Study Estimates Medical Errors May Cause More Than 250,000 Deaths Each Year in the US
Analyzing medical death rate data over an 8 year period, Marin Makary and Michael Daniel (Johns Hopkins University [JHU]) estimated that more than 250,000 deaths/year are due to medical error in the US.
Their figure, published May 3 in The BMJ, surpasses the US Centers for Disease Control and Prevention’s (CDC’s) third leading cause of death—respiratory disease, which kills almost 150,000 people per year.
A press release from JHU says that the researchers examined 4 separate studies that analyzed medical death rate data from 2000 to 2008. Then, using hospital admission rates from 2013, they extrapolated that based on a total of 35,416,020 hospitalizations, 251,454 deaths stemmed from a medical error. That number of deaths translates to 9.5% of all deaths each year in the US.
Medical errors should be a top priority for research and resources, conclude Makary and Daniel.
However, some researchers state that Makary and Daniel’s analysis has misrepresented the true situation. In a response posted to the article posted online to the BMJ, Kevin Stewart (Clinical Director at the Royal College of Physicians, London) wrote that the authors’ extrapolations from the literature are unrealistic, based on flawed assumptions, and that “the single case study which they cite is unrepresentative of the majority of preventable hospital deaths”.
Stewart explains that the author’s figure of around 250,000 preventable deaths in the US was based on 2 studies Medicare patients, who, being mostly elderly, would be expected to have a higher rate of adverse events, with more preventable deaths than the general hospital population, and that 1 study was performed at tertiary care centers, which have a high rate of deaths due to adverse events (almost certainly due to a more complex case mix). All 9 deaths in these hospitals were also deemed preventable (most studies find 40%–60% “preventable” deaths), but on the basis of these 9 cases, the authors concluded that 400,000 deaths occurred across the US, which lacks credibility.
Medscape explained that one reason medical error deaths are not accurately measured is that information about errors is not captured on death certificates, which are the documents the CDC uses for ranking causes of death and setting health priorities. Death certificates depend on International Classification of Diseases (ICD) codes for cause of death, so causes such as human and system errors are not recorded on them.
The New York Times gave an example in which a poorly performed diagnostic test caused a liver injury that led to cardiac arrest, but the cause of death was listed as “cardiovascular” when the cause was a medical error. Diagnostic errors, communication breakdowns, the failure to do necessary tests, medication dosage errors and other improper procedures were all considered medical errors in the study.
It’s not just the US—according to the World Health Organization, 117 countries code their mortality statistics using the ICD system as the primary health status indicator.
Makary and Daniel call for better reporting to help capture the scale of the problem and create strategies for reducing it.
“Top-ranked causes of death as reported by the CDC inform our country’s research funding and public health priorities,” Makary said in the JHU press release. “Right now, cancer and heart disease get a ton of attention, but since medical errors don’t appear on the list, the problem doesn’t get the funding and attention it deserves.”
Medscape reported that the new estimates are considerably higher than those in the 1999 Institute of Medicine (IOM) report entitled “To Err Is Human”. The Washington Post explained that this report called preventable medical errors an “epidemic” and led to significant debate about what could be done. The report, based on 1 study, estimated deaths because of medical errors as high as 98,000 a year.
Makary and Daniel state that the data used for that IOM report were “limited and outdated.” The research of Makary and Daniel involved a comprehensive analysis of 4 large studies, including those performed by the Health and Human Services Department’s Office of the Inspector General and the Agency for Healthcare Research and Quality from 2000 through 2008.
Makary and Daniel propose several changes, including making errors more visible so their effects can be understood. Discussions about prevention often occur in limited and confidential forums, such as a department’s morbidity and mortality conference. They also propose changing death certificates to include not just the cause of death, but a field asking whether a preventable complication stemming from the patient’s care contributed to the death.
Makary and Daniel suggest that hospitals carry out a rapid and efficient independent investigation into deaths to determine whether error played a role. A root cause analysis approach would help while offering the protection of anonymity, they say.
Standardized data collection and reporting are also needed to build an accurate national picture of the problem, say experts.
Kenneth Sands (director of health care quality at Beth Israel Deaconess Medical Center) told The Washington Post that the surprising thing about medical errors is the limited change that has taken place since the 1999 IOM report came out. Only hospital-acquired infections have shown improvement. “The overall numbers haven’t changed, and that’s discouraging and alarming,” he said.
Sands said that one of the main barriers is the tremendous diversity and complexity in the way health care is delivered.
“There has just been a higher degree of tolerance for variability in practice than you would see in other industries,” he explained. That makes it tricky to figure out where errors are occurring and how to fix them.
Human error is inevitable, the Makary and Daniel acknowledge, but “we can better measure the problem to design safer systems mitigating its frequency, visibility, and consequences.”
They add that most errors aren’t caused by bad doctors but by systemic failures and should ‘not be addressed with punishment or legal action.