21 Oct Fact Checking: Part 2
Okay. Here we go.
I started by searching "third leading cause of death in america." One of the first websites that returned was The Healthy Skeptic. It references an article called "Is US Health Really the Best in the World?" authored by Barbara Starfield MD MPH and published in the July 23, 2000 issue of JAMA. Turns out that most of the references to healthcare being the third leading cause of death in the US trace back to this article.
In the article Starfield highlights the following numbers (bulleted below) which are estimates of "the combined effects of errors and adverse effects that occur because of iatrogenic* damage not associated with recognizable error." My interpretation of that sentence is that these numbers detail both deaths which have been specifically attributed to and documented as errors (doing the wrong thing) and deaths in which an outcome perceived as avoidable resulted from a medical intervention.
• 12000 deaths/year from unnecessary surgery
• 7000 deaths/year from medication errors in hospitals
• 20000 deaths/year from other errors in hospitals
• 80000 deaths/year from nosocomial infections in hospitals
• 106000 deaths/year from nonerror, adverse effects of medications
So the most interesting part there is the deaths perceived as avoidable. One assumes that an error has clearly defined criteria (although we might be surprised) but how does "could have been avoided" get defined and how is data collected? One of the great things about medical journal articles is that they reference each other. So one of the first things we're going to do is see what we can understand about where these numbers come from.
The statistic of 12,000 unnecessary deaths per year comes from an article titled "Unnecessary Surgery" written by Lucien L Leape and published in the Annual Review Public Health in 1992, issue 13. In the first part of the article he defines an unnecessary surgery. I'm going to include the full paragraph because it mentions a specific example and I think the example is helpful to understanding how usefulness in this context is often tied to research evidence.
In contrast, Webster's definition of unnecessary, "useless," is easy to use, as it can be based entirely on objective data. No operation is necessary if it is ineffective, i.e. if it does not accomplish its objective for a given clinical situation.(1) For example, if the objectives of coronary artery bypass graft (CABG) surgery are to relieve pain and prolong life, CABG is ineffective and, therefore, unnecessary-for an asymptomatic patient with coronary artery disease that causes blockage of only one of the three coronary arteries, because studies have shown that CABG does not increase longevity in patients with single vessel disease. An unnecessary operation, then, is one that is ineffective or useless. An operation is also unnecessary if it confers no clear advantage over a less risky alternative. In both instances, the operation does not represent a net benefit to the patient. The patient will not be better off. This is the definition we will use.
(1) Rarely is an operation totally ineffective. Internal mammary ligation for the treatment of angina pectoris and glomectomy for asthma are examples. These operations were ultimately discredited by randomized trials. More commonly, an operation is effective for its initial use, but as experience is gained, the indications are broadened to conditions for which it is useless.
Leape goes on to say that unnecessary is also defined at a moment in time. A surgery that right now is thought to be necessary because there are no other options or because enough data hasn't been collected might come to be seen as unnecessary in the future. Of course we can't classify interventions by future information that can't be known, but it is an important nod to the dynamic nature of medical knowledge.
Having defined unnecessary surgery, Leape turns his attention to the question of data collection which gets to the issue of assumption. When someone dies, we have no established system for pausing to reflecting on all the moments that lead up to this one and collectively deciding if there was a moment when a different decision might have lead to a different outcome. There isn't time for that. If the moment was a true error - the wrong medication or the wrong amount of medication or the wrong surgery - it will likely get noticed and recorded. But all the other decision moments, the choice to recommend a surgery instead of another less invasive intervention or the use of one medication over another that caused an adverse effect that wasn't predicted, those often pass with little reflection and are not recorded in a consistent manner across the country. So researchers do their best by looking at all the data we do have and speculating about relationships between that data to build an argument, in this case, about unnecessary surgeries.
For the next post, I'll go into the evidence Leape presents to build a case for capturing data about unnecessary surgery. Stay tuned...
*Iatrogenic means "of or relating to illness caused by medical examination or treatment."