five procedures). Two factors were strongly predictive of an adverse event in all surgical admissions: age >70 years [odds ratio (OR) 1.9, 95% confidence intervals (CI) 1.3–2.6] and duration of operation (P = 0.005). Other predictive factors were: contaminated surgical site (OR 2.1, 95% CI 1.2–3.7) and anaemia (OR 1.8, 95% CI 1.1–2.8). Predictive factors of individual procedures included: urine retention (transurethral resection of the prostate); extended duration of operation and asthma (hysterectomy); acute admissions and extended duration of operation (cholecystectomy); and warfarin type drugs, ethanol abuse, failed prostheses, GI ulcer/inflammation, rheumatoid arthritis, and ischaemic heart disease (hip and knee joint arthroplasty).

Conclusions

The results of this study suggest that five factors should be routinely monitored for patients undergoing these procedures: age >70 years, type of procedure, duration of operation >2 h, contaminated surgical site and anaemia.

Introduction

The purpose of this study was to identify the risk factors associated with surgical adverse events in five high volume surgical procedures. It was part of the Newcastle Surgical Outcomes Study of Adverse Events and was conducted in two teaching hospitals in New South Wales, Australia. An adverse event is defined as ‘an unintended injury or complication which results in disability, death or prolongation of hospital stay, and is caused by health care management rather than the patient's disease’ [1].

One criticism of previous studies of adverse events is that there has been insufficient consideration of patient factors (comorbidity) that may increase the risk of an adverse event. Predictors of morbidity and mortality have been reported in the literature, but not for adverse events, which by definition require that the health care provided caused the poor outcome. This study obtained comorbidity data and other potential risk factors to determine whether they may be predictive of surgical adverse events. Ethics approval was obtained for the conduct of this study from the University of Newcastle Human Research Ethics Committee and the Hunter Area Research Ethics Committee.

Methods

A literature review was conducted to determine potential risk factors for an adverse event and to compile a list of risk factors including comorbidities, lifestyle factors, relevant medications, and other relevant factors. Risk scoring systems were also reviewed. Jones and de Cossart [2] found that there are several systems for assessing a patient's risk of complications in general surgery including: the American Society of Anesthesiologists (ASA) grading, Goldman Cardiac Risk Index, and combinations of these two including the Charlson Comorbidity Index, Prognostic Nutritional Index, Hospital Prognostic Index, Acute Physiology and Chronic Health Evaluation System and its modifications, the Physiological and Operative Severity Score for enUmeration of Mortality and morbidity (POSSUM) and its modifications, Mortality Prediction Model, various multiple organ failure prediction systems, sepsis scores and trauma scores. Jones and de Cossart [2] reviewed 18 of these systems by comparing their components and outcomes and found that the prediction performance of ‘POSSUM is the most appropriate of the currently available scores for general surgical practice’. ASA grading (4–5) has been reported by other authors as predictive of adverse outcomes [3], mortality [4, 5] and morbidity [5, 6].

The POSSUM method contains 12 physiological parameters and six operative parameters [7]. The physiological factors are age, cardiac history, respiratory history, blood pressure, pulse rate, Glasgow coma score, haemoglobin level, white cell count, urea concentration, sodium level, potassium level, electrocardiograph; and the operative factors are operative severity, multiple procedures, total blood loss, peritoneal soiling, presence of malignancy, and mode of surgery. Many of these factors have been identified in the literature as predictors of morbidity or mortality and in two studies of adverse events [3, 8]: age >70 years [3–6, 9]; congestive cardiac failure [3, 4, 8], recent or history of myocardial infarction [3, 9] and arrhythmia [9]; chronic obstructive pulmonary disease or lung disease [5, 8, 9]; operation category (major or minor) [10]; and cancer, tumours or malignancy [4, 5, 8]. Other factors found are: wound category [6, 11]; duration of operation [10]; renal disease [9]; low serum albumin [4–6]; diabetes mellitus [9]; low haematocrit [5, 6]; cerebrovascular disease [9]; liver disease [8]; emergency surgery [3–6, 9] and obesity [9].

The results of these 12 studies may not be generalizable to all surgical admissions and the methodologies are not comparable. Many of these studies used risk adjusted models or severity of illness adjustment to determine predictors of morbidity or mortality [3–6, 10]. This study contributes to this work by determining comorbidities and related factors that may be predictors of surgical adverse events.

