It’s hard to argue that in the wake of an EMR install, some processes are likely to slow down or even break. What makes the following study interesting is that it attempts to do something intriguing — sorting out how EMRs affect emergency department throughput they’re implemented.
A new study appearing in the Journal of the American Medical Informatics Association concludes that EMR installs have a distinct impact on ED processes, as well as patient length of stay.
Researchers with the Cincinnati Children’s Hospital Medical Center attempted to track the impact an EMR install had there by looking at how often non-acute patients got routed to alternate sites. Specifically, they looked at how often potential victims of the H1N1 flu virus were routed to non-acute sites before and after the Center did its implementation.
The hospital phased its EMR rollout in over two years, setting the ED section of the rollout for November 2009, a date which overlapped with the peak of t he H1N1 outbreak in its region. During that period, the hospital set up an overflow clinic — staffed by non-ED providers — to deal with patients who had flu-like illnesses.
The overflow clinic began seeing 50 to 60 patients per day, 10 to 20 percent of the ED’s volume, within two weeks, but plunged to pre-surge levels by November 2009, researchers reported in JAMIA. While 10 percent or more of patients were being diverted prior to the ED rollout, that total fell to 5 percent afterwards, the study concluded.
Another intriguing finding was that length of stay in the ED went up markedly during the implementation. LOS in the emergency department was 24 to 53 minutes for admissions, and 9 to 19 minutes for discharges prior to the EMR rollout; during EMR implementation, LOS for both groups was greater than the pre-overflow clinic block and the H1N1 overflow clinic block by 32 to 62 minutes for admits and 35 to 44 minutes for discharges. If reproducible, those are some serious numbers.
Of course, there’s a long ton of confounding factors here, including but not limited to the fact that patient flow in the H1N1 outbreak may be significantly different in important ways than the standard patient population. For all I know, the H1N1 diversions were not too hard to identify, which could mean that the EMR would have had a worse impact if the virus wasn’t raging.
That being said, the question it asks — what impact the EMR rollout has on the “front door to the hospital” — is one that deserves more attention. Rest assured that if I get more data on this subject I’ll be reporting it here.