Evaluating alternative risk-adjustment strategies for surgery

BACKGROUND: Comparison of institutional health care outcomes requires risk adjustment. Risk-adjustment methodology may influence the results of such comparisons. METHODS: We compared 3 risk-adjustment methodologies used to assess the quality of surgical care. Nurse reviewers abstracted data from a continuous sample of 2,167 surgical patients at 3 academic institutions. One risk adjustor was based on medical record data (National Surgical Quality Improvement Program [NSQIP]) whereas the other 2, the DxCG and Charlson Comorbidity Index (CCI), primarily used International Classification of Disease-9 (ICD-9) codes. Risk-assessment scores from the 3 systems were compared with each other and with mortality. RESULTS: Substantial disagreement was found in the risk assessment calculated by the 3 methodologies. Although there was a weak association between the CCI and DxCG, neither correlated well with the NSQIP. The NSQIP was best able to predict mortality, followed by the DxCG and CCI. CONCLUSION: In surgical patients, different risk-adjustment methodologies afford divergent estimates of mortality risk.     

An Efficient Risk Adjustment Model to Predict Inpatient Adverse Events after Surgery

Background

Risk adjustment is an important component of surgical outcomes and quality analyses. Current models include numerous preoperative variables; however, the relative contribution of these variables may be limited. This research seeks to identify a model with the fewest number of variables necessary to perform an adequate risk adjustment to predict any inpatient adverse event for use in resource-limited settings.

Methods

All patients from the National Surgical Quality Improvement Program (NSQIP) database from 2005 to 2010 were included. Outcomes were inpatient mortality or any surgical complication captured by NSQIP. Models were built by sequential addition of preoperative risk variables selected by their area under the receiver operator characteristic curve (AUC).

Results

Among 863,349 patients, the single variable with the highest AUC was American Society of Anesthesiologists (ASA) classification (AUC = 0.7127). AUC values reached 0.7923 with five variables (ASA classification, wound classification, functional status prior to surgery, albumin, and age) and 0.7945 with six variables. The sixth variable was one of the following: alkaline phosphatase, weight loss, principal anesthesia technique, gender, or emergency status. The model with the highest discrimination that did not require laboratories included ASA classification, functional status prior to surgery, wound classification, and age (AUC = 0.7810). Including all 66 preoperative variables produced little additional gain (AUC = 0.8006).

Conclusions

Six variables are sufficient to develop a risk adjustment tool for inpatient surgical mortality and morbidity. This research has important implications for the field of surgical outcomes research by improving efficiency of data collection. This limited model can aid the expansion of risk-adjusted analyses to resource-limited settings worldwide.     

Preoperative risk factors and surgical complexity are more predictive of costs than postoperative complications: a case study using the National Surgical Quality Improvement Program (NSQIP) database

OBJECTIVE: This single-center study tested the hypothesis that preoperative risk factors and surgical complexity predict more variation in hospital costs than complications. BACKGROUND: Complications after surgical operations have been shown to significantly increase hospital cost. The impact on complication-related costs of preoperative risk factors is less well known. METHODS: The National Surgical Quality Improvement Program (NSQIP) preoperative risk factors, surgical complexity, and outcomes, along with hospital costs, were analyzed for a random sample of 5875 patients on 6 surgical services. Operation complexity was assessed by work RVUs (Centers for Medicare and Medicaid Services Resource Based Relative Value Scale). The difference in mean hospital costs associated with all variables was analyzed. Multiple linear regression was used to determine the cost variation associated with all variables separately and combined. RESULTS: Fifty-one of 60 preoperative risk factors, work RVUs, and 22 of 29 postoperative complications were associated with higher variable direct costs (P < 0.05). Linear regressions showed that risk factors predicted 33% (P < 0.001) of cost variation, work RVUs predicted 23% (P < 0.001), and complications predicted 20% (P < 0.001). Risk factors and work RVUs together predicted 49% of cost variation (P < 0.001) or 16% more than risk factors alone. Adding complications to this combined model modestly increased prediction of costs by 4% for a total of 53% (P < 0.001). CONCLUSION: Preoperative risk factors and surgical complexity are more effective predictors of hospital costs than complications. Preoperative intervention to reduce risk could lead to significant cost savings. Payers and regulatory agencies should risk-adjust hospital cost assessments using clinical information that integrates costs, preoperative risk, complexity of operation, and outcomes.     

