Adoption of electronic health records by admitting physicians: A heuristic approach

Background: Although hospital based electronic health records (EHRs) are generally perceived to improve care, physician resistance may hinder EHR adoption. Purpose: This study uses constructs from diffusion of innovations and resource dependence theories to predict adoption and speed of adoption of an EHR by admitting physicians from three of ten hospitals in a highly integrated health system in Virginia. Functions evaluated: computerized physician order entry (CPOE), electronic history and physical (EH&P) and electronic discharge summary (EDS). The study tested hypotheses that adoption would be associated with: working at larger, academic hospitals; financial alignment; larger physician groups; office EHR; youth; males; medical specialty; high volume; hospital based; high inpatient ratio; and high loyalty. Methods: Administrative data collected for 326 physicians admitting at least ten patients during the six months following EHR activation represented over 80% of the total admissions. Logistic Regression and Cox Regression were used to evaluate how well variables predicted adoption (80% utilization) and adoption speed. Results: The Logistic Regression model predicted significant proportions of variation in adoption of CPOE (66%), EH&P (34%) and EDS (40%). CPOE adoption was more likely (p< .05) for physicians who were male, had a high inpatient ratio, lower patient volume and community hospital setting. EH&P and EDS adoption were more likely for physicians with alignment and large, academic hospital setting. The Cox Regression model predicted significant proportions of variation in speed of adoption of CPOE (10%), EH&P (14%) and EDS (19%). The overall model for CPOE was significant (p=.006); no individual predictors were significant. Physicians who were aligned or worked at the large, academic hospital adopted EH&P and EDS faster. Conclusion: Personal factors: loyalty, age and gender were generally not predictive. Organizational factors: hospital setting and financial alignment were most predictive of adoption. Study results may help administrators improve EHR installations.