Blog: Differing EHR definitions mean data headaches
A reader was confused by an article we ran saying 48% of doctors were e-prescribing using an electronic health-record system. Another article, published Dec. 6, based on an annual survey of ambulatory care physicians by the National Center for Health Statistics at the Centers for Disease Control, placed "basic" EHR adoption at 39.6%. A basic EHR includes e-prescribing capabilities. That seemed like an 8-point spread in physician EHR use.
To make matters worse, the NCHS surveyors calculated two other EHR penetration rates (PDF) for ambulatory care physicians of 71.8% and 23.5%, while Dr. Farzad Mostashari, head of HHS' Office of the National Coordinator for Health Information Technology, quoted in a third article, said 31% of physicians and other eligible professionals have been paid federal EHR incentive money, which means they'd also have to be e-prescribers.
So, that leaves us with quite a range of physician EHR users—71.8% to 23.5%— and several rates in between.
Why so many?
Here's some background and an explanation.
One of Dr. David Blumenthal's great, unheralded contributions to healthcare IT was coming up with common criteria for determining what constitutes a useful EHR. Before Blumenthal, a host of healthcare organizations, Modern Healthcare included, tried to measure EHR adoption using surveys with diverse definitions.
Those efforts yielded scattershot results because surveyors either didn't define EHRs or described them so loosely that variability made the results meaningless.
In an attempt to set a baseline for President George W. Bush's federal EHR promotional campaign, Dr. David Brailer, Bush's first ONC chief, contracted in 2005 with the then Blumenthal-led Institute for Health Policy at Massachusetts General Hospital to define an EHR, design a statistically valid national EHR survey and carry it out (with some help from the Harvard School of Public Health, George Washington University and RTI International).
Blumenthal's people came up with a list of features for an effective EHR. One was e-prescribing.
But when the ONC-funded survey was finally conducted in 2007, it was already three years into the federal health IT campaign. When it turned out that just 4% of physicians in ambulatory care had what Blumenthal's group defined as a "fully functional" EHR, Blumenthal's coalition decided to grade the country on a curve. In their 2008 public report card, they added that 13% of physicians had met a subset of their EHR criteria, and voila, the "basic" EHR was born.
Meanwhile, Blumenthal's group shared its work with the National Center for Health Statistics. The NCHS had been conducting its National Ambulatory Medical Care Survey of office-based physicians, and since 2001, had included a simple question on EHRs.
In the beginning, the NCHS surveyors asked docs only whether they used an EMR system. The NCHS defines "any EMR/EHR system," as a medical or health record system that is all or partly electronic and is not a billing system. Then, in 2006, the NCHS expanded its survey, adding criteria for both a "basic" and a "fully functional" EHR.
That year, the NCHS's surveyors had even worse luck in conjuring up "fully functional" EHR users than would Blumenthal's people a year later. In 2006, just 3.1% of physicians reported to the NCHS they had a fully functional system and only 10.5% had a "basic" system, but 29.2% replied yes when asked whether they had "any" EHR—hopeful news, if you weren't persnickety about what "any" EHR could do.
By the way, the 48% EHR/e-prescribing rate reported by Mostashari was based on data coming from yet another source—the privately owned Surescripts e-prescribing network.
I hope that helps explains why there are so many EHR rates out there. Once we have definitions, we can't let them go.
Follow Joseph Conn on Twitter: @MHJConn.