An analysis of delayed breast reconstruction outcomes as recorded in the American College of Surgeons National Surgical Quality Improvement Program

The American College of Surgeons National Surgical Quality Improvement Program (NSQIP) is a prospective, multi-institutional clinical registry established as an auditing instrument to monitor and improve patient care in surgery. To date no publication has queried the NSQIP dataset from a plastic surgery perspective. A retrospective analysis was carried out of all delayed breast reconstruction cases recorded by the NSQIP from 2005 to 2008 (n=645). The 30-day morbidity was 5.7%, with wound infections being the most common complication. Logistic regression analysis identified BMI >25 and preoperative radiotherapy as independent risk factors for overall morbidity and wound infection. The NSQIP does not allow for capture of procedure-specific outcomes and this results in underreporting of overall morbidity compared to the literature; this may limit the capability to assess risk factors for complications. As breast reconstruction modules for NSQIP are currently under development, modifications to capture procedure-specific outcomes are recommended.     

How well do pediatric anesthesiologists agree when assigning ASA physical status classifications to their patients?1

BACKGROUND: The scope and application of the American Society of Anesthesiologists Physical Status (ASA PS) classification has been called into question and interobserver consistency even by specialist anesthesiologists has been described as only fair. Our purpose was to evaluate the consistency of the application of the ASA PS amongst a group of pediatric anesthesiologists. METHODS: We randomly selected 400 names from the active list of specialist members of the Society for Pediatric Anesthesia. Respondents were asked to rate 10 hypothetical pediatric patients and answer four demographic questions. RESULTS: We received 267 surveys, yielding a response rate of 66.8% and the highest number of responses in any study of this nature. The spread of answers was wide across almost all cases. Only one case had a response spread of only two classifications, with the remaining cases having three or more different ASA PS classifications chosen. The most variability was found for a hypothetical patient with severe trauma, who received five different ASA PS classifications. The Modified Kappa Statistic was 0.5, suggesting moderate agreement. No significant difference between the private and academic anesthesiologists was found (P = 0.26). CONCLUSIONS: We present the largest evaluation of interobserver consistency in ASA PS in pediatric patients by pediatric anesthesiologists. We conclude that agreement between anesthesiologists is only moderate and suggest standardizing assessment, so that it reflects the patient status at the time of anesthesia, including any acute medical or surgical conditions.     

The frail fail: Increased mortality and post-operative complications in orthopaedic trauma patients

ObjectiveThe burgeoning elderly population calls for a robust tool to identify patients with increased risk of mortality and morbidity. This paper investigates the utility of the MFI as a predictor of morbidity and mortality in orthopaedic trauma patients.DesignRetrospective review of the NSQIP database to identify patients age 60 and above who underwent surgery for pelvis and lower extremity fractures between 2005 and 2014.Main outcomes and measuresFor each patient, an MFI score was calculated using NSQIP variables. The relationship between the MFI score and 30-day mortality and morbidity was determined using chi-square analysis. MFI was compared to age, American Society of Anesthesiologists physical status classification, and wound classifications in multiple logistic regression.ResultsStudy sample consisted of 36,424 patients with 27.8% male with an average age of 79.5 years (SD 9.3). MFI ranged from 0 to 0.82 with mean MFI of 0.12 (SD 0.09). Mortality increased from 2.7% to 13.2% and readmission increased from 5.5% to 18.8% with increasing MFI score. The rate of any complication increased from 30.1% to 38.6%. Length of hospital stay increased from 5.3 days (±5.5 days) to 9.1 days (±7.2 days) between MFI score 0 and 0.45+. There was a stronger association between 30-day mortality and MFI (aOR for MFI 0.45+: 2.6, 95% CI: 1.7–3.9) compared to age (aOR for age: 1.1, 95% CI: 1.1-1.1) and ASA (aOR 2.5, 95% CI: 2.3–2.7).Conclusions and relevanceMFI was a significant predictor of morbidity and mortality in orthopaedic trauma patients. The use of MFI can provide an individualized risk assessment tool that can be used by an interdisciplinary team for perioperative counseling and to improve outcomes.     

