There’s been a study that’s been pandered around making the assertion that Physicians lose 48 minutes a day to EHR. This story in Medical Economics is just one example of many. A comment on that story from Dr. Rah describes generally my feelings about the study:
I find it disappointing that such drivel is even reported. #1. A 2012 survey! Data is > 2 yrs old. #2. 411 respondents is a very small N; hardly significant in that there are at least a million users now of EMRs. #3. You can do better–why report such meaningless info??
Of course, this only begins to describe the flaws in this study. First, they were just asking physicians for their perceived views on how long something took with the EHR as opposed to actual time. As humans, we’re really bad at judging the amount of time that’s passed. Not to mention that many of the respondents were trainees who had no history with which to compare. I could go on and on, but I’ll stop there.
I’m not arguing whether EHR saves doctors time or whether it takes more time. I’ve seen places where both sides of the coin have occurred. So, I think that you could write an article that EHR saves doctors time and another article that talks about how EHR takes more time. You can find both experiences out there. There are hundreds of factors at play that influence the answer to this question.
One thing I don’t think anyone would disagree with is that meaningful use has required a lot more time from doctors. So, when you layer on a new EHR with the meaningful use requirements, then you’re probably going to be spending more time documenting in the EHR. Although, is that the EHR’s fault or meaningful use?
It would be nice for someone to do a high quality study on EHRs and the time a doctor spends. However, when you think about the factors that could influence the time spent: EHR software, specialty, location, tech skill of doctor, meaningful use, not meaningful use, etc etc etc, you can see why we haven’t seen a proper study on the impact of EHR on efficiency. There are too many variations for which you’d have to test.