Despite initial challenges, EHRs are being associated with reduced hospital mortality rates

Matured-system electronic health records (EHRs) are leading to decreased mortality rates in hospitals. Although all hospitals may not see equal benefits, smaller and nonteaching hospitals stand to benefit the most according to a recent study published in Health Affairs.

Further, these researchers found that hospital characteristics, including size and teaching status, may play an important role in the effects of EHR adoption on mortality rates.

Critics of EHRs deny that EHRs improve clinical care. These results, however, fly in the face of that. Researchers from the University of California, San Francisco, CA—who conducted the study—stressed that these findings highlight the importance of allowing hospitals and staffs to work with the technology and adopt new capabilities.

“In other industries, widespread digitization took a decade to realize improvements,” said senior author Julia Adler-Milstein, PhD, associate professor, Department of Medicine, Philip R. Lee Institute for Health Policy Studies. “It’s a major transformation of the health-care system to go from paper to digital. We are seeing those rewards, but it has taken time and work.”

Dr. Adler-Milstein and colleagues included data from 3,249 hospitals throughout the country and assessed 30-day mortality rates for 15 common conditions in patients aged ≥ 65 years. Data collection began in 2008, when national data were first collected about EHR adoption.

To hone in effectively on the impact of EHR adoption, they assessed three specific phases: baseline EHR functions, maturation of baseline functions, and adoptions of new EHR functions.

“Hospitals implement functionality over time, because it’s really hard to go from fully paper to fully electronic overnight,” said Dr. Adler-Milstein. “We measured EHR adoption in a way that was truer to the way adoption likely occurred. As hospitals added functionalities over time, there was benefit from each of those new features.”

Mean baseline number of basic EHR functions was 5.7, and the mean number of new functions each hospital adopted per year was 0.6. The average baseline risk-adjusted 30-day mortality rate was 13.46 deaths per 100 admissions.

The data showed that baseline adoption was associated with a 0.11 percentage point higher mortality rate per function, but the maturation of these baseline functions was associated with a 0.09 percentage point lower mortality rate per function per year. Finally, they found that the adoption of new EHR functions was associated with a 0.21 percentage point reduction in mortality rate per year per function.

These relationships were driven mainly by the effects in small and nonteaching hospitals, according to Dr. Adler-Milstein et al.

“The fact that small and nonteaching hospitals realized the majority of improvement in mortality rates from EHR adoption is an interesting and important finding,” wrote the authors.

Dr. Adler-Milstein and colleagues hypothesized that smaller and nonteaching hospitals got the most benefit because they, by nature, have fewer efforts to improve hospital quality than larger, teaching hospitals.

When analyzing the difference-in-differences in predicted annual changes in 30-day mortality, they found that average adopters performed significantly better than nonadopters, with 0.67 fewer deaths per 100 admissions. Upon analyzing predicted changes in mortality rates according to hospital characteristics, they found that average adopters performed significantly better than nonadopters if they were urban/suburban or rural, small or medium, nonteaching, and safety-net or nonsafety-net hospitals. For large hospitals and major/minor teaching hospitals, mortality rate differences for average adopters compared with nonadopters were not significant.

“Given our findings that small and nonteaching hospitals have worse mortality rates with a greater number of baseline EHR functions but then realize improvement from maturation and adoption of new functions, another explanation may be that these hospitals had limited resources and experience to support a high-quality initial EHR implementation. Therefore, they may have done a worse job with implementation of baseline functions but learned how to improve new implementation efforts over time,” they wrote.

“There’s been a lot of frustration with EHRs,” concluded Dr. Adler-Milstein. “Our study shows they are improving care. It just may not be as much as providers or policymakers wanted—or come as quickly, or as easily, as they would have liked.”

This study was funded by the John A. Hartford Foundation.