Association between hospital case volume and mortality in non-elderly pneumonia patients stratified by severity: a retrospective cohort study

Abstract

Background: The characteristics and aetiology of pneumonia in the non-elderly population is distinct from that in the elderly population. While a few studies have reported an inverse association between hospital case volume and clinical outcome in elderly pneumonia patients, the evidence is lacking in a younger population. In addition, the relationship between volume and outcome may be different in severe pneumonia cases than in mild cases. In this context, we tested two hypotheses: 1) non-elderly pneumonia patients treated at hospitals with larger case volume have better clinical outcome compared with those treated at lower case volume hospitals; 2) the volume-outcome relationship differs by the severity of the pneumonia. Methods: We conducted the study using the Japanese Diagnosis Procedure Combination database. Patients aged 18–64 years discharged from the participating hospitals between July to December 2010 were included. The hospitals were categorized into four groups (very-low, low, medium, high) based on volume quartiles. The association between hospital case volume and in-hospital mortality was evaluated using multivariate logistic regression with generalized estimating equations adjusting for pneumonia severity, patient demographics and comorbidity score, and hospital academic status. We further analyzed the relationship by modified A-DROP pneumonia severity score calculated using the four severity indices: dehydration, low oxygen saturation, orientation disturbance, and decreased systolic blood pressure. Results: We identified 8,293 cases of pneumonia at 896 hospitals across Japan, with 273 in-hospital deaths (3.3%). In the overall population, no significant association between hospital volume and in-hospital mortality was observed. However, when stratified by pneumonia severity score, higher hospital volume was associated with lower in-hospital mortality at the intermediate severity level (modified A-DROP score = 2) (odds ratio (OR) of very low vs. high: 2.70; 95% confidence interval (CI): 1.12–6.55, OR of low vs. high: 2.40; 95% CI:0.99–5.83). No significant association was observed for other severity strata. Conclusions: Hospital case volume was inversely associated with in-hospital mortality in non-elderly pneumonia patients with intermediate pneumonia severity. Our result suggests room for potential improvement in the quality of care in hospitals with lower volume, to improve treatment outcomes particularly in patients admitted with intermediate pneumonia severity.