Inpatient Gastroenterology Consultation and Outcomes of Cirrhosis-Related Hospitalizations in Two Large National Cohorts

Background Little is known about use of specialty care among patients admitted with cirrhosis complications. Aims We sought to characterize the use and impact of gastroenterology/hepatology (GI/HEP) consultations in hospitalized patients with cirrhosis. We studied two national cohorts—the Veterans Affairs Costs and Outcomes in Liver Disease (VOCAL) and a nationally representative database of commercially insured patients (Optum Clinformatics™ DataMart). Methods Cirrhosis-related admissions were classified by ICD9/10 codes for ascites, hepatic encephalopathy, alcohol-associated hepatitis, spontaneous bacterial peritonitis, or infection related. We included 20,287/222,166 index admissions from VOCAL/Optum from 2010 to 2016. Propensity-matched analyses were conducted to balance clinical characteristics. Mortality and readmission were evaluated using competing risk regression (subhazard ratios, sHR), and length of stay (LOS) was assessed using negative binomial regression. Results GI/HEP consultations were completed among 37% and 42% patients in VOCAL and Optum, respectively. In propensity-matched analyses for VOCAL, GI/HEP consultation was associated with adjusted estimates of increased LOS (1.55 + 1.03 additional days), 90-day mortality (sHR 1.23, 95% CI 1.14–1.36), and lower 30-day readmissions (sHR 0.82, 95% CI 0.75–0.89). In Optum, inpatient consultation was associated with higher LOS (1.13 + 1.01 additional days), higher 90-day mortality (sHR 1.57, 95% CI 1.43–1.72), and higher 30-day readmission risk (sHR 1.04, 95% CI 1.02–1.05). Post-discharge primary and specialty care was higher among admissions receiving GI/HEP consultation in both cohorts. Conclusions Use of GI/HEP consultation for cirrhosis-related admissions was low. Patients who received consultation had higher disease severity, and consultation was not associated with lower mortality but was associated with lower 30-day readmissions in the VA cohort only.