The five procedures used in this study were transurethral resection of prostate, hysterectomy, hip and knee arthroplasty, cholecystectomy and herniorrhaphy. These procedures were selected because they are normally elective high volume surgical procedures. They have been previously reported to have high adverse event rates and these adverse events were reported to be highly preventable [12]. A structured medical record review form was used to collect all comorbidity and drugs of potential clinical significance from medical records. Drugs of potential clinical significance were non-steroidal anti-inflammatories, aspirin, hormone replacement therapy, antibiotics, oral contraceptive pill, steroids, anticoagulants, insulin, immuno-suppressive/immuno-modifier agents, Tamoxifen and other hormonal antineoplastic agents, methotrexate, opioid/narcotic analgesics, tranexamic acid, monoamine oxidase inhibitors, anatabuse, antimetabolites, dilantin, desmopressin, methadone and stimulants.

Retrospective medical record review was conducted in two stages. The first stage involved a trained Registered Nurse screening the medical record using the screening criteria used in the Quality in Australian Health Care Study [1]. If one or more criteria were positive, the record was reviewed by a surgical reviewer from the relevant specialty in the hospital who had not performed the procedure for that admission. They were provided with training sessions and a manual to assist them to review medical records using a structured questionnaire based on the Quality in Australian Health Care Study. The structured review process included a series of questions designed to assist reviewers to determine the outcomes of admissions and whether they were caused by health care management, thus meeting the definition of an adverse event.

Retrospective record review determined adverse events and their risk factors of 1177 surgical admissions. Risk factors that had <10 cases were not used. The remaining risk factors were analysed using bivariate (χ2) analyses and factors with a P-value <0.25 were entered into a stepwise multiple logistic regression model. This was done for all admissions and each procedural group separately.

Results

There were 1177 admissions (619 females, 558 males; age ranged from 17 to 93 years, mean 61 years, median 65 years). There were 272 adverse events and the rate was 23.1% for all procedures (Table 1); and there was a statistically significant association between types of procedures and adverse event rates, P < 0.001. The types of adverse events included 54 patients who had bleeding problems, 11 patients had pulmonary emboli, five had deep vein thromboses and six patients died (procedures were two transurethral resection of prostate, two joint replacements, one cholecystectomy, one hernia) (causes were two pulmonary emboli, two sepsis, one acute myocardial infarction, one haemorrhage). Eighty-four patients had some form of postoperative infection – including wound infection, a deep, cavity or joint infection, pneumonia, sepsis and urinary tract infections. The admissions that had one or more adverse events (272) were associated with 89 (33%) unplanned readmissions requiring 709 additional days in hospital. Fifty-five (20%) patients had additional surgery (seven returned to theatre during the admission for the procedure).

Table 1

Admissions for procedural groups and the proportion associated with an adverse event.

Procedure Total admissions Adverse events % (95% CI) 
Transurethral resection of prostate 208 41 19.7 (14.3–25.1) 
Hysterectomy 243 82 33.7 (27.8–39.7) 
Total joint replacement 420 82 19.5 (15.7–23.3) 
Cholecystectomy 212 37 17.5 (12.3–22.6) 
Herniorrhaphy 94 30 31.9 (22.5–41.3) 
Total 1177 272 23.1 (20.7–25.5) 
Procedure Total admissions Adverse events % (95% CI) 
Transurethral resection of prostate 208 41 19.7 (14.3–25.1) 
Hysterectomy 243 82 33.7 (27.8–39.7) 
Total joint replacement 420 82 19.5 (15.7–23.3) 
Cholecystectomy 212 37 17.5 (12.3–22.6) 
Herniorrhaphy 94 30 31.9 (22.5–41.3) 
Total 1177 272 23.1 (20.7–25.5) 

