On March 18th, 2019 Fred Schulte and Erika Fry published the article “Death by 1,000 Clicks” on Kaiser Health News (KHN), a collaboration between KHN and Fortune Magazine. The article reviewed the findings of a three-month long investigation into the United States’ implementation of electronic health systems. Their key take-aways are summarized below.

How much time does a doctor spend each day clicking through EMR records?

Read the whole article at khn.org/news/death-by-a-thousand-clicks/

Part 1 – Patient Harm

The most troubling consequence of a troubled system is, of course, patient harm. While the adoption of EHR systems may have alleviated certain risks, such as mis-reading a doctor’s handwriting, it has resulted in many cases of misdiagnosis, improper medication, or mishandling of a case due to how the EHR functions.

One example was the case of Fabian Ronisky, who in 2015 was admitted to a hospital in Santa Monica. A doctor suspected meningitis and typed in an order for a critical lab test. However, the test never made it to the lab, and a proper diagnosis was delayed by days. Ronisky ended up suffering irreversible brain damage, which he alleges could have been minimized had the EHR functioned properly.

The EHR in question was Epic, which had been installed at the Santa Monica hospital four month earlier. Although the doctor who had ordered the test could see it on Epic’s screen in Ronisky’s chart, Epic didn’t fully interface with the lab’s software, and so the ordered test simply never went anywhere. In the lawsuit, Epic denied liability or defects, saying that the doctor had been at fault for not pushing the right buttons, and that the hospital had set up the interface between the EHR and the lab software.

The lack of proper interfacing has been found to cause many incidents of harm, including giving a patient a medication they are allergic to, or reading a test result that has been returned but is not actually complete. A whistleblower case against eClinicalWorks cited many problems with the medication lists, such as prescribed drugs not showing up while discontinued drugs were shown as current. One patient’s medication profile could be accompanied by a physician’s note for a different patient. And nearly 30,000 prescriptions lacked stop and start dates in the system.

The private health care analytics firm Quantros said it logged 18,000 EHR-related safety events from 2007 to 2018. 3 percent of those events resulted in patient harm, including seven deaths. A 2016 study by The Leapfrog Group, a patient- safety watchdog, found that the medication-ordering function in hospital EHRs failed to flag potentially harmful drug orders in 39% of cases in a test simulation. 13% of those mistakes could have been fatal if they occurred with a patient.

Several other studies and cases cited examples of EHR software failing to correctly match records to patients, failing to fire critical medication warnings, or burying the important information under a sea of functionally useless alerts, warnings, and pop-ups. While errors in in the software’s basic code can cause direct patient harm, the lack of usability can also lead to mistakes, frustration, and burnout among the medical practitioners struggling to put this software to use in meaningful ways.

Our summary coverage will continue on this topic next month.