Linear regression (3040 hospitals) of VTE rate on percentage of surgical patients receiving timely VTE prophylaxis did not show a significant association (95% CI, −1.65% to 13.35%; P = .13). The horizontal red line represents the simple linear regression line. AHRQ indicates Agency for Healthcare Research and Quality; PSI, Patient Safety Indicator; SCIP, Surgical Care Improvement Project; and VTE, venous thromboembolism.
The Kruskal-Wallis rank test was used to determine the equality of population distributions (P = .04). None of the Bonferroni-corrected rank sum tests for pairwise comparisons of quintiles were statistically significant (all P > .005). The horizontal line in the middle of each box indicates the median, whereas the top and bottom borders of the box mark the 75th and 25th percentiles, respectively. The whiskers above and below the box are Tukey fences marking the most extreme observed values within 1.5 times the interquartile range above the 75th and below the 25th percentiles. AHRQ indicates Agency for Healthcare Research and Quality; PSI, Patient Safety Indicator; SCIP, Surgical Care Improvement Project; and VTE, venous thromboembolism.
JohnBull EA, Lau BD, Schneider EB, Streiff MB, Haut ER. No Association Between Hospital-Reported Perioperative Venous Thromboembolism Prophylaxis and Outcome Rates in Publicly Reported Data. JAMA Surg. 2014;149(4):400-401. doi:10.1001/jamasurg.2013.4935
Venous thromboembolism (VTE), consisting of deep vein thrombosis (DVT) and pulmonary embolism, is an important cause of postoperative mortality and long-term morbidity. Because many events are preventable, VTE prophylaxis performance and postoperative VTE are used as measures of hospital quality of care and patient safety. Two such metrics are reported on the Centers for Medicare & Medicaid Services Hospital Compare website (http://www.medicare.gov/hospitalcompare/search.html) with the stated goal of helping consumers make decisions about where to receive their health care.
The Surgical Care Improvement Project (SCIP) VTE-2, a process measure, captures the percentage of a hospital’s surgical patients who received any VTE prophylaxis within 24 hours of surgery. Previous studies have shown that performance on the SCIP VTE-2 measure is not associated with VTE among Medicare patients in 6 high-risk surgical procedures,1 nor with outcomes in a Veterans Affairs cohort.2 In October 2011, Hospital Compare made additional VTE outcomes data available to consumers. We hypothesized that there is no association between the process measure (SCIP VTE-2) and the outcome (Agency for Healthcare Research and Quality Patient Safety Indicator [PSI]–12, a risk-adjusted postoperative VTE rate based on administrative data).
The SCIP VTE-2 and PSI-12 data were downloaded from the Centers for Medicare & Medicaid Services website. Average annual prophylaxis performance was linked to the 2-year VTE rate for each hospital (from July 1, 2009, to June 30, 2011). The PSI-12 is adjusted for age, sex, age × sex interactions, diagnosis-related group, and modified comorbidity index prior to public reporting. Simple linear regression was used to test for an association between VTE prophylaxis and VTE rate. Sensitivity analyses were performed that excluded outliers and that used a log-transformed VTE rate because the data were not normally distributed. Hospitals were categorized by quintile of prophylaxis performance, and VTE rates were compared using a nonparametric Kruskal-Wallis rank test and Bonferroni-corrected rank sum tests for multiple pairwise comparisons. Hospitals with 100% prophylaxis performance were compared to the lowest quintile using a rank sum test. Statistical analyses were conducted using Stata version 11.2 (StataCorp).
There were 3040 hospitals with complete prophylaxis and VTE data. Average annual prophylaxis performance was negatively skewed (median rate, 94.5%). The median risk-adjusted VTE rate was 4.13 per 1000 surgical discharges. Prophylaxis performance was not associated with VTE rate (P = .13) on linear regression (Figure 1). Regression results were unchanged in sensitivity analyses that excluded outliers (P = .51) and that used a log-transformed VTE rate (P = .90). Among quintiles, the VTE rates appeared to be similar (Figure 2). Although the global Kruskal-Wallis test was significant (P = .04), no pairwise comparison by quintile was statistically significant. Hospitals reporting 100% VTE prophylaxis performance (n = 141) and hospitals in the bottom quintile of prophylaxis performance (n = 618) had nearly identical median VTE outcome rates (4.18 vs 4.17; P = .98, determined by rank sum test).
Our findings show that there is no association between reported VTE prophylaxis and outcome rates for surgical patients, when compared across the range of process measure performance. There was no difference in quintiles, or between extremes at 100% performance and the lowest quintile. This may reflect the current low, minimum standard for VTE prophylaxis because the SCIP VTE-2 gives credit for even the most basic, suboptimal prophylaxis. Perhaps a more rigorously applied benchmark to report adherence to optimal, risk-appropriate prophylaxis would have a greater chance of differentiating hospitals and improving outcomes. Furthermore, VTE rates are greatly influenced by surveillance bias,3 which likely weakens any association between process and outcome.
To be useful to consumers, hospital quality measures should be associated with important patient outcomes.4 Biased information, presented as evidence of hospital quality, should not inform the decision-making process. Hospital Compare–reported SCIP VTE-2 prophylaxis rates are of little value to consumers choosing a hospital for surgical treatment. Our results suggest a need to explore alternate public-reporting strategies or metrics, such as those that combine process and outcome to identify preventable harm5,6 and quantify VTE-related quality of care.
Corresponding Author: Elliott R. Haut, MD, Division of Acute Care Surgery, Department of Surgery, Johns Hopkins University School of Medicine, 1800 Orleans St, Sheikh Zayed 6107C, Baltimore, MD 21287 (firstname.lastname@example.org).
Published Online: February 5, 2014. doi:10.1001/jamasurg.2013.4935.
Author Contributions: Dr Haut and Mr JohnBull had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: JohnBull, Lau, Schneider, Haut.
Acquisition of data: JohnBull, Lau, Schneider.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: JohnBull.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: JohnBull, Lau, Schneider.
Obtained funding: Haut.
Administrative, technical, and material support: JohnBull, Lau.
Study supervision: Lau, Schneider, Haut.
Conflict of Interest Disclosures: Dr Haut is the primary investigator of the Mentored Clinician Scientist Development Award K08 1K08HS017952-01 from the Agency for Healthcare Research and Quality entitled “Does Screening Variability Make DVT an Unreliable Quality Measure of Trauma Care?” Dr Haut receives royalties from Lippincott Williams & Wilkins for a book he coauthored (ie, Avoiding Common ICU Errors). He has received honoraria for various speaking engagements regarding clinical and quality and safety topics, and has given expert witness testimony in various medical malpractice cases. Dr Streiff has received research funding from sanofi-aventis and Bristol-Myers Squibb, honoraria for CME lectures from sanofi-aventis and Ortho-McNeil Pharmaceutical, consulted for sanofi-aventis, Eisai, Daiichi-Sankyo, and Janssen Healthcare Innovation, and has given expert witness testimony in various medical malpractice cases. Mr Lau and Drs Streiff and Haut are supported by a contract from the Patient-Centered Outcomes Research Institute entitled “Preventing Venous Thromboembolism: Empowering Patients and Enabling Patient-Centered Care via Health Information Technology.” No other disclosures were reported.