There were 243 comorbidities and clinically significant drugs but 166 risk factors had <10 cases and were not used. The remaining 77 risk factors were analysed individually as potential predictors of an adverse event. There were 37 factors that were not statistically significant (P-values >0.25): antibiotics, fibroids/adenomyosis/polyps, uterine prolapse, gall bladder empyema/porcelain/mucocele/gangrenous, pancreatitis, arrhythmia, rheumatoid arthritis, aspirin, congestive cardiac failure/left ventricular failure, chronic airways limitation, oral contraceptive pill, diabetes mellitus, disc degeneration, diverticulosis, emphysema, ex-smoker, haematuria, cardiac murmur, hypertension, hypothyroidism, immuno-suppressed, elevated liver function tests, non-steroidal anti-inflammatory drugs, oedema, previous anaesthetic problems, previous deep vein thrombosis, previous postoperative bleeding, previous pneumonia, previous postoperative infection, previous postoperative vomiting, previous pulmonary or other embolus, smoker, short of breath, tamoxifen, thrombocytopenia, recent upper respiratory tract infection and urinary tract infection. The 40 factors that had P-values <0.25, using bivariate analysis, are listed in Table 2.

Table 2

Bivariate analysis of factors as predictors of an adverse event (1177 admissions)

Risk factora Patients P-value (bivariate analysis) OR (95% CI) 
Contaminated surgical site 337 29 <0.001 2.0 (1.5–2.7) 
Cancer 137 12 0.001 2.0 (1.3–2.9) 
Anaemia 102 0.001 2.1 (1.4–3.3) 
Age >70 years 399 34 0.001 1.6 (1.2–2.2) 
Osteoporosis 52 0.003 2.4 (1.3–4.2) 
Angina 47 0.004 2.4 (1.3–4.3) 
Frail 23 0.005 3.1 (1.4–7.2) 
Warfarin/Coumadin/Ticlid 30 0.008 2.6 (1.3–5.5) 
Hypo-albuminaemia 13 0.008 4.0 (1.3–11.9) 
Chronic renal failure 38 0.015 2.2 (1.2–4.4) 
Hepatitis B/C positive 14 0.016 3.4 (1.2–9.7) 
Redo procedures 83 0.017 1.8 (1.1–2.9) 
Ischaemic heart disease 180 15 0.017 1.5 (1.1–2.2) 
Urine retention 29 0.018 2.4 (1.1–5.1) 
Previous stroke 59 0.020 1.9 (1.1–3.3) 
Ethanol abuse 36 0.023 2.2 (1.1–4.3) 
Acute admissions 74 0.024 1.8 (1.1–2.9) 
Preoperative bladder catheterization 44 0.034 2.0 (1.0–3.7) 
Peripheral vascular disease 35 0.046 2.0 (1.0–4.1) 
Asthma 124 11 0.060 1.5 (1.0–2.2) 
Elevated creatinine and urea 26 0.060 2.1 (1.0–4.7) 
Thrombocytosis 14 0.078 2.5 (0.9–7.4) 
Urinary incontinence 71 0.104 1.5 (0.9–2.6) 
Gastrointestinal ulcer/inflammation 53 0.113 1.6 (0.9–2.9) 
Osteoarthritis 504 43 0.143 0.8 (0.6–1.1) 
Insulin 26 0.159 1.8 (0.8–4.1) 
Hypokalaemia 23 0.180 1.8 (0.8–4.3) 
Obesity 577 49 0.196 0.8 (0.6–1.1) 
Cardiac valve replacement 10 0.203 2.2 (0.6–8.0) 
Febrile 10 0.203 2.2 (0.6–8.0) 
Pacemaker 10 0.203 2.2 (0.6–8.0) 
Steroids 39 0.249 1.5 (0.7–3.0) 
Hormone replacement therapy 101 0.250 1.3 (0.8–2.1) 
Risk factora Patients P-value (bivariate analysis) OR (95% CI) 
Contaminated surgical site 337 29 <0.001 2.0 (1.5–2.7) 
Cancer 137 12 0.001 2.0 (1.3–2.9) 
Anaemia 102 0.001 2.1 (1.4–3.3) 
Age >70 years 399 34 0.001 1.6 (1.2–2.2) 
Osteoporosis 52 0.003 2.4 (1.3–4.2) 
Angina 47 0.004 2.4 (1.3–4.3) 
Frail 23 0.005 3.1 (1.4–7.2) 
Warfarin/Coumadin/Ticlid 30 0.008 2.6 (1.3–5.5) 
Hypo-albuminaemia 13 0.008 4.0 (1.3–11.9) 
Chronic renal failure 38 0.015 2.2 (1.2–4.4) 
Hepatitis B/C positive 14 0.016 3.4 (1.2–9.7) 
Redo procedures 83 0.017 1.8 (1.1–2.9) 
Ischaemic heart disease 180 15 0.017 1.5 (1.1–2.2) 
Urine retention 29 0.018 2.4 (1.1–5.1) 
Previous stroke 59 0.020 1.9 (1.1–3.3) 
Ethanol abuse 36 0.023 2.2 (1.1–4.3) 
Acute admissions 74 0.024 1.8 (1.1–2.9) 
Preoperative bladder catheterization 44 0.034 2.0 (1.0–3.7) 
Peripheral vascular disease 35 0.046 2.0 (1.0–4.1) 
Asthma 124 11 0.060 1.5 (1.0–2.2) 
Elevated creatinine and urea 26 0.060 2.1 (1.0–4.7) 
Thrombocytosis 14 0.078 2.5 (0.9–7.4) 
Urinary incontinence 71 0.104 1.5 (0.9–2.6) 
Gastrointestinal ulcer/inflammation 53 0.113 1.6 (0.9–2.9) 
Osteoarthritis 504 43 0.143 0.8 (0.6–1.1) 
Insulin 26 0.159 1.8 (0.8–4.1) 
Hypokalaemia 23 0.180 1.8 (0.8–4.3) 
Obesity 577 49 0.196 0.8 (0.6–1.1) 
Cardiac valve replacement 10 0.203 2.2 (0.6–8.0) 
Febrile 10 0.203 2.2 (0.6–8.0) 
Pacemaker 10 0.203 2.2 (0.6–8.0) 
Steroids 39 0.249 1.5 (0.7–3.0) 
Hormone replacement therapy 101 0.250 1.3 (0.8–2.1) 

