Skip to main content.

Helen Haskell on Tracking Medical Errors: How We Err When Counting the Casualties of Medical Care

Helen Haskell on Tracking Medical Errors: How We Err When Counting the Casualties of Medical Care

Picture of William Heisel

It has been 12 years since Helen Haskell lost her 15-year-old son, Lewis Blackman, because of a series of medical errors, and she has worked hard in that time to prevent similar errors. She founded Mothers Against Medical Error (MAME). She helped with a campaign to get a hospital infection public disclosure bill passed into law in South Carolina. And she lobbied the South Carolina legislature to have the Lewis Blackman Hospital Safety Act passed in 2005 in honor of her son.


Helen Haskell, William Heisel, reporting on health, medical errors, Lewis Blackman, patient safety

When Haskell’s son was suffering from the effects of a surgery gone wrong, Haskell tried to get the attention of physicians. She was punted to a medical resident, and only found out later that the resident had very little experience. The Blackman Act requires that staff wear tags indicating whether they are physicians, residents, or medical students. It also requires nurses to contact attending physicians when a patient needs them.

Haskell contacted me when I was writing about Sorry Works! with some thoughts on the number of people harmed by medical professionals, and I asked her to write some guest posts about the subject. The first is below.

We will not be able to effectively prevent medical harm until we have a better handle on just how much harm there is. 

The source that is most often cited for the frequency of medical harm in the United States  – the 1999 Institute of Medicine (IOM) report To Err is Human – relied on data from the 1980s and 1990s for its broad estimate of 44,000 to 98,000 annual deaths from medical error.  These numbers came from two studies – one in New York and one in Utah and Colorado – that were initially undertaken to measure the feasibility of malpractice insurance reform. They looked at a wide range of adverse medical events, both fatal and non-fatal, and they also found big differences between populations. The death rate in the New York study, for example, was more than twice that of Utah and Colorado.

Overall, the New York study found that adverse medical events caused harm to 3.7% of all hospital patients. In the Utah-Colorado study, the rate was 2.9%. Only a fraction of these adverse events were considered to be due to error.

The IOM arrived at its totals by applying the medical error death rates in each study to the total number of hospital admissions in the United States in 1997.  The errors counted in the IOM report included only those that both met the strict standard of legal liability and also led to death. This amounted to about one death in every 350 to 750 hospital admissions.

As shocking as these numbers were, many people have felt that they did not capture the extent of medical harm in the United States. 

For one, the investigations were restricted to hospitals, which provide only a portion of the medical care in the US.  What about the millions of patients who undergo procedures in surgery centers and outpatient clinics? What about residential care and medical treatment in facilities like nursing homes and dialysis clinics? What about the decisions made in doctors’ offices that lead to patient harm or death?

In addition, by focusing just on medical errors for which a health care provider might be legally liable, the IOM left the impression that millions of other medical injuries might be somehow inevitable.  But how do we know what an error is? And who decides? Then there is the issue of the accuracy of medical records, which a 2008 Harvard study suggested might be missing as many as half the adverse events that harm patients.

The biggest problem with the IOM numbers, however, is simply their age.

Medical research moves slowly. Even by the time of its publication, the IOM report’s numbers were far from current. The New York numbers were 15 years old. And the Utah-Colorado numbers were seven years old.  Today, those numbers are 28 years and 20 years old. 

Health writers would not write about cancer statistics from 1984 as if they were current or diabetes rates from 1992 as if were just released. So why do these dusty IOM numbers – with all their limitations – continue to hold sway with journalists?

In part it’s because the IOM numbers caused such a stir in 1999 that they became widely cited in the medical literature, reinforcing the old estimate even in very recent publications.  Also, by rounding up the 98,000 estimate, writers have created an ominous and easy-to-remember total: 100,000 deaths every year. It sounds scary.

More recent estimates, though, tell us that the true number of patients harmed every year by the medical system is probably even scarier. I’ll write about those studies – and some of the issues raised by the studies – in the next post.

Related Posts:

Easing the Harm From Medical Errors: Q&A with Sorry Works! Founder Doug Wojcieszak

Apology as Cure: Finding the Secret Ingredients to Make "Sorry" Really Work

Apology as Cure: Dig into Data to Find Number of Patients Harmed

Apology as Cure: Should Laws Change to Encourage Doctor to Admit Medical Errors?

Comments

Picture of <span class="username">Guest (not verified)</span>

I found the "Harvard" study of 2008 somewhat confusing and never could get clear answers from the first author on my questions. That said, according to the abstract, the medical records showed 11 serious, preventable adverse events, and patients could recall an additional 21 serious, preventable events. This means the medical record under-reported such errors by MORE than a factor of 2...it's more like a factor of 3: (11+21)/3 = 11

Picture of <span class="username">Guest (not verified)</span>

The recent deFeijter article (PLoS ONE 7(2)) reported internal reporting systems (IRS) do not accurately capture all adverse events or incidents. Using information from multiple sources: patient complaints, retrospective chart review and from healthcare workers is preferable. Patient complaints revealed more incidents in relation to diagnosis (surprising!!), general care and procedure/treatment than the information entered by healthcare workers into the IRS.

Leave A Comment

Announcements

The GOP’s health reform bills call for a massive rollback of Medicaid, and changes via waivers could reshape many state programs. This webinar will give participants the policy primer they need to understand such historic changes and highlight story ideas reporters can pursue. Sign up here!

Do you have a great idea for a data-informed health reporting project?  We'll give you four days of intensive training, a $2,000 grant and six months of expert mentoring to help produce it. Click here to find out more.

CONNECT WITH THE COMMUNITY

Member Activities

Katharine Gammon has shared a blog post

Read it.

Anna Romano has shared a blog post

Read it.

Barrington Salmon has shared a blog post

Read it.

Richard Cohen joined the community

Connect with Richard Cohen

Jesus ramirez commented on a post

Join the conversation.
More Member Activities

Follow Us

Facebook


Twitter

CHJ Icon
ReportingHealth