Hospitalizations During Systemic Therapy for Metastatic Lung Cancer

IMPORTANCE: Understanding the risk of hospitalization due to treatment-related toxic effects is essential for patients, their clinicians, and health systems. Unplanned hospitalizations represent potential gaps in patient care; definition of these gaps allows characterization and identification of areas for quality improvement. OBJECTIVE: To compare the rates of hospitalization in patients with metastatic non-small-cell lung cancer (mNSCLC) receiving chemotherapy in the "real world" vs clinical trial settings and to identify factors associated with hospitalization. EVIDENCE REVIEW: A systematic review of Medline and EMBASE was conducted for records dating from database inception (1946 and 1974, respectively) through December 2014 to identify articles reporting rates of hospitalization during chemotherapy in patients with cancer. Both observational studies and clinical trials were eligible. This report focuses on patients with mNSCLC receiving chemotherapy because data were available for this clinical scenario in both the clinical trial and observational setting, allowing comparison. Summary statistics were used to describe results, and the χ2 test was used to compare hospitalization rates. FINDINGS: Of the 74 articles reporting hospitalization rates during chemotherapy, 10 studies, all published after 2004, examined chemotherapy in mNSCLC, 5 randomized clinical trials (3962 patients) and 5 observational studies (8624 patients). Chemotherapy regimens included doublet therapy, single-agent therapy, or chemotherapy type unspecified. The real world cohort was older (71 vs 63 years). All real world studies reported on comorbidities, while clinical trials reported performance status. The aggregate hospitalization rate among real world patients was significantly higher than among trial patients (51% vs 16%) (odds ratio, 7.7; 95% CI, 7.0-8.4; P < .001). Performance status and type of chemotherapy were associated with hospitalization during chemotherapy in clinical trials, while type of chemotherapy was a risk factor in observational studies. CONCLUSIONS AND RELEVANCE: Clinical trials in mNSCLC consistently report significantly lower rates of hospitalization than reports of real world cohorts of patients undergoing similar therapies. However, very few clinical trials report hospitalization information.