aSeven risk factors that were procedure-specific, are not shown in this table because they were indications of the procedure: failed prosthesis/internal fixation (P = 0.002), gallstones/polyp/adenomyomatosis (P = 0.006), endometriosis/pelvic pain (P = 0.008), pelvic mass (P = 0.012), prostatism (P = 0.073), dysfunctional uterine bleeding/dysmenorrhoea (P = 0.158) and gall bladder inflammation/biliary colic (P = 0.250).

The logistic regression model, found 10 factors that were statistically significant (Table 3). The c-index for the logistic regression was 0.71 and the Hosmer–Lemeshow statistic had a P-value of 0.79, which shows the model is a good fit to the data. Procedural group and duration of operation were also significant factors. The adverse event rate increased with increased duration of operation (Table 4). There were two factors that were strongly predictive of an adverse event: age >70 years, odds ratio (OR) = 1.9 and duration of operation, P = 0.005. Risk factors found to be significant in this study that were not identified in previous studies were: hepatitis B/C positive, ethanol abuse, warfarin type drugs, angina, osteoporosis, asthma, and failed previous prosthesis or internal fixation. Results of analysis of risk factors for each procedure are described later. Risk factors identified as significant for each procedure are summarized in Table 5.

Table 3

Results of stepwise multiple logistic regression model to determine statistically significant risk factors of an adverse event (1173 admissionsa)

Factor Patients Logistic regression modelb
 
  OR (95% CI) P-value 
Hepatitis B/C positive 14 3.7 (1.1–12.0) 0.027 
Warfarin/Coumadin/Ticlid 30 2.8 (1.2–6.0) 0.011 
Ethanol abuse 36 2.6 (1.2–5.4) 0.015 
Failed prosthesis/internal fixation 42 2.3 (1.1–4.8) 0.025 
Contaminated surgical site 337 2.1 (1.2–3.7) 0.013 
Osteoporosis 52 2.0 (1.0–3.7) 0.033 
Angina 47 2.0 (1.0–3.8) 0.037 
Age >70 years 399 1.9 (1.3–2.6) 0.001 
Anaemia 102 1.8 (1.1–2.8) 0.017 
Asthma 124 1.6 (1.0–2.5) 0.030 
Factor Patients Logistic regression modelb
 
