| Description |
Although Electronic Health Record (EHR) systems have recently achieved widespread adoption in the U.S., our understanding of their impact on care outcomes is still limited. Current literature has produced mixed results due to the use of non-standardized measurements and weak research designs. In this dissertation, 4 studies are conducted to develop a systematic methodology for detecting near real-time performance changes during EHR implementations. It also explores factors that can affect outcomes during a commercial EHR implementation. The first study assesses the current state of the literature on health IT adoption to identify the most commonly reported outcome measures and proposes a taxonomy to classify these measurements. The second study expands the first study by identifying additional measures through semistructured interviews with experienced clinical and administrative leaders from a large care delivery system. We also collect input from national informatics experts who suggested additional relevant measures. The third study is a robust longitudinal analysis including several measures from our larger inventory that were used for monitoring a large-scale commercial EHR implementation and detected patterns of impact and mixed time-sensitive effects across geographically dispersed settings from an integrated care delivery system. The fourth study is a qualitative analysis guided by the quantitative results of the third study. We identified several factors that may have contributed to performance changes detected by our methodology. In summary, this dissertation will help the broader medical and informatics communities by informing what and how to continuously monitor future similar implementations. First, it contributes to the identification of relevant outcomes likely impacted by health IT interventions. Second, it combines these outcome measures with a robust interrupted time-series design, producing a systematic methodology that allows earlier and potentially more precise detection of unexpected effects, and implementation of effective response to mitigate negative impacts. Last, the identification of factors that may impact outcomes during and following an EHR implementation and covariates to measure them will empower researchers in charge of future evaluations, hopefully increasing the understanding of the full impact of health IT interventions. |
