Sepsis Inpatient Diagnostic Testing at Mid-Size Academic and Non-academic Health Systems

Abstract

Abstract Introduction/Objective Comparison of clinical laboratory test and radiology imaging utilization for sepsis inpatients across multiple academic and non-academic health systems would be useful for healthcare resource utilization benchmarking. Methods/Case Report We compared laboratory and imaging utilization for adult patients hospitalized for sepsis procedures at mid-size (250-450 hospital beds), academic (N=12; 31,353 hospitalizations) and non-academic (N=18; 54,192 hospitalizations), hospitals using the Vizient Clinical Data Base™ over 3 years (2019-2021). We used Medicare Severity Diagnosis Related Groups (MSDRG), used by the US Centers for Medicare and Medicaid Services (CMS), to identify hospitalizations for sepsis (MSDRG triplet 870/871/872). We stratified patients by severity comorbid conditions and complications into high, moderate, and low severity groups. We compared the mean number of clinical laboratory (CPT codes 80000-89999) and imaging studies (CPT codes 70000-79999) per hospitalization and per hospital day. We also measured aggregate clinical Resource Intensity Weight (RIW) per hospitalization to quantify diagnostic resource consumption for each encounter. The RIW is based on the US Centers for Medicare and Medicaid Services (CMS) Ambulatory Payment Classification (APC) weights. Results (if a Case Study enter NA) Mean laboratory tests were significantly greater (p<.01) at academic hospitals per encounter (152.2 vs. 95.4 tests, 59.5% higher) and per hospital day (14.5 vs. 9.9 tests, 46.5% higher). Laboratory RIW was also higher per hospitalization (21.0 vs. 13.5 RIW, 55.6% higher) and per hospital day (2.0 vs. 1.4 RIW, 42.9% higher) at academic hospitals. Mean imaging studies at academic hospitals were higher per case (7.8 vs. 7.2 studies, 8.3% higher) but not different per day (0.80 for both hospital groups). Imaging RIW was not significantly different between academic and non-academic hospitals either by encounter (11.1 vs. 9.9, 12.1% higher) or per day (1.2 for both hospital groups). Imaging comparisons lacked statistical significance. Analysis of utilization after stratification for severity levels revealed a similar pattern with greater differences in the high severity group. Conclusion Mean laboratory tests and laboratory RIW per hospitalization are greater at academic hospitals. These differences persist after adjustment for hospital length of stay and severity level. Differences in imaging studies and imaging RIW were not significant.