Two new studies of electronic medical records (EMRs) have come to different conclusions: one found that they don’t save money or staffing time, but the other suggested that at least one part of EMR — electronic prescribing — does save lives.
In the electronic prescribing study, at a single California hospital, the mean monthly adjusted mortality rate decreased by 20% after implementation of a computerized physician order entry (CPOE) system.
The unadjusted death rate declined from 1.008 to 0.716 deaths per 100 discharges per month (95% CI 0.8% to 40%; P=0.03), according to an article prepared for the July issue of Pediatrics.
“With observed versus expected mortality-rate estimates, these data suggest that our CPOE implementation could have resulted in 36 fewer deaths over the 18-month postimplementation time frame,” Christopher A. Longhurst, MD, of Lucile Packard Children’s Hospital at Stanford University in Palo Alto, and colleagues reported.
“The potential implications of these findings on national mortality statistics in children are dramatic,” they wrote. “Future research should focus on replicating these findings in other inpatient settings and populations, and evaluating the cost-effectiveness of this intervention.”
To determine the effect of CPOE implementation on the hospital-wide mortality rate, the researchers conducted a cohort study using historical controls at the main campus of the Lucile Packard Children’s Hospital.
At the time of the study, the hospital had 271 beds, including 193 on the main campus.
Patients were included if they were admitted between Jan. 1, 2001, and April 30, 2009. Obstetrics patients were excluded from analysis.
The hospital began implementing EMR in 2004, starting with nonmedication, unit-clerk-based order entry. The second phase, activated in late 2007, included CPOE with clinical decision support and comprehensive nursing and support service documentation.
The pediatric intensive care unit implemented the CPOE system about 10 months after the other units; as of the study’s end, only the cardiovascular intensive care unit had not yet been activated. The primary outcome measure was the hospital-wide mortality rate, excluding obstetrics.
The researchers found no significant difference between pre-CPOE and post-CPOE populations on the basis of patient gender (47.4% females versus 48.0% females; P=0.34).
Statistically significant differences were noted for age, severity of illness as represented by case-mix index, and race/ethnicity, with fewer individuals identified as white or of unknown ethnicity and more Hispanic and Asian individuals in the postintervention group.
There are several possible reasons for the decline in mortality following the introduction of CPOE, the authors noted.
Standardization of patient orders, which may have resulted in better communication with staff
Making orders, vital-sign documentation, and medication-administration information remotely accessible in real time, described as “a fundamental shift in basic care processes”
Introduction of CPOE itself, which eliminated the redundant transcription of medication orders by pharmacists, thus improving accuracy and decreasing turnaround times
The authors noted several limitations to their study.
First, they wrote, “it is possible that the reduced mortality rate was simply the result of differences in the preintervention and postintervention populations and is independent of the CPOE intervention.”
Also, they noted, the results cannot necessarily be generalized to other hospitals.
In the economics study, EMR implementation was associated with a 6% to 10% higher cost per discharge in a hospital’s medical-surgical acute units, and it increased registered nurse hours per patient day by 15% to 26%, according to Michael F. Furukawa, PhD, of Arizona State University, and colleagues.
“The results imply that EMR may increase the demand for skilled nurses, which could have implications for nurse labor markets,” they wrote online in the journal HSR: Health Services Research.
“Contrary to expectation, we found little support for the proposition that EMR generates significant cost savings to hospitals through reductions in length of stay and the demand for nurses.”
While interest in EMR adoption is high, little is known about the impacts of the EMR on nurse staffing and patient outcomes in community hospital settings, the authors noted.
To find out if instituting an EMR could reduce length of stay (LOS) and improve efficiency, the authors examined data from the 1998-2007 HIMSS Analytics Databases. The sample included medical-surgical acute units within short-term, general acute care hospitals in California.
Federal government, specialty, children’s, and long-term acute hospitals were excluded, as were hospitals with incomplete financial reports. The final data set included 326 hospitals and comprised 2,828 hospital-year observations.
The authors found that EMR implementation increased significantly from 1998 to 2007. In 1998, only 33.9% of hospitals had at least started implementation of EMR, compared with 80.8% by 2007.
Although EMRs increased costs per discharge across several stages of implementation, costs varied by stage, the researchers found. Later stages of implementation were associated with 5.9% to 10.3% higher cost per discharge, with the cost increases attributable to both higher cost per patient day and higher LOS.
Increases in LOS also varied depending on the state of implementation. Early-stage implementation increased LOS by 2.1%, compared with later-stage implementation, where LOS rose by 3.7% to 4.4%.
All three stages of EMR implementation increased nurse staffing levels, the study showed.
Total nursing hours increased 13.3% to 14.6% in early-stage EMR implementation, compared with 11.2% to 21.6% during the middle stage and 16.0% to 19.4% during the late stage.
The increase in total nursing hours per patient day was attributed to high staffing levels for both RNs and aides, they noted.
The researchers offered several explanations for their findings. The fact that EMR did not decrease LOS suggests that per-diem payments may work against such reductions.
And the association of EMR with increased staffing may also reflect the unintended consequences of poor implementation or cultural resistance to change, they wrote.
The authors noted several limitations of their study, including possible bias from measurement error, quality improvement initiatives that might impact staffing or patient outcomes, and California’s minimum nurse staffing regulation, which mandated staffing ratios for medical-surgical acute units in 2004 and 2005.
In addition, the database is self-reported and primarily used for market research, which might lead to overestimates of actual rates of EMR implementation.
Joyce Frieden is a MedPage Today News Editor.