Adrenalectomy in Veterans Affairs and selected university medical centers: results of the patient safety in surgery study

BACKGROUND: Data from the Patient Safety in Surgery Study were used to compare preoperative risk factors, intraoperative variables, and surgical outcomes of adrenalectomy procedures performed in 81 Veterans Affairs (VA) hospitals with those performed in 14 private-sector (PS) hospitals. STUDY DESIGN: This study is a retrospective review of prospectively collected data on all patients undergoing adrenalectomy in the VA and PS for fiscal years 2002 through 2004. Bivariate analysis compared VA and PS preoperative risk factors, intraoperative variables, and 30-day morbidity and mortality. Regression risk-adjustment analysis was used to compare 30-day postoperative morbidity in the VA and PS. RESULTS: During the 3 years studied, 178 VA patients and 371 PS patients underwent adrenalectomy procedures with a median per site of 2 (range 1-9) and 21 (range 8-70) procedures per VA and PS hospital, respectively. The VA patients had considerably more comorbidities than PS patients. The unadjusted 30-day morbidity rate was significantly higher in VA (16.29%) than PS (6.74%) hospitals (p = 0.0003); after controlling for the higher rate of comorbidities, the adjusted odds ratio for morbidity in the VA versus the PS hospitals was no longer significant (odds ratio = 1.328; 95% CI, 0.488-3.613). Unadjusted mortality rate was VA 2.81%, PS 0.27%, p = 0.0074. The low event rate overall precluded risk adjustment for mortality. CONCLUSIONS: The VA adrenalectomy population has more preoperative risk factors and substantially higher unadjusted 30-day postoperative morbidity and mortality rates than the PS population. After risk adjustment, there is no significant difference in morbidity between the VA and the PS. A larger study population is needed to compare risk-adjusted mortality between the VA and PS.     

A pilot study to predict high cost patients in the pre-operative setting

Introduction: The National Surgical Quality Improvement Program (NSQIP) previously demonstrated that certain clinical variables predict surgical mortality and morbidity. We examined whether these clinical variables could also predict the cost of care in the private sector. Methods: All 1,008 patients enrolled into the NSQIP at our institution between October 2001 and June 2002 were tracked for cost incurred by the health system using our internal cost accounting database. The original data collection of clinical data and outcomes was via a trained nurse reviewer through direct chart review and patient interview. A model for predicting cost was created via multiple linear regression systematically testing 44 preoperative variables against log-transformed cost. This model was then recalculated using the statistically significant variables from the initial model with the inclusion of a variable denoting occurrence of a complication. Testing was also performed to fit the model to only those without post-operative complications as well as those that survived. Results: While 31 variables were significant when tested separately, after considering interaction, a single model was derived from the 15 statistically significant (p < 0.05) preoperative variables predicted 65% of the variation in hospital costs (adjusted R² = 0.650). Top predictors of increased costs included: operation requiring inpatient stay, high ASA classification, low albumin, use of general anesthesia, high surgical complexity, and high BUN. Though complications were significantly correlated with increased costs, including whether complications actually occurred only improved the best overall model’s predictive capability by an additional 2% of the variation in costs for the entire population (adjusted R² = 0.669). Testing the model’s output in the subset of patients that avoided complications yielded an R² of 0.438. Conclusions: High cost patients can be predicted in the preoperative setting. These factors predicted higher costs even cases that did not have complications as viewed after the fact. It may be feasible to create benchmarking studies that “risk-adjust” costs as they relate to specific patient populations, which will allow for comparisons across institutions of cost-effectiveness. Institutions treating higher risk patients should seek increased reimbursement for these populations in order to match costs with revenues.     

Evaluation of surgical risk in elderly patients: a review of 207 cases

Thorough evaluation of surgical risk represents the sine qua non for a correct therapeutic choice particularly in the elderly who are frequently affected by multiple pathologies. The aim of this study was to evaluate the prognostic value of two of the most common classification systems for predicting surgical risk (ASA and Reiss scores) and of other laboratory parameters. A consecutive series of 207 patients aged 70 or above were analysed retrospectively, considering age, ASA and Reiss scores, elective or emergency surgery, operative time, leucocytes, haemoglobin, creatinine, and albumin levels. Morbidity and mortality rates were compared in relation to these parameters. Emergency surgery was associated with significantly higher morbidity (P = 0.006 chi-square) and mortality (P = 0.001 chi-square) than elective surgery. No differences in morbidity were noted in association with the ASA classification (P = 0.07 chi-square), though there was a significant difference (P = 0.001 chi-square) in mortality. Significant differences in both morbidity (P = 0.04 chi-square) and mortality (P = 0.001 chi-square) were found to be associated with the Reiss classification. Multivariate analysis showed that ASA score (P = 0.006), Reiss score (P = 0.004), operative time (P = 0.005), and haemoglobin level (P = 0.01) were independent prognostic factors. The results of the study confirm the prognostic value of multiparametric classifications such as the ASA and Reiss score in elderly patients, even if the addition of other prognostic factors may be expected to improve the sensitivity.     

