July 2021. As electronic health record adoption has expanded, an early “advanced use digital divide” was found among critical access hospitals, which were systematically less likely to have adopted advanced uses of EHRs for both patient engagement and clinical data analytics. Our study updates the evidence on this gap through 2018, and shows that while CAHs have caught up with other hospitals in terms of overall EHR adoption, they still lag in patient engagement capabilities and have fallen further behind in clinical data analytics. Only 47% of CAHs had achieved advanced use in patient engagement by 2018, and only 32% had achieved advanced use in clinical data analytics. This gap prevents CAH patients from reaping the full benefits of a digitized health care delivery system, and policies should target support for CAH advanced use via technical resources, vendor partnerships, and standards.
March 2021. Only 37% of office-based physicians are able to send data electronically to outside providers to whom they refer patients. Within this group, little is known about the factors related to varying levels of use of these technologies. We investigated variation in the percentage of referrals that office-based providers sent with electronic care summary data. Providers with exchange capabilities only used these tools in about half of referrals, leaving millions of transitions of care without accompanying information for the consulting provider to review.
November 2020. Electronic health record log data is increasingly being used in health services and informatics research. Measures are often study-specific and not replicable across institutions or vendors, which limits the generalizability of findings. However, most major EHR vendors calculate use measures from log data and make those measures available to practice and hospital managers, often for identifying inefficiencies. We surveyed the landscape of vendor-provided EHR use measures for outpatient providers, and compared the existing measures to previously proposed measures. Use of these measures can improve standardization in research using these data for insights into clinician workload, administrative burden of EHRs, and provider burnout.
September 2020. Since its inception in 2017, the Merit-Based Incentive Payment System has reported high rates of physician attestation and resulting payment bonuses from Medicare. We analyzed detailed reporting data from the first year of MIPS to understand variation how providers achieved their final MIPS scores, finding that 46% of physicians skipped at least one of the three component quality reporting categories. Despite this, many of these physicians still received payment bonuses for 2017. The removal of incentives to participate across all three program categories could result in providers being less prepared for penalties set to go into effect in 2022.
Full study with Jordan Everson
August 2020. Adoption of electronic health records has increased physician documentation burden, raising concerns about physician burnout deriving from after-hours EHR use. We analyzed two years of longitudinal survey data of EHR users at a large health system, to compare physicians’ rates of reported off-hours use to other EHR users. 25% of physicians reported EHR use on days off, and 50% reported use in the evenings, compared to 5% and 41% of nurses, respectively. These findings suggest that the burden of after-hours EHR use falls primarily on physicians, however other EHR users also report after-hours and off-day EHR use.
June 2020. Integration of prescription drug monitoring program (PDMP) databases with electronic health records (EHRs) is an important tool in combating the opioid crisis in the US. We assessed nationwide hospital adoption of three EHR capabilities related to opioid prescribing and PDMP integration: ability to electronically prescribe controlled substances, checking of state PDMPs directly from EHRs, and integration of PDMP data into EHRs. Less than one third of hospitals reported the ability to check state PDMPs, and only 14% reported full PDMP data integration. More troublingly, hospitals in high opioid-prescribing counties were less likely to report these capabilities.