There’s a specialty in medicine where health IT has already revolutionized their practice. In this specialty, successful, profitable software companies sell generally appreciated products that increase practice revenue and shorten their workdays. What’s the specialty? Radiology.
I’ve worked with a couple of PACS (picture archive and communications systems) and RIS (radiology information systems) vendors and people selling associated products and seen their financials. These guys offer inter-operable systems that allow digital radiology systems (x-ray, CT, MRI, etc) installed in hospitals to share the images with radiologists working anywhere. The radiologists get easy access to systems, all the utilities they need, and reporting templates. Many of the systems include workflow utilities that push radiologists to finish cases on a schedule, resulting in shorter work days or more revenue.
Contrast this with the typical medical records system. Research I’ve seen suggest that for first 18 months, most clinicians are less productive with EMRs. There have been articles published in scientific journals detailing inefficiency, risks, and even increased neonatal mortality with a commercial, off-the-shelf CPOE system. In two EMR deployments I’ve been involved with, the consensus was that each added an hour a day to the physician’s work.
Why is this? Most HIT systems are designed around the abnormal situation: preventing errors, streamlining emergency access to data, making medical information available if the patient is hospitalized (or re-admitted). Many of the docs I work with now resent their hospital’s EMR system. Some reject it, sticking with paper, others accept it but gain no benefit. Some physicians accept and even like their EMR, but they gain no workflow or efficiency benefit. When I ask a doc to tell me how many patients with a specific diagnosis they have, usually they can only give me a count of the primary diagnoses. And for all the work on interoperability, many find out that one of their patients is hospitalized from the patient’s family — if they find out at all.
There’s a parallel for this that is illustrative. In the 1980’s American and Japanese automobile manufacturers each experimented with robots. Americans viewed robots as superior to humans, and designed robots that did incredibly complicated things, but slowly. Japanese viewed robots as faster than humans, and designed robots that did simple things quickly. (Really, the difference was that American’s made robots whose motion was so complex that it had to be modeled using quasi-static assumptions, thus slowly, whereas the Japanese planned their robots’ simple motions using full dynamic models.) By using robots for the routine, hum-drum tasks of automobile manufacture, the Japanese workers were more engaged with high-level, thinking tasks, while their American counterparts found the most interesting tasks were being done, slowly, by robots.
We need to better design our health care systems to optimize the routine parts of healthcare rather than targeting the uncommon risks. With PACS, we’ve cut everything out of the job of radiologists except doing radiology. For most EMR users, there’s been no similar benefit. We need to change this. Physicians don’t want a system that warns them of every minor (and well-known) potential risk (we’ve all head of “alert fatigue”) but they do want a system that helps them do their high-level work better, faster, or more profitably.