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. National VA Surgical Quality Improvement Program.

OBJECTIVE: To provide reliable risk-adjusted morbidity and mortality rates after major surgery to the 123 Veterans Affairs Medical Centers (VAMCs) performing major surgery, and to use risk-adjusted outcomes in the monitoring and improvement of the quality of surgical care to all veterans. SUMMARY BACKGROUND DATA: Outcome-based comparative measures of the quality of surgical care among surgical services and surgical subspecialties have been elusive. METHODS: This study included prospective assessment of presurgical risk factors, process of care during surgery, and outcomes 30 days after surgery on veterans undergoing major surgery in 123 medical centers; development of multivariable risk-adjustment models; identification of high and low outlier facilities by observed-to-expected outcome ratios; and generation of annual reports of comparative outcomes to all surgical services in the Veterans Health Administration (VHA). RESULTS: The National VA Surgical Quality Improvement Program (NSQIP) data base includes 417,944 major surgical procedures performed between October 1, 1991, and September 30, 1997. In FY97, 11 VAMCs were low outliers for risk-adjusted observed-to-expected mortality ratios; 13 VAMCs were high outliers for risk-adjusted observed-to-expected mortality ratios. Identification of high and low outliers by unadjusted mortality rates would have ascribed an outlier status incorrectly to 25 of 39 hospitals, an error rate of 64%. Since 1994, the 30-day mortality and morbidity rates for major surgery have fallen 9% and 30%, respectively. CONCLUSIONS: Reliable, valid information on patient presurgical risk factors, process of care during surgery, and 30-day morbidity and mortality rates is available for all major surgical procedures in the 123 VAMCs performing surgery in the VHA. With this information, the VHA has established the first prospective outcome-based program for comparative assessment and enhancement of the quality of surgical care among multiple institutions for several surgical subspecialties. Key features to the success of the NSQIP are the support of the surgeons who practice in the VHA, consistent clinical definitions and data collection by dedicated nurses, a uniform nationwide informatics system, and the support of VHA administration and managerial staff.     

Patient factors,comorbidities, and surgical characteristics that increase mortality and complication risk after spinal arthrodesis: a prognostic study based on 5,887 patients

Background context

The impact of patient factors and medical comorbidities on the risk of mortality and complications after spinal arthrodesis has not been well described. Prior works have been limited by small sample size, single center data, or the inability to be broadly generalized.

Purpose

To determine if there is an association between the patient demographic factors, comorbidities, nutritional status, and surgical characteristics and the occurrence of mortality and complications after spinal arthrodesis.

Study design

Retrospective review of prospectively collected data in the National Surgical Quality Improvement Program (NSQIP).

Patient sample

Patients who underwent spinal arthrodesis and had data registered with the NSQIP between 2005 and 2010.

Outcome measures

Primary outcomes were death or any complication after spinal arthrodesis. Secondary measures were the development of a specific complication, including wound infection, thromboembolic disease, or cardiac arrest/myocardial infarction.

Methods

The data set of the NSQIP from 2005 to 2010 was queried to identify all patients who underwent spinal arthrodesis. Demographic information, body mass index (BMI), medical comorbidities, arthrodesis procedure, operative time, American Society of Anesthesiologists (ASA) classification, and preoperative albumin were recorded for all patients identified. Mortality, the development of postoperative complications, and the presence of specific complications were also abstracted. Risk factors for mortality and complications were initially evaluated using chi-square and univariate logistic regression analyses. The risk factors that maintained p values less than .2 in univariate analysis were then combined in a multivariate fashion that identified significant, independent, predictors of mortality and complications while controlling for other factors present in the model. Sensitivity analysis was also performed, discriminating between the impact of risk factors on major and minor complications and the relative contribution to overall risk of morbidity. Multivariate analysis resulted in odds ratios (ORs) with 95% confidence intervals (CIs) for each risk factor. Only those predictors with ORs and 95% CI exclusive of 1.0 and p values less than .05 were considered statistically significant.