  OR (95% CI) P-value 
Hepatitis B/C positive 14 3.7 (1.1–12.0) 0.027 
Warfarin/Coumadin/Ticlid 30 2.8 (1.2–6.0) 0.011 
Ethanol abuse 36 2.6 (1.2–5.4) 0.015 
Failed prosthesis/internal fixation 42 2.3 (1.1–4.8) 0.025 
Contaminated surgical site 337 2.1 (1.2–3.7) 0.013 
Osteoporosis 52 2.0 (1.0–3.7) 0.033 
Angina 47 2.0 (1.0–3.8) 0.037 
Age >70 years 399 1.9 (1.3–2.6) 0.001 
Anaemia 102 1.8 (1.1–2.8) 0.017 
Asthma 124 1.6 (1.0–2.5) 0.030 

aThere were four admissions that did not have contaminated surgical site or operating time recorded in the medical record.

bProcedure category was included in the model and was statistically significant, P = 0.002. Duration of operation was included in the model using four categories (<60, 60–119, 120–179, >180 min) and was statistically significant, P = 0.005.

Table 4

Duration of operation, adverse event rates and ORs for an adverse event

Duration of operationa (min) Adverse events (admissions) Adverse event rate (%) OR – bivariate (95% CI) Adjusted ORb (95% CI) 
<60 68 (345) 19.7 0.8 (0.6–1.0) 1.0 
60–119 136 (647) 21.0 0.8 (0.6–1.0) 1.2 (1.0–1.6). 
120–179 45 (144) 31.25 1.6 (1.1–2.4) 2.2 (1.6–3.1) 
>180 22 (39) 56.4 4.6 (2.4–8.8) 5.5 (3.3–9.2) 
Duration of operationa (min) Adverse events (admissions) Adverse event rate (%) OR – bivariate (95% CI) Adjusted ORb (95% CI) 
<60 68 (345) 19.7 0.8 (0.6–1.0) 1.0 
60–119 136 (647) 21.0 0.8 (0.6–1.0) 1.2 (1.0–1.6). 
120–179 45 (144) 31.25 1.6 (1.1–2.4) 2.2 (1.6–3.1) 
>180 22 (39) 56.4 4.6 (2.4–8.8) 5.5 (3.3–9.2) 

aThere were two admissions that did not have operating time recorded in the medical record.

bThe adjusted ORs are after including procedure in the model (P ≤ 0.001).

Table 5

Risk factors by procedure and ORs of an adverse event

Procedure Risk factors OR (95% CI) 
Transurethral resection of prostate Urine retention 3.5 (1.4–8.6) 
Hysterectomy Asthma 2.9 (1.3–6.7) 
 Duration of operation (min)  
  60–119 0.3 (0.1–0.7) 
  120–179 2.0 (0.6–5.0) 
   >180 3.3 (0.9–16.7) 
Hip and knee joint arthroplasty Warfarin type drugs 5.1 (1.4–19.5) 
 Ethanol abuse 4.6 (1.4–14.3) 
 Failed prosthesis/internal fixation 3.7 (1.6–8.2) 
 Gastrointestinal ulcer/inflammation 3.4 (1.3–8.3) 
 Rheumatoid arthritis 2.9 (1.0–7.5) 
 Age >70 years 1.9 (1.1–3.3) 
 Ischaemic heart disease 1.8 (1.0–3.3) 
Cholecystectomy Age >70 years 3.4 (1.4–8.1) 
 Acute admission 2.8 (1.2–6.4) 
 Duration of operation (min)  
  60–119 1.0 (0.3–3.3) 
  120–179 2.5 (0.4–10.0) 
   >180 10.0 (1.0–125.0) 
Herniorrhaphy Age >70 years 3.1 (1.2–8.0) 
Procedure Risk factors OR (95% CI) 
Transurethral resection of prostate Urine retention 3.5 (1.4–8.6) 
Hysterectomy Asthma 2.9 (1.3–6.7) 
 Duration of operation (min)  
  60–119 0.3 (0.1–0.7) 
  120–179 2.0 (0.6–5.0) 
   >180 3.3 (0.9–16.7) 
Hip and knee joint arthroplasty Warfarin type drugs 5.1 (1.4–19.5) 
 Ethanol abuse 4.6 (1.4–14.3) 
 Failed prosthesis/internal fixation 3.7 (1.6–8.2) 
 Gastrointestinal ulcer/inflammation 3.4 (1.3–8.3) 
 Rheumatoid arthritis 2.9 (1.0–7.5) 
 Age >70 years 1.9 (1.1–3.3) 
 Ischaemic heart disease 1.8 (1.0–3.3) 
Cholecystectomy Age >70 years 3.4 (1.4–8.1) 
 Acute admission 2.8 (1.2–6.4) 
 Duration of operation (min)  
  60–119 1.0 (0.3–3.3) 
  120–179 2.5 (0.4–10.0) 
   >180 10.0 (1.0–125.0) 
Herniorrhaphy Age >70 years 3.1 (1.2–8.0) 

