Wiklund's Perspective: Patient Deaths Due to Medical Error

By Michael Wiklund

Major initiatives to improve patient safety have been driven in part by alarming reports on the number of fatalities due to medical errors committed in hospitals. In recent times, a stay in the hospital has been characterized as risky on the level of driving drunk. In 1999, the Institute of Medicine (IOM) published To Err is Human, which estimated the number of annual deaths in the U.S. due to error at up to 98,000. Seventeen years later, researchers at Johns Hopkins estimated the annual death rate in the U.S. to be at least 250,000:

“… 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, which the researchers say now translates to 9.5 percent of all deaths each year in the U.S.”  

As you can imagine, these estimates were shocking and a strong call to action at both points in time spanning 17 years. In parallel, the troubling reports were inflammatory to some clinicians who could recall isolated, “sentinel” events but not a continuing swarm of error-related deaths over decades of work at various hospitals. 

I confess to accepting the high estimates at face value, thankful for a foundation for discussing the dirge of medical error and how human factors might help the situation. I assumed that the estimates were reasonably accurate, at least in terms of their order of magnitude. Afterall, the studies were performed by highly respected institutions and experts and still have merit in the eyes of many. Therefore, in many human factors (HF) talks and workshops, I cited the death counts as evidence that manufacturers should applying HF rigorously during the medical product development process. I and many others in my professional evangelized HF as a reliable way to reduce the rate of use errors that could be harmful. I proffered that demonstrably good user interfaces protected against use errors. My routine experience watching clinicians commit use errors while interacting with prototype and marketed medical devices affirmed my point of view.  (Spoiler alert. I continue to believe heartily in the value of HF.) 

Fast forward to late 2022. Multiple media sources reported on the work of Kristen Panthagani, MD, PhD who studied the death rate research and now challenges the stunning estimates. She reviewed the basis of the recent estimate of 251,000 deaths per year in the U.S. and argued that it was an inappropriate extrapolation from a Canadian study that cited a single death out of 503 patients who visited two selected emergency departments during a specific time period. She challenged the Agency for Healthcare Research and Quality (AHRQ) findings as follows:  

“[AHRQ misused] this single death to estimate a death rate across the entirety of the US. Dividing 1 death by 503 patients, they estimate the death rate to be 0.2%. They then multiple [sic] by total ER visits in the US: 130 million visits x 0.2% = ~250k deaths.” 

So now, what’s an HF practitioner to think? Did a reasonably clear picture of medical error rates and consequences just get blurred? Maybe so, depending on what we are to accept as reasonable estimation or not. What is clear to me is that there is no system in place to determine exact number of deaths due to medical error. The incident rate is hardly as clear-cut as tallying airline accident deaths worldwide in 2021, for example, which is cited precisely as 134 due to 15 separate accidents. Unlike in aviation, not every death is investigated in terms of its root cause; far from it. Causes of death are routinely reported, but many reporting systems are not engineered to describe explicit, HF-related causes. Yes, there are cases in which medical error is cited as a definitive cause, but these cases are described as greatly underreported.  

My take is that we do not need a perfect estimate of deaths due to medical error in the U.S. or anywhere else. We know that “to err is human,” just as the IOM book titled reminds us. We know that medical products are imperfect at accommodating every potential user in every potential use environment in every potential use scenario. Sometimes a medical product’s user interface will induce a use error like making the wrong tube connection, setting the wrong medication dose, or misinterpreting an alarm. Therefore, it continues to make sense to apply HF to mitigate the risk of use errors that can cause harm. The mission remains valid. I am not aware of any parties that oppose a continuing investment in HFE by the medical industry, although some might question the scale of the efforts and the standard of acceptance of the work. In any case, I and many of my colleagues in human factors believe firmly that society would benefit from a more comprehensive medical error surveillance system that not only tracks medical errors (similar to MAUDE), but also draw conclusions about root causes and suggests future risk mitigations. 

Finally, let’s suppose that an annual death rate of 250,000 people in the U.S. is an overestimate by 100%, 200%, or even 500%. If the latter, you would still be seeing 50,000 deaths per year, which is more than the total number of motor vehicle traffic deaths in 2020, which was itself an alarming 40,698. 

In my view, the call to action – to apply human factors rigorously during medical product development – remains strong. 

Michael Wikund is General Manager of Emergo by UL's Human Factors Research & Design division.