Electronic records, medical errors -- and the inescapable human factor


This article was originally on a blog post platform and may be missing photos, graphics or links. See About archive blog posts.

We all know that medical errors happen -- and most of us have read that electronic medical records will help deal with that problem.

But electronic record-keeping doesn’t remove the human factor. A study published in the Archives of Internal Medicine, has found that electronically alerting doctors of suspicious test results doesn’t mean doctors will a) open the electronically delivered alert or b) act on it if they do.


And though you’d think that alerting more than one physician would reduce the chances that a patient would slip through the cracks, the study found just the opposite: Suspicious test results were less likely to be acted on, not more, presumably because each physician assumed the other doctor had taken the necessary steps.

The study, by Dr. Hardeep Singh of the Veterans Affairs Medical Center in Houston and co-workers, was conducted in a Veterans Affairs outpatient facility from November 2007 to June 2008. The VA has a fairly sophisticated electronic medical records system. The study focused on reports of imaging exams -- CT scans, MRIs, mammograms, sonograms and radiograms.

Under the system, when a result is abnormal and needs follow-up, an ‘alert’ window comes up on the doctor’s computer screen. It stays there for two weeks. Then it goes away.

During the time period studied, there were 123,638 imaging studies. Of those, there were 1,196 alerts indicating something was potentially abnormal.

Of the alerts, 217 (18.1%) were unopened after two weeks.

And of the 1,196 alerts, 92 (7.7%) didn’t receive timely follow-up, such as a call to a patient or ordering of more tests. (The definition of timely follow-up was within four weeks.)

Perhaps surprisingly, the rates of poor follow-up were about the same for reports that were unopened and reports that were opened.

But in cases when a radiologist actually got on the phone and talked to a physician about a test, follow-up was more likely. (This may have partly been due to those results being more serious than other abnormal tests, the authors said.)


The study doesn’t say how this electronic system compares to ones where it’s all paperwork and word of mouth: surprisingly, there aren’t good comparative data to be had, the authors note. But it does have a few suggestions about what it found:

1) Something needs to be done about the information overload doctors experience. If you’re peppered with alerts from left, right and center, you’re more likely to ignore or miss some.

2) There has to be clear understanding about who, if more than one doctor is notified, is the responsible party for taking the next step.

3) Alerts that haven’t been opened should stay up on doctors’ screens for longer, ‘perhaps even indefinitely, and should require the healthcare provider’s signature and statement of action before they are allowed to drop off from the screen.’

4) The technology should do more than track whether doctors have opened alerts. They should track whether follow-up action was taken.

-- Rosie Mestel