Transurethral resection of the prostate

There were 208 transurethral resections of the prostate admissions with 41 adverse events. The principal admission diagnoses included: prostatism/benign prostatic hypertrophy/hyperplasia/prostatomegaly/with or without bladder neck obstruction/contracture (161), prostate cancer (30), bladder cancer (6), urine retention (without diagnosis of prostatism) (8), and other (3). There were 13 factors that had P-values <0.25 using bivariate analysis. In the logistic regression model, urine retention was the only statistically significant risk factor for an adverse event: OR 3.5 [95% confidence intervals (CI) 1.4–8.6]. Duration of operation was also included in the model to control for duration of operation and was not a statistically significant factor.

Hysterectomy

There were 243 hysterectomy admissions with 82 adverse events. The principal admission diagnoses included: dysfunctional uterine bleeding/menorrhagia/metrorrhagia (72), uterine cancer (47), uterine/vaginal prolapse (38), uterine fibroids/benign tumour/other cyst (28), pelvic/abdominal mass (22), cervical cancer (10), ovarian cancer (9), endometriosis (9), dysmenorrhoea (5) and others (3). There were 10 factors with a P-value <0.25. There were two factors that were statistically significant in the logistic regression model: asthma, OR 2.9 (95% CI 1.3–6.7) and duration of operation. ORs for duration of operation 60–119, 120–179 and >180 min were: 0.3, 2.0 and 3.3, respectively. Some hysterectomy admissions required extensive surgery including pelvic lymphadenectomy, which increased operating time.

Hip and knee joint arthroplasty

There were 420 joint admissions with 82 adverse events. The admission diagnoses included: osteoarthritis (354), failed or loose prostheses (36), rheumatoid arthritis (8), avascular necrosis (7), fractured neck of femur or girdle stones (4), infected knee prostheses (3), painful joint (3), and other (5). There were 22 factors that had a P-value <0.25.

In the logistic regression model, there were seven factors that were statistically significant. Warfarin type drugs: OR 5.1 (95% CI 1.4–19.5), ethanol abuse: OR 4.6 (95% CI 1.4–14.3), failed prosthesis/internal fixation: OR 3.7 (95% CI 1.6–8.2), gastrointestinal ulcer/inflammation: OR 3.4 (95% CI 1.3–8.3), rheumatoid arthritis: OR 2.9 (95% CI 1.0–7.5), age >70 years: OR 1.9 (95% CI 1.1–3.3) and ischaemic heart disease: OR 1.8 (95% CI 1.0–3.3). Duration of operation was also included in the model and was not a statistically significant factor. Failed prostheses are associated with redo procedures (38 of 42 redo procedures) and the association with an adverse event may reflect increased complexity of these procedures.

Cholecystectomy

There were 212 cholecystectomy admissions with 37 adverse events. The principal admission diagnoses included: cholelithiasis or choledocholithiasis (101), cholecystitis (with or without gallstones) (51), pancreatitis or biliary tree obstruction (19), biliary colic or pain (16), gall bladder empyema/porcelain gallbladder/gangrenous cholecystitis/abscess/necrotizing pancreatitis (12), mucocele of gall bladder (5), cancer (4) and others (4). Nine factors had P-values <0.25. The surgical site was classified as contaminated where spillage of contents of gall bladder occurred during its removal, and fluid was not considered sterile, e.g. empyema of gall bladder.

In the logistic regression model, there were three statistically significant factors. Age >70 years: OR 3.4 (95% CI 1.4–8.1), acute admission: OR 2.8 (95% CI 1.2–6.4) and operation time. ORs of duration of operation of 60–119, 120–179 and >180 min were: 1.0, 2.5, 10.0, respectively. Extended operating time is due to more complex procedures.

Herniorrhaphy

There were 94 herniorrhaphy admissions with 30 adverse events. The admission diagnoses included: inguinal/femoral hernia (81), strangulated/incarcerated hernia (8), recurrent hernia (4), and cancer (1). There were two factors with P-values <0.25. In the logistic regression model, only age >70 years was statistically significant: OR 3.1 (95% CI 1.2–8.0). Duration of operation was included in the model and was not statistically significant.

