The association of hospital research publications and clinical quality.

ABSTRACT

OBJECTIVE: To assess the quantity and impact of research publications among US acute care hospitals; to identify hospital characteristics associated with publication volumes; and to estimate the independent association of bibliometric indicators with Hospital Compare quality measures. DATA SOURCES Hospital Compare; American Hospital Association Survey; Magnet Recognition Program; Science Citation Index Expanded. STUDY DESIGN: In cross-sectional studies using a 40% random sample of US Medicare-participating hospitals, we estimated associations of hospital characteristics with publication volumes and associations of hospital-linked bibliometric indicators with 19 Hospital Compare quality metrics. DATA COLLECTION/EXTRACTION METHODS: Using standardized search strategies, we identified all publications attributed to authors from these institutions from January 1, 2015 to December 31, 2016 and their subsequent citations through July 2020. PRINCIPAL FINDINGS: Only 647 of 1604 study hospitals (40.3%) had ≥1 publication. Council of Teaching Hospitals and Health Systems (COTH) hospitals had significantly more publications (average 599 vs. 11 for non-COTH teaching and 0.6 for nonteaching hospitals), and their publications were cited more frequently (average 22.6/publication) than those from non-COTH teaching (18.2 citations) or nonteaching hospitals (12.8 citations). In multivariable regression, teaching intensity, hospital beds, New England or Pacific region, and not-for-profit or government ownership were significant predictors of higher publication volumes; the percentage of Medicaid admissions was inversely associated. In multivariable linear regression, hospital publications were associated with significantly lower risk-adjusted mortality rates for acute myocardial infarction (coefficient -0.52, p = 0.01), heart failure (coefficient -0.74, p = 0.004), pneumonia (coefficient -1.02, p = 0.001), chronic obstructive pulmonary disease (coefficient -0.48, p = 0.005), and coronary artery bypass surgery (coefficient -0.73, p < 0.0001); higher overall Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) ratings (coefficient 2.37, p = 0.04); and greater patient willingness to recommend (coefficient 3.38, p = 0.01). CONCLUSIONS: A minority of US hospitals published in the biomedical literature. Publication quantity and impact indicators are independently associated with lower risk-adjusted mortality and higher HCAHPS scores.