Influence of Insurance Status on Disease Severity and Risk Measures in Acute Care Surgery

Author(s):
Isaac Sears; Chibueze Nwaiwu; Andrew Stephen; Daithi Heffernan

Background:

Quantitative measures such as the Charlson Comorbidity Index (CCI) and All Patient Refined Data Reference Groups (APR-DRGs) are commonly used to stratify patients according to the severity of their disease for the purpose of determining risk of serious complications based on documented comorbidities. However, it is more likely that increasing numbers of comorbidities may imply better access to healthcare to be able to have diagnoses made rather than merely degrees of illness. We hypothesized that among patients requiring acute care surgery, insurance status, as a proxy for healthcare access, would be associated with CCI score and APR-DRGs independent of the actual risk of complications

Methods:

Patients 18–65 years with acute appendicitis, cholecystitis, or perforated diverticulitis from 2010-2019 were identified in the National Inpatient Sample. We identified two subgroups based on insurance status: a “private insurance” (insured) group and a “self-pay” (uninsured) group. We compared the CCI score, APR-DRG severity, and APR-DRG risk of mortality between groups after adjusting for confounders. We computed adjusted odds ratios (ORs) for the following complications: in-hospital surgical site infection, in-hospital mortality, prolonged length of stay (hospital stay >90th percentile for each condition), and need for re-operation.

Results:

Overall, 232,410 admissions were included; 189,037 were insured, and 43,373 were uninsured. Compared to the uninsured group, the insured group had higher demographic-adjusted average CCI score (0.31 vs. 0.19, p < 0.001) and APR-DRG severity (1.52 vs. 1.47, p < 0.001). APR-DRG risk of mortality was higher in the insured group though not significantly different (1.11 vs. 1.09, p = 0.57). Conversely, the insured group had lower adjusted odds of in-hospital surgical site infections (OR 0.80, 95% CI 0.68-0.95, p = 0.012), in-hospital mortality (OR 0.56, 95% CI 0.36-0.88, p = 0.012) and prolonged length of stay (OR 0.79, 95% CI 0.76-0.82, p < 0.001). The adjusted OR for reoperation was not significant (OR 1.00, 95% CI 0.97-1.02, p = 0.72).

Conclusions:

Although the CCI scores and APR-DRGs implied that the insured group was sicker, the uninsured group suffered more complications. This phenomenon likely reflects better healthcare access with comorbidity identification and documentation among insured patients. Uninsured patients may harbor hidden, undocumented comorbidities that impact care of patients with surgical emergencies.