G. Shadid1, S. E. Sharath1, J. C. Choi1, A. Roth1, C. Balentine2, M. Roberts3, E. Faridmoayer1, P. Kougias1 1SUNY Downstate Health Sciences University, Department Of Surgery, Brooklyn, NEW YORK, USA 2University of Wisconsin Madison School of Medicine and Public Health, Department Of Surgery, Madison, WISCONSIN, USA 3SUNY Downstate Health Sciences University, College Of Nursing, Brooklyn, NY, USA
Introduction: Studies have described venous thromboembolism (VTE) risks separately by race/ethnicity and by hospital structure, that is a hospital’s available resource reserves to screen for and implement VTE prophylaxis. These separate associations seem to indicate potential for interactive or mediating effects. The objective of this study was to examine whether hospital structure could act as a mediator or effect modifier on the association between race/ethnicity classification and post-operative VTE risk in a national multicenter cohort of patients undergoing operations for oncologic indications (Figure 1).
Methods: Data from the National Inpatient Sample, identified using International Classification of Diseases, Ninth/Tenth Revisions codes, were restricted to major surgeries for oncologic indications (January 1993-December 2020). Hospital structure (HS) was defined as a composite continuous measure of hospital size, teaching status, and facility-level payer mix. The primary composite outcome, VTE incidence, included postoperative deep vein thrombosis and pulmonary embolism. Adjusted logistic regression with 4-way decomposition was used to break down the joint and mediating effects of HS on race/ethnicity and VTE risk.
Results: We identified 1,459,532 cases performed over 28 years. The cohort was composed of individuals identifying as White/Caucasian (947,527; 80.1%) compared to Black/African American (121,734; 10.4%), Hispanic (68,435; 5.9%), Asian (28,732; 2.5%), and Native American (3,434; 0.4%). The unadjusted VTE rate was 1.0% (14,789). VTE risk was independently associated with increasing hospital size, teaching status, and increasing private payor proportions (aOR = 1.06, p<0.001). Unadjusted VTE incidence was highest among Black/African American individuals (1,638; 1.4%) – a characteristic that was consistent across both DVT (1,045; 0.9%) and PE (795; 0.7%). Proportions of excess risk/protection were largely and consistently explained by race/ethnicity classification rather than HS (either as a mediator or effect modifier). Adjusted total risks relative to White patients were: Black = 1.26 (p<0.001; HS explains <5% variation), Hispanic = 1.00 (p<0.001), Asian = 0.64 (p<0.001; HS explains <5% variation), and Native Americans = 0.66 (p<0.001; HS explains <5% variation).
Conclusion: Our findings suggest minimal HS influence in mitigating VTE incidence by race/ethnicity – indicating that the race/ethnicity effect is not influenced by hospital structure. Future research should further evaluate the reasons behind the increased in-hospital DVT rate for African American patients.
