A recent analysis by Malarky and Daniel published in the British Medical Journal (BMJ) suggested that more than 250,000 deaths in hospitalized patients in the United States are a result of medical errors. If correct, this would place medical errors third behind only heart disease (611,000) and cancer (585,000) as causes of death in the United States. The analysis suggests that deaths from medical errors are 7-8 times more common than gun-related deaths. Rather than relying on death certificate information, the authors call for improved tracking of medical related deaths. They suggest adding an additional field to death certificates to record "whether a preventable complication stemming from the patient's medical care contributed to the death." Arguing for a more scientific approach to medical error, they suggest ICD coding should be improved to capture medical error including the physiologic cause of death as well as the related issue with the delivery of care.
Patient safety has been a major priority in healthcare for the past several years. Hospitals are scrutinized by the Joint Commission on a regular basis with a strong emphasis on promoting a culture of safety. Most institutions perform detailed root cause analyses for so-called sentinel events. A sentinel event is an unexpected event resulting in death or serious physical or psychological injury to a patient. Before the 250,000 number is widely disseminated as fact, a critical review of the BMJ report is warranted.
Where does the 250,000 number come from?
The authors in the BMJ paper pooled data from 4 prior studies which reported medical error related deaths in diverse populations, ranging from 10 hospitals in North Carolina to the all Medicare hospitalizations from 2000-2002. Extrapolating estimates from these studies to the total number of 2013 U.S. hospital admissions, they calculated a mean death rate from medical error of 251,454 per year.
Problems with the analysis
- Definition of a medical error
- Distinguishing preventable from non preventable adverse effects
- Distinguishing adverse effects (complications) from errors
- Assessing causality
- Pooling data from different populations/subgroups and with differing assumptions
Assessment of the impact of interventions to promote a "culture of safety" requires an accurate metric for sentinel events, particularly deaths. A closer look at the studies used in the BMJ report raises serious questions about the validity of the 250,000 number. Lumping medical error related deaths with unpreventable adverse effect related deaths only confuses the statistics. One study included in the BMJ analysis counted all adverse effects (100%) as potentially preventable and related to medical error. As all experienced clinicians know, many adverse effects are not related to medical error and certainly not preventable. To count these scenarios as "errors" could seriously overestimate the number of deaths attributed to medical mistakes. Even attributing a death to a clearly recognizable medical error isn't straightforward. The Medicare study used in the pooled analysis included patients over 65 years of age, many of whom likely had multiple comorbid conditions. A medication error in such an individual might be catastrophic. However, it's also likely that many such patients' deaths were unrelated to the medical errors identified.
In sum, the BMJ number of 250,000 medical error related deaths is really a wild guess. The real message of the paper is the need for better data. The evolving robust "culture of safety" in hospitals, promoted by the Joint Commission requires more accurate metrics to assess efficacy of patient safety strategies. As the authors note, "measuring the consequences of medical care on patient outcomes is an important prerequisite to creating a culture of learning from our mistakes."
The next steps
While it may not be 250,000, any number of medical error-related deaths is too high. Patients, caregivers, and health care organizations need to share in the effort to reduce the risk.
- Better definitions and data.
- Educated consumers (patients)
- Smarter EMRs with decision support
- Involve all caregivers
Engaging and empowering patients and their families, particularly in the hospital setting, in their care can be a first line of defense against some medical mistakes, e.g medication allergies or even wrong site surgeries. Providing busy clinicians with state of art technology and seamless point of care decision support is likely to reduce risk of error. I've previously written about the serious risk of EMR-related patient prescription errors. Health IT and healthcare systems urgently need to address transitions of care, i.e., from acute care hospital to home care. These are points of risk for miscommunication with potential for serious consequences. In a complex health system, including all caregivers in the "culture of safety" is essential to success.
I welcome your comments and feedback.