Depressive Symptoms and Short-term Outcomes after Medical Hospitalization

This thesis consists of 2 closely linked projects that evaluate the importance of depressive symptoms (hereafter, for brevity, referred to as “depression”) for short-term prognosis after general medical hospitalizations. Project 1 systematically reviews the literature to examine whether depression identifies increased risk of short-term adverse events (such as the composite endpoint of death or readmission within 30-days) post-discharge. Using prior literature and potential confounders identified in Project 1, Project 2 uses primary data collection to overcome limitations of previous studies and evaluates the prognostic utility and independent association between depression as defined by the PHQ-9 and the carefully adjusted risk of short-term adverse events post-discharge from general internal medicine (GIM) wards. In Project 1, 18 papers were reviewed, including published and unpublished data retrieved for up to 3397 patients with 30-day follow-up. It was determined that depression affected one-third of patients hospitalized for an acute medical reason (median 32%) and predicted both higher rates of readmission (20% vs. 14%, RR 1.73, 95%CI 1.16-2.58, n=2433) and mortality (3% vs. 2%, RR 2.13, 95%CI 1.31-3.44, n=3397) within 30-days post-discharge, with similar results observed at 90-days; that relatively few valid studies have examined short-term outcomes, and that most prior studies are limited to disease-specific patient populations (vs. unselected general medical cohorts) and lack repeated depression measures post-discharge. In Project 2, it was found that depression was common (26% of inpatients), largely unrecognized, often persisted for months post-discharge, and independently predicted an adjusted 2-fold increased risk of short-term death or readmission post-discharge over and above the best available risk-prediction tools. The findings from my thesis support the concept that clinicians should systematically screen patients for depression before discharge using validated tools and highlights two future directions for research: (1) designing post-discharge interventions that target depressed patients and (2) determining if treating depression reduces rates of short-term death or readmission in these patients.