Discussion

Although there have been many publications on how to determine the impact of comorbidity on patient outcomes, this is the first study to identify those comorbidities that increase the risk of an adverse event. In previous studies there have been differences in the patient profiles (veterans surgical admissions, acute hospital admissions, surgical admissions, longitudinal study groups and medicare populations) and in the outcomes studied (mortality, morbidity, length of stay, readmission, postoperative morbidity/mortality, pulmonary complications, adverse outcomes and adverse events as measured by Davis et al. [8] and Rosen et al. [3]). Further, there were differences in the comorbidities selected, and while some were combined into more general categories, no set exceeded 67 comorbidities. For these reasons it is very difficult to make comparisons between these studies.

In this study the patient profile is very specific – surgical admissions of five common substantially elective procedures. The outcome studied was the occurrence of an adverse event, a poor outcome associated with the care provided. There were 243 categories of comorbid conditions or potentially significant drugs obtained from the medical record. The results of this study are therefore not comparable with previous work and have identified comorbidities not previously reported. These factors should be used in developing a set of comorbidities that predict the occurrence of an adverse event.

There were 12 risk factors identified for all procedural groups combined. These factors may be generalizable to other elective surgery admissions: hepatitis B/C positive, ethanol abuse, prior use of warfarin type drugs, failed prosthesis/internal fixation, contaminated surgical site, osteoporosis, angina, age >70 years, anaemia, asthma and duration of operation. Five of these factors have been identified by previous studies as risk factors: contaminated surgical site [6, 11], increasing age/age >70 years [3–7, 9], anaemia [7], operation category (type of procedure) [10] and duration of operation [10]. Extended duration of operation included technically difficult and complex procedures such as redo/conversion procedures and pelvic lymphadenectomy. Seven factors were not found in previous studies and may reflect the unique profile of patients in this study: hepatitis B/C positive, warfarin type drugs, ethanol abuse, failed prosthesis, osteoporosis, angina and asthma.

There were also risk factors identified for specific procedural groups. For transurethral resection of the prostate admissions: pre-existing urine retention; for hysterectomy: asthma and duration of operation; for joint procedures: warfarin type drugs, age >70 years, ethanol abuse, failed prosthesis, gastrointestinal ulcer/inflammation, rheumatoid arthritis, and ischaemic heart disease; for cholecystectomy: age >70 years, acute admissions and duration of operation; and for herniorrhaphy: age >70 years. It is likely that gastrointestinal ulcer or inflammation is an indication of extended use of non-steroidal anti-inflammatory drugs in joint patients prior to a decision for surgery. The increased complexity involved in a revision procedure may contribute to the risk for an adverse event for patients with failed prostheses. It is not clear why asthma was found to be significant for hysterectomy admissions.

These results confirm that adverse event rates vary with the type of procedure and may be increased by the presence of comorbidities or risk factors, but these factors vary between procedural groups. The most frequently identified factors were age >70 years and duration of operation, however, they were not statistically significant for all procedural groups.

When these results are compared with previous published work on predictors of mortality and morbidity, they support the following previously identified comorbidities: age, contaminated surgical site, procedure, duration of operation and anaemia. The comorbidity profile of admissions in this study may not be comparable with previously published studies since other studies included more than five procedures [4–7, 10], Veterans Affairs studies had older, male patients [4–6] and other studies included medical patients [8].

The following factors: contaminated surgical site, age >70 years, anaemia, operation category (type of procedure) and duration of operation; could be used to routinely identify surgical patients ‘at risk’ of an adverse event at admission, and provide the basis for developing preventative management strategies. Operation time of >2 h has previously been identified as a risk factor for nosocomial wound infection by McLaws et al. [11], and the increase in the OR reported for duration of operation 120–179 min in Table 4 supports the use of this indicator as a risk factor. The ORs reported in this table have a pattern similar to those reported in a previous study where the ORs for an adverse event increased as operating time increased [10].