Results

In all, 5,887 patients who underwent spinal arthrodesis were identified. The average age of patients was 55.9 (±14.5) years. Twenty-five (0.42%) patients died after surgery, whereas 608 (10%) sustained a complication. Wound infection was the most common specific complication occurring in 2% of the cohort. Age (p=.03) and pulmonary conditions (p=.002) were found to have a significant association with the risk of mortality. Age exceeding 80 years was found to carry the highest risk of mortality. Age, pulmonary conditions, BMI, history of infection, ASA classification more than 2, neurologic conditions, resident (i.e., trainee) involvement, and procedural times exceeding 309 minutes increased the risk of complications. Body mass index, ASA classification more than 2, resident involvement, and procedural times exceeding 309 minutes were associated with the risk of infection. Although limited to univariate analysis, serum albumin 3.5 g/dL or less increased the risk of mortality, complications, wound infection, and thromboembolic disease. The OR for postoperative mortality among patients with albumin 3.5 g/dL or less was 13.8 (95% CI, 4.6–41.6; p<.001).

Conclusions

Several factors, including patients' age, BMI, ASA classification more than 2, pulmonary conditions, procedural times, and nutritional status likely influence the risk of postoperative morbidity to varying degrees. The risk factors identified here may be more generalizable to the American population as a whole because of the design and methodology of the NSQIP in comparison with previously published studies.     

Bariatric operations in Veterans Affairs and selected university medical centers: results of the patient safety in surgery study

BACKGROUND: The objectives of this study were to evaluate outcomes and predictors of morbidity in patients undergoing Roux-en-Y gastric bypass (RYGB) during the Patient Safety in Surgery (PSS) Study. STUDY DESIGN: National Surgical Quality Improvement Program data on PSS patients undergoing RYGB were analyzed for unadjusted and adjusted outcomes. Gender groups acted differently and were analyzed separately. Multivariable regression modeling was used to analyze hospital type as a predictor of risk. Stepwise logistic regression was performed to determine patient factors predictive of postoperative morbidity. RESULTS: A total of 2,438 patients (2,064 private sector [PS], 374 Veterans Affairs [VA]) were identified for analysis. Adjusted odds ratio for postoperative morbidity for VA versus PS female patients was 1.14 (95% CI, 0.63-2.05), and for male patients 2.29 (95% CI, 1.28-4.10). Stepwise logistic regression showed that independent risk factors predictive of morbidity were open procedure, higher American Society of Anesthesiologists class, higher body mass index, diabetes, alcohol consumption, leukocytosis, SGOT > 40 U/L, smoking history, and older age. Importantly, male gender was not significant (p = 0.13) in the regression analysis. Subsequent and unrelated to this study, the VA has restructured its bariatric surgical program, including regionalization of centers, with a substantial lowering of associated mortality and morbidity. CONCLUSIONS: The VA male subset showed higher risk-adjusted postoperative morbidity compared with the PS male subset. The VA and PS female subsets had equivalent risk-adjusted postoperative morbidity. A systematic approach to quality-improvement processes resulted in improved bariatric surgical outcomes in the VA. Male gender might not be an independent risk factor in RYGB patients.     

Outcomes of simultaneous laparoscopic cholecystectomy and ventral hernia repair compared to that of laparoscopic cholecystectomy alone

Background

Although incidental hernias frequently are found and repaired during laparoscopic cholecystectomy (LC), the outcomes of simultaneous LC and laparoscopic ventral hernia repair (LVHR) have not been scrutinized. In this study we evaluated short-term outcome data comparing simultaneous LC and LVHR against LC alone.

Methods

The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database (2005–2009) was queried using primary procedure and secondary current procedural terminology (CPT^®) codes for LC and LVHR. Outcomes analyzed included separate LC and LVHR and simultaneous laparoscopic cholecystectomy and ventral hernia repair (LC/LVHR). The 30 day clinical outcomes along with postoperative hospital length of stay (LOS) were assessed using the χ² test and analysis-of-variance test with p values < 0.01 set as significant. We also performed forward stepwise multivariable regression taking in to consideration over 50 ACS NSQIP risk factors to adjust for patient risk.

