Are There Heterogeneous Effects of Electronic Medical Records Adoption on Patient Health Outcomes?

The increasing availability of electronic medical records (EMR) has generated substantial optimism concerning possible consequent improvements in productivity, costs, and quality within the healthcare sector.  This optimism has proven substantial enough to spur U.S. policy to create incentives for adoption through the 2009 HITECH Act.  The recent attention on the possible impacts of EMRs has also spurred a great deal of research aiming to identify whether tangible improvements actually have taken place.  Thus far, the findings largely have been mixed. 

This paper aims to assess how EMR adoption affects patient health outcomes. It extends prior work in two important ways.  First, unlike the existing literature that typically focuses on patient mortality, we utilize patient safety indicators (PSIs) to measure outcomes. PSIs, developed by the Agency for Healthcare Research and Quality (AHRQ), are intended to measure in-hospital complications and adverse events, health outcomes that are less severe than mortality, but still very important for patient well-being.

Second, we allow for the possibility that the impact of EMR adoption may differ between seniors and non-seniors. Prior studies, mainly focused on the senior population, have found, at best, a modest impact of EMR adoption on health outcomes. However, there are reasons to believe EMR adoption might have differential impact on seniors and non-seniors. On one hand, EMRs may have more impact on higher severity patients and may therefore be less important for the care of non-seniors. Alternatively, aspects of EMRs that provide the clinician with decision support such as reminders and treatment protocols may have more of an impact for more straightforward cases rather than cases with multiple interacting comorbidities. EMRs may also have a larger effect on a younger population due to differing physician/patient dynamics. For example, existing evidence has shown that duration of a patient’s relationship with his/her physician is increasing with age, potentially dampening the information value of EMRs for older patients.

We merge data on hospital EMR adoption from the Health Information and Management Systems Society (HIMSS) Analytics Database with the Nationwide Inpatient Sample (NIS), a nationally representative database of hospital inpatient discharges collected by the AHRQ’s Healthcare Cost and Utilization Project (HCUP). We adopt a fixed effects Poisson model taking into consideration that our variables of interest are discrete in nature (the number of patients experiencing an adverse event). The fixed effects model controls for the potential bias caused by unobserved heterogeneity and allows us to exploit within hospital variation in EMR adoption. We use this model to estimate the overall effect of EMR adoption on PSIs, and differential effects for seniors and non-seniors.

Our results suggest that the adoption of EMRs significantly improves patient safety by lowering the occurrence of adverse events. More interestingly, we find that EMRs leader to larger improvements of care for the non-senior population, consistent with the hypothesis that the adoption of EMRs may have a larger effect on the younger population. Additional analysis explores the mechanism behind these differential effects.