Profiles in social vulnerability: The association of social determinants of health with postoperative surgical outcomes

Background: The association of different social vulnerability subthemes (ie, socioeconomic status, household composition and disability, minority status and language, and housing and transportation) with surgical outcomes remains poorly defined. The current study aimed to identify distinct profiles of social vulnerability among Medicare beneficiaries and define the association of these profiles with postoperative outcomes. Methods: The Medicare 100% Standard Analytic Files were used to identify patients undergoing lung resection, coronary artery bypass grafting, abdominal aortic aneurysm repair, and colectomy between 2013 and 2017. A cluster analysis was performed based on ranked scores across the 4 subthemes of the Centers for Disease Control and Prevention social vulnerability index. The likelihood of complications, extended length of stay, readmission, and mortality were assessed relative to cluster vulnerability profiles. Results: Among 852,449 Medicare beneficiaries undergoing surgery, median social vulnerability index among patients in the cohort was 49 (interquartile range: 24-74); cluster analysis revealed 5 vulnerability profiles that had heterogeneity in the vulnerability subthemes, even among patients with similar overall social vulnerability index scores. Postoperative outcomes differed across the 5 vulnerability profiles, with patients in the profiles characterized by higher overall vulnerability having worse postoperative outcomes. In particular, risk of complications (profile 1, 31.9% vs profile 5, 34.0%), extended length of stay ( profile 1, 21.7% vs profile 5, 24.3%), 30-day readmission (profile 1, 12.6% vs profile 5, 13.2%), and 30-day mortality (profile 1, 4.0% vs profile 5, 4.7%) was greater among patients with the highest vulnerability (all P < .01). Of note, surgical outcomes varied among patients who resided in communities with similar average social vulnerability index scores (social vulnerability index 49-54). In particular, patients in social vulnerability profile 4 had 26% increased odds of 30-day mortality compared to social vulnerability profile 2 patients (odds ratio 1.26, 95% confidence interval 1.21-1.30). Additionally, profile 3 patients had 15% higher odds of 30-day mortality versus profile 2 patients (odds ratio 1.15, 95% confidence interval 1.10-1.20). Conclusion: Postoperative outcomes differed across patients based on cluster vulnerability profiles. Despite similar overall aggregate social vulnerability index scores, cluster analysis was able to discriminate various social determinants of health subthemes among patients who resided in "average" vulnerability communities that stratified patients relative to risk of adverse postoperative events. (C) 2021 Elsevier Inc. All rights reserved.