Privacy, Network Effects and Electronic Medical Record Technology Adoption

An Electronic Medical Record (EMR) allows medical providers to store and transfer patient information using computers rather than paper records. The ability to transfer existing medical information about a patient to another hospital electronically could be beneficial or detrimental to hospitals. On one hand, the ability to transfer patient records electronically can improve patient safety and reduce the time medical staff spends on paperwork. On the other hand, creating transferable records could make it easier for patients to leave the hospital and seek treatment elsewhere. In both cases, one hospital’s EMR adoption decision will rest on the adoption decisions of other hospitals. If hospitals benefit from accessing electronic information about their patients from other hospitals, they will react positively to the adoption of their local competitor. If, on the other hand, hospitals are worried about losing patients to other hospitals, they may be less likely to adopt EMR if their local competitors have already adopted it. We empirically distinguish between these two competing theories by exploiting variation in state privacy laws governing the transfer of patient information. Our results indicate that there are positive network effects in the diffusion of EMR, and that when medical records can be transferred freely, the presence of other hospital adopters encourages adoption. We present empirical evidence, exploiting both time-series and cross-sectional variation, that the enactment of state privacy laws restricting the transfer of medical information from hospitals inhibits at least 25 percent of the network effects that would have otherwise promoted a hospital’s adoption of EMR. We also show that privacy laws affect hospital choices over compatibility of EMR software. The size of the network effects is identified by using characteristics of the other hospitals in the region as instruments for the installed base. We conclude that policymakers face sizable tradeoffs between offering strong privacy protection and promoting the network gains from EMR technology. ∗Economics Department, University of Virginia, Charlottesville, VA †MIT Sloan School of Business, MIT, Cambridge, MA. ‡We thank HIMSS for providing the data used in this study, and Pierre Azoulay, Pam Jia, Enrico Moretti, Ariel Pakes, Fiona Scott Morton and the participants of the MIT IO lunch, for helpful comments. All errors are our own.