The effect of electronic medical record application on the length of stay in a Chinese general hospital: a department- and disease-focused interrupted time-series study

Peng Yang, Yi Cao, Danhong Liu, Yuxiang Bai, Feng Pan, Yongyong Xu

A key purpose of electronic medical records (EMR) introduced in medical institutions is to improve work efficiency. The average length of stay (LOS) is just an important indicator to evaluate work efficiency of medical care in hospitals. Recently, there have been reports about effects of EMR application on LOS in medical institutions, but they have been mostly based on the overall analysis of a region or a hospital and not of specific clinical departments and diseases or based on longer time periods. Therefore, in this study, we selected four clinical departments and four diseases with the largest number of inpatients from January 2004 to December 2012 in a Chinese 3A general hospital and used an interrupted time-series method by the departments and diseases to analyze the relationship of EMR application and LOS. Through our analyses, we concluded that, under unadjusted condition, LOS were all reduced (P < 0.001) after EMR application in four departments and for four diseases. After adjustment by gender, age or admission condition, LOS still all decreased after EMR application (P < 0.05) regardless of departments or diseases. The trend changes in LOS reversed from increasing to decreasing in the orthopedics department (coefficient: 0.016 to -0.079), the cardiovascular surgery department (coefficient: 0.007 to -0.126) and all departments overall (coefficient: 0.004 to -0.070), as well as for the intervertebral disc disorders (coefficient: 0.026 to -0.068). Furthermore, the decreasing trend gained a larger slope in the cardiology department (coefficient: -0.017 to -0.023), the neurology department (coefficient: -0.012 to -0.043) and for the coronary heart disease (coefficient: -0.010 to -0.018), the ventricular septal defect (coefficient: -0.024 to -0.059), and the cerebral infarction (coefficient: -0.031 to -0.040). Together, these findings indicate that EMR application coincided with a decrease in LOS and may have a contribution to the decrease.