In 1999, an Institute of Medicine (IOM) report “To Err is Human: Building a Safer Health System” brought about an awareness about medical errors in healthcare organizations that had not previously existed. The report was based upon analysis of multiple studies by a variety of organizations and concluded with the shocking figure that between 44,000 to 98,000 people die each year as a result of preventable medical errors.
Since the release of that report, the push for improved patient safety has continued across the healthcare ecosystem. But just last week, a study by researchers at Johns Hopkins Medicine, reported on by National Public Radio (NPR) online, revealed that medical errors rank as the third-leading cause of death in the U.S., and highlights how shortcomings in tracking vital statistics may hinder research and keep the problem out of the public eye, as reported in a news story by Healthcare Informatics Editor-in-Chief Mark Hagland. The study was published an article The BMJ (formerly the British Medical Journal).
Per Hagland’s report, the NPR article notes that, “Based on an analysis of prior research, the Johns Hopkins study estimates that more than 250,000 Americans die each year from medical errors. On the CDC's official list, that would rank just behind heart disease and cancer, which each took about 600,000 lives in 2014, and in front of respiratory disease, which caused about 150,000 deaths,” for an estimated average of 400,000 deaths a year. “
What’s more, The BMJ article itself mentions the aforementioned report, per Halgand: “The most commonly cited estimate of annual deaths from medical error in the U.S.—a 1999 IOM report—is limited and outdated. The report describes an incidence of 44,000-98,000 deaths annually. This conclusion was not based on primary research conducted by the institute but on the 1984 Harvard Medical Practice Study and the 1992 Utah and Colorado Study. But as early as 1993, Leape, a chief investigator in the 1984 Harvard study, published an article arguing that the study’s estimate was too low, contending that 78 percent rather than 51 percent of the 180,000 iatrogenic deaths were preventable (some argue that all iatrogenic deaths are preventable). This higher incidence (about 140,400 deaths due to error) has been supported by subsequent studies which suggest that the 1999 IOM report underestimates the magnitude of the problem.”
As such, various health IT leaders point to the role that technology such as electronic health records (EHRs) can play in preventing medical errors. James Merlino, M.D., president at South Bend, Ind.-based clinical improvement software company Press Ganey, and former chief experience officer and associate chief of staff at the Cleveland Clinic, for instance, says “Every month you’re seeing evidence that [technology] decreases medical errors and thus improves safety. Merlino, who is a former clinician himself, feels that organizations do a good job in teaching providers how to use EHRs, but are not great in figuring out how to best integrate them into their workflows. “When they can integrate it into their workflows, it’s not only a good tool for providing additional capabilities, but it helps them deliver higher quality and safer care,” Merlino says.
James Merlino, M.D.
Merlino says what strikes him about the BMJ study, and what should scare all patients, is indeed that the 1999 IOM report “underestimated the severity of the problem.” He says, “Stepping back from the technology aspect, organizations need to make safety a non-comprisable core value. You can be the best surgeon operating at the greatest health center in the world, but if you make a safety error and kill a patient, what’s the point?”
Human Problem or Tech Problem?
Paul Dexter, M.D., research scientist, Regenstrief Institute, Inc., and chief medical information officer (CMIO) of Indianapolis-based Eskenazi Health, feels that because clinicians have not grown too comfortable with their use of EHRs still, providers have not been able to optimize the best ways to use IT to decrease medical errors. “We still have user-friendliness issues to work out. Clinicians are not terribly satisfied with these EHRs. We have a long evolution to go yet,” Dexter says.
Nonetheless, Dexter points out that while humans are imperfect, they have a certain rate of making errors, for example, diagnostic errors, as it’s estimated that 10 percent of initial diagnoses are wrong, due to biases and the like. “But that’s not to say that by far the vast majority of clinicians are well-intentioned, well-meaning, focused, and hard working,” he says. “The focus needs to be on the work environment in which they are working.”
Paul Dexter, M.D.
Dexter notes the IOM report which says that humans will make errors as they are the ones working with the patients, and as such, they want a safety net. “Whether that safety net is a pharmacist calling up the clinician when they see an error or having technology in place to catch an error even earlier in the process, you want to make the system similar to the aviation industry so that it’s basically harder to make a mistake than not. You want to have multiple points so that if you do make an error, there is a good chance of intercepting it. And if that fails, you have yet another technology that might catch it,” Dexter says. He notes that of course, human error is inevitable. “You need to recognize that errors will happen and not go out of our way to shame the person to make that error. You have to be transparent, make it clear that when an error happens, learn from it, and use technology to improve the processes of care.”
Both Dexter and Merlino do put a level of onus on software vendors too, however. “Poor user interfaces can be confusing and lend themselves to error, whereas intuitive user-centered designs can decrease the rate of errors,” Dexter says. Merlino points to software from Epic Systems as an example of a vendor stepping up to the plate, as it has the capability to bring up early warning solutions. So if a patient is in the hospital and the organization is on the Epic system, if the patient is demonstrating markers of sepsis, the system gives the provider the ability to call up the warning system to the nurse, and display a care path or order set, so you get quicker validation of what’s happening with the patient, as well as putting a solution set in front of the caregiver or nurse on what to do next, Merlino explains. “Other vendors are doing it too, and it’s [an example] of EHRs being more than just a repository of information, but a decision aid to help us make faster, more efficient decisions. That’s very exciting.”
Regarding the long-debated issue of a universal patient identifier, which in theory, would solve the problem of being unable to track patients as they move from institution to institution, thus reducing errors, Dexter says that in order to make sure that an organization receiving the patient can get all the records from all previous organizations that a patient has been to previously, a universal patient identifier “would be [awfully] helpful.” He adds, “There have been lots of patient matching algorithms to bypass that, but to me, not that’s not as good as a universal patient identifier.” Dexter does admit that he doesn’t know whether clinicians, who already are burdened with an excessive amount of information, would look at the record that came from the outside source at the right moment in time and process it correctly, thus preventing an error. That remains an unknown and would need rigorous testing, he says.
Reflecting back on the IOM report, Merlino says that while there has been real progress made since 1999, there have not been enough significant changes to organizations’ approaches to patient safety, noting that it will require both process and culture shifts. “You need to think about how to learn from other industries,” he says. “Look at what the airlines have done, what nuclear power has done, and how the military functions. We need to be inserting concepts of higher liability into our processes, and that will be critical in our drive towards zero harm.”
And Dexter notes that as healthcare continues to shift towards value-based care and incentives for quality care, if an organization is responsible for a population of patients, if a mistake is made leading a patient to be hospitalized for longer periods of time, the organization would be incentivized from multiple levels to prevent that from happening. “This is part of the broader learning health system,” Dexter says. “You want to learn from every patient and every mistake to generate new knowledge so you can improve care and safety.”