Results

A total of 82,837 patients underwent LC and/or LVHR of which 357 (0.4 %) underwent simultaneous LC/LVHR. Patients who underwent LC/LVHR were more likely to have surgical site infections, suffer sepsis or septic shock, and have pulmonary complications, including pneumonia, reintubation or prolonged ventilator requirements, than LC-alone patients. No difference was noted in 30 day mortality, rates of deep vein thrombosis/pulmonary embolism (DVT/PE), renal insufficiency, or stroke. After multivariable adjustment for over 50 ACS NSQIP risk factors, concurrent LC/LVHR continued to pose a higher risk for these outcomes relative to LC only.

Conclusions

Simultaneous LC/LVHR results in greater postoperative morbidity in terms of surgical site infections, sepsis, and pulmonary complications when compared to LC alone. In light of this increased short-term morbidity, consideration should be given toward performing LC and LVHR independently in patients requiring both procedures. Prospective studies with long-term follow-up are required to better understand the implications of simultaneous LC/LVHR.     

Surgical morbidity and mortality among American Indian and Alaska Native veterans: a comparative analysis

BACKGROUND: Few studies have examined surgical risk factors and outcomes in American Indians and Alaska Natives (AI/ANs). My colleagues and I sought to determine if prevalence of preoperative risk factors for morbidity and mortality differed between male AI/AN and Caucasian surgical patients, and to determine if AI/ANs had an increased risk of surgical morbidity or mortality. STUDY DESIGN: We obtained data from the Veterans Affairs National Surgical Quality Improvement Program on major, noncardiac, surgical procedures performed between 1991 and 2002 for all AI/AN men (n = 2,155) and a random sample of Caucasian men (n = 2,264), matched by facility. Chi-square and t-test analyses were used to assess differences in preoperative risk factors between the two groups. Logistic regression was used to determine whether AI/AN race was independently associated with 30-day morbidity (defined as 1 or more of 21 postoperative complications) or 30-day all cause mortality after adjustment for major risk factors. RESULTS: Prevalence of major preoperative risk factors for morbidity and mortality often differed between the groups. Compared with Caucasians, AI/AN race did not predict morbidity (adjusted odds ratio, 0.92; 95% CI, 0.75-1.13), but AI/ANs were at higher risk for 30-day all cause postoperative mortality (adjusted odds ratio, 1.56; 95% CI, 1.04-2.35). CONCLUSIONS: Our results add postoperative mortality to health disparities experienced by AI/ANs. Future research should be conducted to identify other factors that contribute to this disparity.     

Use of national surgical quality improvement program data as a catalyst for quality improvement

BACKGROUND: Semiannually, the National Surgical Quality Improvement Program (NSQIP) provides its participating sites with observed-to-expected (O/E) ratios for 30-day postoperative mortality and morbidity. At each reporting period, there is typically a small group of hospitals with statistically significantly high O/E ratios, meaning that their patients have experienced more adverse events than would be expected on the basis of the population characteristics. An important issue is to determine which actions a surgical service should take in the presence of a high O/E ratio. STUDY DESIGN: This article reviews case studies of how some of the Department of Veterans Affairs and private-sector NSQIP participating sites used the clinically rich NSQIP database for local quality improvement efforts. Data on postoperative adverse events before and after these local quality improvement efforts are presented. RESULTS: After local quality improvement efforts, wound complication rates were reduced at the Salt Lake City Veterans Affairs medical center by 47%, surgical site infections in patients undergoing intraabdominal surgery were reduced at the University of Virginia by 36%, and urinary tract infections in vascular patients were reduced at the Massachusetts General Hospital by 74%. At some sites participating in the NSQIP, notably the Massachusetts General Hospital and the University of Virginia, the NSQIP has served as the basis for surgical service-wide outcomes research and quality improvement programs. CONCLUSIONS: The NSQIP not only provides participating sites with risk-adjusted surgical mortality and morbidity outcomes semiannually, but the clinically rich NSQIP database can also serve as a catalyst for local quality improvement programs to significantly reduce postoperative adverse event rates.     