Published predictors in the literature that did not prove to be significant in this study were: acute admissions, renal disease/dialysis, chronic obstructive pulmonary disease/lung disease, congestive cardiac failure, low serum albumin, myocardial infarction, diabetes mellitus, elevated creatinine and urea, peripheral vascular disease, cerebrovascular accident, dementia, connective tissue disease, ulcer disease, liver disease, hemiplegia, cancer, cardiac history (excepting angina), respiratory history (excepting asthma), arrhythmia, obesity, hypertension and smoking. The database did not contain ASA scores, pain, and low haematocrit (also identified as predictors in the literature) as they were not routinely documented in the medical record.

A potential limitation of this study is the collection of data from medical records where the quality of the documentation may be variable, however, the data are considered to be more accurate than using administrative data for this purpose because they allowed researchers to distinguish pre-existing comorbidities from complications. Although ASA classification of physical status scores were not collected in this study, previous research has reported that grades 4–5 are a significant risk factor [2–6] and researchers may consider adding it to the factors identified in this study. It is possible that an intraoperative adverse event may extend operating time and this may potentially confound results. Also the analyses by procedural groups have smaller sample sizes and hence have reduced power to identify predictor variables.

The five recommended risk factors for predicting an adverse event in surgical admissions are: contaminated surgical site, duration of operation (operation time >2 h), age >70 years, anaemia and operation category (type of procedure).

Funding

This study was supported by a grant provided by the Ministerial Advisory Committee on Quality in Health Care, NSW Department of Health, Australia. This grant also provided a scholarship for A.K.K. and the study was part of a successful PhD candidature.

References

Wilson
RM
Runciman
WB
Gibberd
RW
, et al.  . 
The Quality in Australian Health Care Study
Med J Aust
 , 
1995
, vol. 
163
 (pg. 
458
-
71
)
Jones
H
de Cossart
L
Risk scoring in surgical patients
Br J Surg
 , 
1999
, vol. 
86
 (pg. 
149
-
57
)
Rosen
A
Ash
A
McNiff
K
, et al.  . 
The importance of severity of illness adjustment in predicting adverse outcomes in the medicare population
J Clin Epidemiol
 , 
1995
, vol. 
48
 (pg. 
631
-
43
)
Khuri
S
Daley
J
Henderson
W
, et al.  . 
Risk adjustment of the postoperative mortality rate for the comparative assessment of the quality of surgical care: results of the National Veterans Affairs Surgical Risk Study
J Am Coll Surg
 , 
1997
, vol. 
185
 (pg. 
315
-
27
)
Khuri
S
Daley
J
Henderson
W
, et al.  . 
The Department of Veterans Affairs' NSQIP: the first national, validated, outcome-based, risk-adjusted, and peer-controlled program for the measurement and enhancement of the quality of surgical care
Ann Surg
 , 
1998
, vol. 
228
 (pg. 
491
-
507
)
Daley
J
Khuri
S
Henderson
W
, et al.  . 
Risk adjustment of the postoperative morbidity rate for the comparative assessment of the quality of surgical care: results of the National Veterans Affairs Surgical Risk Study
J Am Coll Surg
 , 
1997
, vol. 
185
 (pg. 
328
-
40
)
Prytherch
D
Whiteley
M
Higgins
B
, et al.  . 
POSSUM and Portsmouth POSSUM for predicting mortality
Br J Surg
 , 
1998
, vol. 
85
 (pg. 
1217
-
20
)
Davis
P
Lay-Yee
R
Fitzjohn
J
, et al.  . 
Co-morbidity and health outcomes in three Auckland hospitals
N Z Med J
 , 
2002
, vol. 
115
 (pg. 
211
-
5
)
Tuula
S
Kurki
M
Kataja
M
Preoperative prediction of postoperative morbidity in coronary artery bypass grafting
Ann Thorac Surg
 , 
1996
, vol. 
61
 (pg. 
1740
-
5
)
Pillai
S
van Rij
A
Williams
S
, et al.  . 
Complexity- and risk-adjusted model for measuring surgical outcome
Br J Surg
 , 
1999
, vol. 
86
 (pg. 
1567
-
72
)
McLaws
M
Murphy
C
Keogh
G
The validity of surgical wound infection as a clinical indicator in Australia
Aust N Z J Surg.
 , 
1997
, vol. 
67
 (pg. 
675
-
8
)
Kable
A
Gibberd
R
Spigelman
A
Adverse events in surgical patients in Australia
Int J Qual Health Care
 , 
2002
, vol. 
14
 (pg. 
269
-
76
)