Recalling To Err’s Impact and a Small (But Telling) IOM Mistake

Michael MillensonThis year marks the 15th anniversary of the Institute of Medicine (IOM)’s To Err is Human report, which famously declared that from 44,000 to 98,000 Americans died each year from preventable mistakes in hospitals and another one million were injured. That blunt conclusion from a prestigious medical organization shocked the public and marked the arrival of patient safety as a durable and important public policy issue.

Alas, when it comes to providing the exact date of this medical mistakes milestone, the IOM itself is confused and, in a painful piece of irony, sometimes just plain wrong. That’s unfortunate, because the date of the report’s release is an important part of the story of its continued influence.

There’s no question among those of us who’d long been involved in patient safety that the report’s immediate and powerful impact took health policy insiders by surprise.

The data the IOM relied upon, after all, came from studies that appeared years before and then vanished into the background noise of the Hundred Year War over universal health insurance. This time, however, old evidence was carefully rebottled in bright, compelling new soundbites.Medical error was an “epidemic” claiming more victims than breast cancer or AIDS, said the IOM, and, drawing upon newspaper and television accounts of medical mistakes, illustrated that point with stories of individual victims. The report also assigned an eye-catching $17 billion to $29 billion estimate to the economic costs of medical error.

Under the Spotlight

The combination of sensational conclusions and an intellectually serious source catapulted To Err is Human into the media spotlight. The evening of its release, the report received prominent treatment in the widely watched TV network news broadcasts. The next day saw a front-page display in the New York TimesWashington Post, and other newspapers. By one estimate, more than 100 million Americans had seen a story about the report.

As Molla Donaldson, one of the reports three co-authors, later wrote:

Governmental agencies, professional groups, accrediting organizations, insurers, and others quickly responded with plans to define events and develop reporting systems. Health care organizations were put on the defensive….[T]he public held organizational leadership, boards, and staff accountable for unsafe conditions.

Congressional hearings, for example, led to funding of a $50 million patient safety center at what is now the Agency for Healthcare Research and Quality. Some state legislatures also passed patient safety reporting requirements.

Though To Err is Human launched the patient safety movement into the public policy mainstream, it also proved a high-water mark, provoking a level of public attention never reached before or since. That remained true even when newer estimates of the preventable error death toll at hospitals rose to between 210,000 and more than 400,000 deaths annually.

So when, exactly, should the release date of this influential report be celebrated, and why is it important?

Finding the date should be simple, but the IOM’s website is filled with contradictory information. For instance, there’s an 8-page summary document featuring a photo of the report’s cover dated November 1999. But a few paragraphs later, readers are told the report was issued in September 1999.

On a different IOM page the release date is listed as “November 1, 1999.” Click to the cover page and there’s a 2000 publication date, the same as the copyright. Many (but not all) academic citations use that. Yet in the print version of the report scanned by Google, the cover page lists four printings, the first of which is February 2001. Did the IOM misunderstand the difference between “printing” and “reprinting”?

To summarize: according to the IOM, To Err is Human first appeared sometime between September 1999 and February 2001, an 18-month span. Which date is right?

The evidence from contemporaneous sources is unequivocal. The press release announcing the report and providing an advance copy came out on Nov. 29, 1999. Donaldson, not surprisingly, got the citation right in her subsequent paper.

What’s in a Date?

The actual date is important for policy context. The Clinton administration, for example, issued an executive order backing new patient safety measures and procedures just a week after the first press coverage. If you believed that the release date was September, it would be a three-month lag.

The release date also teaches that timing counts. As a TV reporter who called me for comment at the time noted, the IOM’s “j’accuse” jumped out during a slow news period between Thanksgiving and Christmas. The ongoing saga of the Bill Clinton-Monica Lewinsky affair had hit a pause, and turmoil in the Russian Federation was equally in a lull. Just a few weeks later, Russian president Boris Yeltsin would resign and be succeeded by his prime minister, a man named Vladimir Putin.

Put differently, To Err is Human shows policy mavens attempting to marshal public opinion that they need a good story and good luck. (For a report with a good story and bad timing ­– released during the height of the Ebola worries ­– see my own recent examination of patient safety by Congressional district.) With both those elements, they may at least start to change the world.

On the other hand, the 2001 Crossing the Quality Chasm report demonstrated that the IOM didn’t fully understand its earlier success. Though arguably more important in the breadth of issues it addressed, the Chasm report linked a laundry list of “to-dos” with bureaucratic prose and few human-interest stories. Its immediate political impact was nil.

To Err is Human focuses on systemic errors that foil the good intentions of individuals. One common problem is mistakes that occur because of a failure of standardization; for example, two clinicians may inadvertently be using two different measures of a medication’s strength. That kind of failure is as good an explanation as any of why the IOM offers three different publication dates for this report (September, November 1, and November 29), three different years (1999, 2000, and 2001) and no explanation of a possible difference between the online and print release dates.

To err is, indeed, human. Celebrating the 15th anniversary of the IOM report by correcting the website to ensure accuracy and consistency of future academic citations and policy history would be…well, not divine. More like the perfect birthday present.

1 reply »