The Association of Attending Physicians' Publications and Patients' Readmission Rates : Evidence from Tertiary Hospitals in China Using a Retrospective Data Analysis

Background: Physicians play a unique role in scientific and clinical research, which is the cornerstone of evidence-based medical practice. In China, tertiary public hospitals link promotions and bonuses with publications. However, the weight placed on research in the clinician's evaluation process and its potential impact on clinical practice have come under controversy. Despite the heated debate about physicians' role in research, there is little empirical evidence about the relationship between physicians' publications and their clinical outcomes. Method: This paper examines the association of the quantity and quality of tertiary hospitals' attending physicians' publications and inpatient readmission rates in China. We analyzed a 20% random sample of inpatient data from the Urban Employee Basic Medical Health Insurance scheme in one of the largest cities in China from January 2018 through October 2019. We assessed the relationship between the quantity and impact factor of physicians' publications and 30-day inpatient readmission rates using logistic regression. There were 111,965 hospitalizations treated by 5794 physicians in our sample. Results: Having any first-author publications was not associated with the rate of readmission. Among internists, having clinical studies published in journals with an average impact factor of 3 or above was associated with lower readmission rates (OR = 0.849; 95% CI (0.740, 0.975)), but having basic science studies published in journals with an average impact factor of 3 or above was not associated with the rate of readmission. Among surgeons, having clinical studies published in journals with an average impact factor of 3 or above was likewise associated with lower readmission rates (OR = 0.708 (0.531, 0.946)), but having basic science studies published in journals with an average impact factor of 3 or above was associated with higher readmission rates (OR = 1.230 (1.051, 1.439)).