Comparison of Risk-Adjusted 30-Day Postoperative Mortality and Morbidity in Department of Veterans Affairs Hospitals and Selected University Medical Centers: General Surgical Operations in Women

Open Access
Publisher Website
Google Scholar
Abstract
Results Data from 5,157 female general surgical VA patients who underwent eligible procedures were compared with those from 27,467 patients in the private sector. Unadjusted 30-day mortality was virtually identical in the two groups (1.3%). The unadjusted morbidity rate was slightly, but notably, higher in the private sector (10.9%) as compared with that observed in the VA (8.5%, p < 0.0001). Predictive models were generated for mortality and morbidity combining both groups; top variables in these models were similar to those described previously in the National Surgical Quality Improvement Program. The indicator variable for system of care (VA versus private sector) was not statistically significant in the mortality model, but substantially favored the VA in the morbidity model (odds ratio = 0.80, 95% CI = 0.71, 0.90). Conclusions The data demonstrate that in female general surgical patients, risk-adjusted mortality rates are comparable in the VA and the private sector, but risk-adjusted morbidity is higher in the private sector. Rates of urinary tract infections in the two populations may account for much of the latter difference. Abbreviations and Acronyms ASA American Society of Anesthesiologists NSQIP National Surgical Quality Improvement Program NVASRS National VA Surgical Risk Study PSS Patient Safety in Surgery RVU relative value unit VA Department of Veterans Affairs Faced with considerable negative press impugning the quality of surgical care offered by the Department of Veterans Affairs (VA), Congress passed Public Law 99-166 in 1985. 1 This bill mandated that the VA compare its surgical outcomes to those experienced in the private sector. To deal with claims by VA surgeons that “their patients were sicker,” the bill also mandated that clinical risk adjustment be integrated into the comparison of outcomes. Given that there was no formalized mechanism with which to comply with these mandates, the VA ultimately funded the National VA Surgical Risk Study (NVASRS), supporting development and validation of the methodology needed to comply with the congressional mandate. Ultimately, the NVASRS provided the VA with a validated tool with which to monitor, compare, and improve the quality of surgery in all VA hospitals performing major surgery. 2-5 Based on the success of the NVASRS, the VA supported establishment of the National Surgical Quality Improvement Program (NSQIP), allowing the structure developed in the NVASRS to be disseminated to all VA hospitals performing major surgery. Since its inception in 1994, the NSQIP has served a critical role in the continuous monitoring and enhancement of the quality of surgical care in the VA health care system. 6 Unfortunately, with the exception of specific subspecialties such as cardiac surgery, no such program allowed assessment and comparison of risk-adjusted surgical outcomes between private sector hospitals. With the latter goal in mind, in 1999, a preliminary attempt was made to implement the NSQIP’s case identification and data collection methodology in three private sector hospitals. 7 This successful trial demonstrated in a preliminary fashion the applicability of the NSQIP to the private sector, and prompted initiation of a larger trial, the Patient Safety in Surgery (PSS) Study, which was conducted in 128 VA hospitals and 14 private academic medical centers between 2001 and 2004. 8 In this article, we report the results observed in female general surgical patients included in the PSS Study and add to the series of reports with which the demands of the 1985 congressional mandate are finally being met. Methods This study used the methods of the NSQIP and the PSS Study, which have been described in detail in other publications. 2-5,9 A brief description of the pertinent methodology is provided in the proceeding text. Patient population In the NSQIP, data are abstracted from all patients having operations using general, spinal, or epidural anesthesia, with the exception of patients undergoing operations in the previous 30-day period or having selected operations with minimal postoperative mortality and morbidity. Trauma and transplantation procedures are also excluded, given their infrequent performance within the VA system. Certain very common operations, eg, inguinal hernia repair, breast lumpectomies, or transurethral resection of prostate or bladder tumors are limited to the first five consecutive patients in each 8-day cycle. In hospitals that perform more than 140 eligible operations per month, operations are sampled by including the first 40 operations (36 in the VA) in each 8-day cycle, beginning each cycle on a different day of the week. Although the VA NSQIP includes nine surgical subspecialties: general, vascular, cardiac, noncardiac thoracic, orthopaedics, urology, neurosurgery, otolaryngology, and plastic surgery, the PSS Study and this analysis were limited to general and vascular surgery. Variable selection NSQIP variables were selected on the basis of clinical relevance and the reliability, availability, and ease of data collection. Preoperative variables included demographics, some lifestyle variables, functional status, American Society of Anesthesiologists (ASA) classification, selected laboratory tests, and selected systemic (pulmonary, cardiac, hepatobiliary, renal, vascular, central nervous system, nutritional, and immunologic) comorbidities. Operative data included Current Procedural Terminology (CPT) codes for the principal operation and any secondary operations, emergency case status, wound classification, anesthesia method, operative times, blood loss, and transfusions. Outcomes variables included 30-day mortality from any cause inside or outside the hospital, length of stay, return to the operating room, and 19 different postoperative complications occurring in the 30-day postoperative period. Because the NSQIP was designed to cover all major operations performed in a VA surgical...