The impact of operative complexity on patient risk factors

OBJECTIVE: The VA National Surgical Quality Improvement Program (NSQIP) formula for risk factors was applied to the University of Texas Health Science Center at San Antonio (UTHSCSA)/University Hospital (UH) database. Its applicability to a civilian organization was established. Logistic regression analysis of the UH database revealed that operative complexity was significantly related to mortality only at high complexity levels. Patient risk factors were the major determinants of operative mortality for most civilian surgical cases. SUMMARY BACKGROUND DATA: Since 1994, the NSQIP has collected preoperative risk factors, intraoperative data, 30-day morbidity, and mortality within the VA health system. The VA formula to predict 30-day postoperative mortality was applied to our UH patients (N = 8593). The c-index of .907, a statistical measure of accuracy, compared favorably to the VA patient c-index of .89. The UH database did not include a surrogate for operative complexity. We were elated by the predictive accuracy but had concern that operative complexity needed further evaluation. METHODS: Operative complexity was ascribed to each of the 8593 UH cases, and logistic regression analyses were compared with and without operative complexity. Operative complexity was graded on a scale of 1 to 5; 5 was the most complex. RESULTS: Without operative complexity, a c-index was .915. With operative complexity: an even higher c-index of .941 was reached. The large volume of level 2-3 operative cases obscured to a degree the effect of operative difficulty on mortality. CONCLUSION: Operative complexity played a major role in risk estimation, but only at the extreme. The dominance of cases of midlevel complexity masked the effect of higher complexity cases on mortality. In any individual case, operative complexity must be added to estimate operative mortality accurately. Patient risk factors alone accounted for operative mortality for operations less than level 4 (95% of patients).     

Relationship of processes and structures of care in general surgery to postoperative outcomes: a hierarchical analysis

BACKGROUND: The majority of studies relating processes and structures of surgical care to outcomes focus on mortality alone, even though morbidity outcomes are frequent, costly, and can have an adverse effect on a patient's short- and longterm survival and quality of life. The purpose of this study was to identify the important processes and structures of surgical care that relate to 30-day, risk-adjusted postoperative morbidity in general surgery. STUDY DESIGN: Department of Veterans Affairs general surgery patients operated on in the period October 1, 2003 to September 30, 2004 at medical centers that participated in the Patient Safety in Surgery (PSS) Study and responded to a process and structure of care survey were included in this study. The patient's risk information was combined with key process and structure variables in a hierarchical maximum likelihood analysis to predict 30-day postoperative morbidity. RESULTS: A number of hospital-level processes and structures of care were identified that predicted 30-day postoperative morbidity. The dominant factor was university affiliation. Affiliated hospitals showed an increase in risk of morbidity even after adjustment for patient risk. CONCLUSIONS: Risk-adjusted morbidity is higher in Veterans Affairs hospitals that are affiliated with university medical centers. These findings mandate additional study to identify the exact factors responsible for this increased morbidity.     

External validation of the Codman score in colorectal surgery: a pragmatic tool to drive quality improvement

Aim

The simple six-variable Codman score is a tool designed to reduce the complexity of contemporary risk-adjusted postoperative mortality rate predictions. We sought to externally validate the Codman score in colorectal surgery.

Methods

The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) participant user file and colectomy targeted dataset of 2020 were merged. A Codman score (composed of six variables: age, American Society of Anesthesiologists score, emergency status, degree of sepsis, functional status and preoperative blood transfusion) was assigned to every patient. The primary outcome was in-hospital mortality and secondary outcome was morbidity at 30 days. Logistic regression analyses were performed using the Codman score and the ACS NSQIP mortality and morbidity algorithms as independent variables for the primary and secondary outcomes. The predictive performance of discrimination area under receiver operating curve (AUC) and calibration of the Codman score and these algorithms were compared.

Results

A total of 40 589 patients were included and a Codman score was generated for 40 557 (99.02%) patients. The median Codman score was 3 (interquartile range 1–4). To predict mortality, the Codman score had an AUC of 0.92 (95% CI 0.91–0.93) compared to the NSQIP mortality score 0.93 (95% CI 0.92–0.94). To predict morbidity, the Codman score had an AUC of 0.68 (95% CI 0.66–0.68) compared to the NSQIP morbidity score 0.72 (95% CI 0.71–0.73). When body mass index and surgical approach was added to the Codman score, the performance was no different to the NSQIP morbidity score. The calibration of observed versus expected predictions was almost perfect for both the morbidity and mortality NSQIP predictions, and only well fitted for Codman scores of less than 4 and greater than 7.

Conclusion

We propose that the six-variable Codman score is an efficient and actionable method for generating validated risk-adjusted outcome predictions and comparative benchmarks to drive quality improvement in colorectal surgery.     

