The earliest data record and processing systems were introduced to the American healthcare industry in the mid 1960s. Their primary purpose was to assist in clinical information management.

Around the same time, physician and researcher Dr. Larry Weed brought about the idea of accumulating patients’ details from various sources (i.e., physician notes, medication history, x-rays etc.), reorganizing and documenting them digitally. He went on to become the creator of the Problem-Oriented Medical Record, the predecessor of present-day electronic health records (EHRs).

The first EHRs was developed by the Regenstrief Institute in 1972 but it was not well received due to its cost. It was not until personal computers became more affordable and the emergence of the internet in the 1990s that set the stage for web based EHRs and gradually eliminate hand written notes and manual filing. Today, EHRs is pretty much adopted everywhere.

EHRs that may kill

The landmark report released by the Institute of Medicine (IOM) in year 1999 found that medical errors contributed to an average of a million in patient injuries and 98,000 fatalities in a year. As such, medical professionals have regarded EHRs as a hope to reduce these numbers.

However, according to scientists at the University of Utah Health, Harvard University and Brigham and Women’s Hospital, in spite of the prevalence and continuous improvement over the years, some EHR systems are still unable to detect damaging drug interactions and other medical errors that could potentially injure or kill patients. The conclusion was drawn after testing several simulated medical records.

What the researchers did was to use scenarios generated by Leapfrog CPOE EHR test, an EHR safety assessment tool, which replicates actual drug orders that may have adverse effects on patients and feed into EHR systems at more than 2000 hospitals in the US to examine how these systems respond. All tests were conducted over a 10-year span from year 2009 to 2018.

For example, in one of the simulated scenarios, a 52-year-old woman with pneumonia was admitted into the hospital. Prior to the episode, she has been taking a blood-thinning medication called warfarin to prevent blood clot from forming in her veins (i.e., Deep vein thrombosis; DVT).

Less than 100% error detection rate

The patient carries on with her warfarin intake three times a day after admission. The large dosage of warfarin was left undetected by the hospital’s EHR system and eventually the patients passed away as a result of hemorrhage. Researchers found that in year 2009, these EHR systems accurately issued warning at 54% of the time and by 2018, they were able to detect errors 66% of the time.

Researchers believe medical institutions have to tailor EHRs to meet their demands and needs because a drug reaction may trigger EHRs warning at one hospital but not the other. Nevertheless, federal regulators only check on EHR systems with factory specifications, but not on any of the alteration that respective hospitals made after installation. Besides, most hospitals do not participate in safety evaluations similar to the ones carried out by the study. This means keeping these systems continuously in check will be a big challenge.

“EHRs are supposed to ensure safe use of medication in hospitals but they’re not doing that. In any other industry, this degree of software failure wouldn’t be tolerated. You would never get on an airplane, for instance, if an airline could only promise it could get you to your destination safely two-thirds of the time” says Dr. David C. Classen, Professor of Internal Medicine at University of Utah Health and the study’s corresponding author.

All findings were published in the May 29 issue of JAMA Open Network.


Author Bio

Hazel Tang A science writer with data background and an interest in the current affair, culture, and arts; a no-med from an (almost) all-med family. Follow on Twitter.