The multimillion-dollar system, manufactured by Epic Systems Corp. and considered by some to be the Cadillac of medical software, had been installed at the hospital about four months earlier. Although the order appeared on Epic’s screen, it was not sent to the lab. It turned out, Epic’s software didn’t fully “interface” with the lab’s software, according to a lawsuit Ronisky filed in February 2017 in Los Angeles County Superior Court. His results and diagnosis were delayed — by days, he claimed — during which time he suffered irreversible brain damage from herpes encephalitis. The suit alleged the mishap delayed doctors from giving Ronisky a drug called acyclovir that might have minimized damage to his brain.
Epic denied any liability or defects in its software; the company said the doctor failed to push the right button to send the order and that the hospital, not Epic, had configured the interface with the lab. Epic, among the nation’s largest manufacturers of computerized health records and the leading provider to most of America’s most elite medical centers, quietly paid $1 million to settle the suit in July 2018, according to court records. The hospital and two doctors paid a total of $7.5 million, and a case against a third doctor is pending trial. Ronisky, 34, who is fighting to rebuild his life, declined to comment.
Incidents like that which happened to Ronisky — or to Annette Monachelli, for that matter — are surprisingly common, data show. And the back-and-forth about where the fault lies in such cases is actually part of the problem: The systems are often so confusing (and training on them seldom sufficient) that errors frequently fall into a nether zone of responsibility. It can be hard to tell where human error begins and the technological shortcomings end.
EHRs promised to put all of a patient’s records in one place, but often that’s the problem. Critical or time-sensitive information routinely gets buried in an endless scroll of data, where in the rush of medical decision-making — and amid the maze of pulldown menus — it can be missed.
Thirteen-year-old Brooke Dilliplaine, who was severely allergic to dairy, was given a probiotic containing milk. The two doses sent her into “complete respiratory distress” and resulted in a collapsed lung, according to a lawsuit filed by her mother. Rory Staunton, 12, scraped his arm in gym class and then died of sepsis after ER doctors discharged the boy on the basis of lab results in the EHR that weren’t complete. And then there’s the case of Thomas Eric Duncan. The 42-year-old man was sent home in 2014 from a Dallas hospital infected with Ebola virus. Though a nurse had entered in the EHR his recent travel to Liberia, where an Ebola epidemic was then in full swing, the doctor never saw it. Duncan died a week later.
Many such cases end up in court. Typically, doctors and nurses blame faulty technology in the medical-records systems. The EHR vendors blame human error. And meanwhile, the cases mount.
Quantros, a private health care analytics firm, said it has logged 18,000 EHR-related safety events from 2007 through 2018, 3 percent of which resulted in patient harm, including seven deaths — a figure that a Quantros director said is “drastically underreported.”
A 2016 study by The Leapfrog Group, a patient-safety watchdog based in Washington, D.C., found that the medication-ordering function of hospital EHRs — a feature required by the government for certification but often configured differently in each system — failed to flag potentially harmful drug orders in 39 percent of cases in a test simulation. In 13 percent of those cases, the mistake could have been fatal.
The Pew Charitable Trusts has, for the past few years, run an EHR safety project, taking aim at issues like usability and patient matching — the process of linking the correct medical record to the correct patient — a seemingly basic task at which the systems, even when made by the same EHR vendor, often fail. At some institutions, according to Pew, such matching was accurate only 50 percent of the time. Patients have discovered mistakes as well: A January survey by the Kaiser Family Foundation found that 1 in 5 patients spotted an error in their electronic medical records. (Kaiser Health News is an editorially independent program of the foundation.)
The Joint Commission, which certifies hospitals, has sounded alarms about a number of issues, including false alarms — which account for between 85 and 99 percent of EHR and medical device alerts. (One study by researchers at Oregon Health & Science University estimated that the average clinician working in the intensive care unit may be exposed to up to 7,000 passive alerts per day.) Such over-warning can be dangerous. From 2014 to 2018, the commission tallied 170 mostly voluntary reports of patient harm related to alarm management and alert fatigue — the phenomenon in which health workers, so overloaded with unnecessary warnings, ignore the occasional meaningful one. Of those 170 incidents, 101 resulted in patient deaths.
The Pennsylvania Patient Safety Authority, an independent state agency that collects information about adverse events and incidents, counted 775 “laboratory-test problems” related to health IT from January 2016 to December 2017.
To be sure, medical errors happened en masse in the age of paper medicine, when hospital staffers misinterpreted a physician’s scrawl or read the wrong chart to deadly consequence, for instance. But what is perhaps telling is how many doctors today opt for manual workarounds to their EHRs. Aaron Zachary Hettinger, an emergency medicine physician with MedStar Health in Washington, D.C., said that when he and fellow clinicians need to share critical patient information, they write it on a whiteboard or on a paper towel and leave it on their colleagues’ computer keyboards.
While the Food and Drug Administration doesn’t mandate reporting of EHR safety events — as it does for regulated medical devices — concerned posts have nonetheless proliferated in the FDA MAUDE database of adverse events, which now serves as an ad hoc bulletin board of warnings about the various systems.
Further complicating the picture is that health providers nearly always tailor their one-size-fits-all EHR systems to their own specifications. Such customization makes every one unique and often hard to compare with others — which, in turn, makes the source of mistakes difficult to determine.
Dr. Martin Makary, a surgical oncologist at Johns Hopkins and the co-author of a much-cited 2016 study that identified medical errors as the third-leading cause of death in America, credits EHRs for some safety improvements — including recent changes that have helped put electronic brakes on the opioid epidemic. But, he said, “we’ve swapped one set of problems for another. We used to struggle with handwriting and missing information. We now struggle with a lack of visual cues to know we’re writing and ordering on the correct patient.”