Complications after Extended (D2) and Superextended (D3) Lymphadenectomy for Gastric Cancer: Analysis of Potential Risk Factors

BACKGROUND: Very few studies from Western centers have compared D2 and D3 dissection in the surgical treatment of gastric cancer. The aim of the prospective observational study reported here was to analyze the postoperative outcome and potential risk factors for complications following D2 and D3 lymphadenectomy. METHODS: A total of 330 consecutive patients, of which 251 submitted to D2 lymphadenectomy and 79 were treated by D3 lymphadenectomy, were enrolled in the study. Twenty potential risk factors for morbidity and mortality were studied by means of univariate and multivariate analysis. RESULTS: Overall morbidity and mortality rates were 34% (111 patients) and 4% (14 patients), respectively. Abdominal abscess, anastomotic leakage, pleuropulmonary diseases and pancreatitis were the most commonly observed complications. No differences in morbidity, surgical morbidity, mortality rates and mean hospital stay between D2 and D3 lymphadenectomy were found. Multivariate analysis revealed that American Society of Anesthesiologists' (ASA) class II/III versus class I, perioperative blood transfusions, and low albumin serum levels were independent predictors of postoperative complications. Age, surgical radicality (R1/R2 vs. R0) and low albumin serum levels independently predicted mortality. Mortality rate was .5% in the 203 patients aged 75 years or younger who underwent curative surgery. Most of deaths were observed in patients older than 75 years with low albumin serum levels or treated by non-curative surgery. CONCLUSIONS: D2 lymphadenectomy represents a feasible procedure associated to acceptable morbidity and mortality rates. In specialized centers, D3 lymphadenectomy may be performed without increasing the risk of postoperative complications and associated deaths in carefully selected patients. These techniques should be avoided in subgroups of patients with a high risk of postoperative mortality.     

The National Surgical Quality Improvement Program in non-veterans administration hospitals: initial demonstration of feasibility

OBJECTIVE: To assess the feasibility of implementing the National Surgical Quality Improvement Program (NSQIP) methodology in non-VA hospitals. SUMMARY BACKGROUND DATA: Using data adjusted for patient preoperative risk, the NSQIP compares the performance of all VA hospitals performing major surgery and anonymously compares these hospitals using the ratio of observed to expected adverse events. These results are provided to each hospital and used to identify areas for improvement. Since the NSQIP's inception in 1994, the VA has reported consistent improvements in all surgery performance measures. Given the success of the NSQIP within the VA, as well as the lack of a comparable system in non-VA hospitals, this pilot study was undertaken to test the applicability of the NSQIP models and methodology in the nonfederal sector. METHODS: Beginning in 1999, three academic medical centers (Emory University, Atlanta, GA; University of Michigan, Ann Arbor, MI; University of Kentucky, Lexington, KY) volunteered the time of a dedicated surgical nurse reviewer who was trained in NSQIP methodology. At each academic center, these nurse reviewers used NSQIP protocols to abstract clinical data from general surgery and vascular surgery patients. Data were manually collected and then transmitted via the Internet to a secure web site developed by the NSQIP. These data were compared to the data for general and vascular surgery patients collected during a concurrent time period (10/99 to 9/00) within the VA by the NSQIP. Logistic regression models were developed for both non-VA and VA hospital data. To assess the models' predictive values, C-indices (0.5 = no prediction; 1.0 = perfect prediction) were calculated after applying the models to the non-VA as well as the VA databases. RESULTS: Data from 2,747 (general surgery 2,251; vascular surgery 496) non-VA hospital cases were compared to data from 41,360 (general surgery 31,393; vascular surgery 9,967) VA cases. The bivariate relationships between individual risk factors and 30-day mortality or morbidity were similar in the non-VA and VA patient populations for over 66% of the risk variables. C-indices of 0.942 (general surgery), 0.915 (vascular surgery), and 0.934 (general plus vascular surgery) were obtained following application of the VA NSQIP mortality model to the non-VA patient data. Lower C-indices (0.778, general surgery; 0.638, vascular surgery; 0.760, general plus vascular surgery) were obtained following application of the VA NSQIP morbidity model to the non-VA patient data. Although the non-VA sample size was smaller than the VA, preliminary analysis suggested no differences in risk-adjusted mortality between the non-VA and VA cohorts. CONCLUSIONS With some adjustments, the NSQIP methodology can be implemented and generates reasonable predictive models within non-VA hospitals.