Monday, May 16, 2016

Deciding whether adverse events are preventable or not

Adverse events and poor outcomes are not always preventable. Deciding whether an adverse event is preventable or not can be difficult.

"To Err is Human: Building a Safer Health System," the original Institute of Medicine report in 1999, stated that  between 44,000 and 98,000 deaths each year were caused by preventable medical errors.

That report was widely cited and spawned a number of studies and reviews claiming that anywhere from 250,000 to 440,000 preventable deaths occur in the United States every year.

I was critical of the 440,000 deaths paper as well as the most recent of these estimates—the one claiming 250,000 deaths due to medical errors per year.

It's not widely known or perhaps simply forgotten, but the 1999 Institute of Medicine report also came under fire. In 2000, two researchers from Dartmouth, Drs. Harold C. Sox Jr, and Steven Woloshin, published a critique called "How many deaths are due to medical error? Getting the number right."

In that paper, the authors pointed out that the IOM was correct about the number of adverse events per hospitalization (2.9-3.7%). However, the IOM's report was based on only two studies which used data from 1984 and 1992 and did not define preventable adverse events or medical errors.

Re-analyses of the two papers judged that 54-69.6% of deaths due to adverse events were preventable, figures that Sox and Woloshin said were subjective and not reliable because the re-analyses were performed by looking only at summaries of patient hospitalizations not the actual records.

As is the case with all subsequent preventable death studies, the IOM report relied on many estimates and extrapolations.

Sox and Woloshin concluded, "It is unfortunate that we do not have a credible estimate of the number of deaths due to medical errors." That statement remains true today.

If you think it's easy to tell a preventable complication from a non-preventable one, read this summary of a case from one of the papers cited in the IOM extrapolation.

A 39-year-old woman employed as an engineering and science technician had a laparoscopic cholecystectomy. Three days later, she developed fever and abdominal pain and was found to have a bile leak and possible infectious peritonitis requiring 4-day hospitalization for observation only. 

Was this complication preventable or not?

Feel free to explain your answer in a comment.

In a few days, I will tell you what the reviewers decided.


Anonymous said...

I know way too many patients who have been the objects of medical mistakes and the like. Doctors have a way way different definition. When we see ego going into making a diagnosis and the resulting errors, doctors don't see it but we sure do.

Anonymous said...

Your question is very much rhetorical, but I think that, underlying all this, there's the fact that payers are more interested in driving a hard bargain than in science or even in basic fairness.

If you complain that complications that happened under your watch weren't reimbursed, much of the lay public will tend to see you as money grubbing at best and criminally incompetent at worst--even if the complication wasn't entirely unavoidable in the first place.

Even if your complaint is entirely justified, it will tend to corroborate the kind of sensational headline you mentioned in your previous post, giving Joe Public a warm fuzzy feeling of superiority over those pompous ass docs.

In summary, there's a lot of interest--billions of dollars' worth of interest--in putting physicians on the defensive. Maybe this is biasing some conclusions out there.

artiger said...

Regarding the bile leak example, based on the information given, it could go either way. It depends on if the leak was a draining duct of Luschka, clips falling off the cystic duct stump, or the common bile/hepatic duct being cut. The first is likely something that would not have been visible during surgery, and usually is not avoidable. The second could go either way. The third is an error in a lot of cases, although we don't know if the patient had severe acute cholecystitis, which would make the situation more troublesome.

Since the example only mentioned a four day hospitalization for observation only (i.e., no interventions or re-operation), I'd have to conclude that it was a leak that stopped, although that's a big assumption. That's not an error. It's impossible to know without looking at the whole record, which is the point Scalpel was making above.

Skeptical Scalpel said...

I am simply pointing out that deciding if an error was preventable is not so easy and can be subject to bias.

It's MDs who are writing these papers with wild estimates of the number of preventable deaths. Is it a conspiracy of us against ourselves?

Artiger, you have saved me the trouble of explaining why the summary I cited is a good example of a case that can't be decided by a brief summary. Of course the complication could be preventable or not depending on what caused the leak.

Anonymous said...

As a general surgeon with 12 years of experience: I have never had a cystic duct leak or a major bile duct injury in my practice. I have had some duct-of-Luschka leaks, however; roughly the 1/200 rate quoted in the literature. In recent years, my bile leak rate is down. I think it is because of my increased use of blunt dissection to free the gallbladder from the liver bed, and decreased use of cautery (especially when I can't visualize a nice areolar plane.) Sometimes I essentially peel the gallbladder off of the liver. Staying out of the liver (and also staying out of the gallbladder), is a big deal in my opinion. I think it reduces the rate of bile leaks and of major bleeds from the liver bed.
Regarding the cystic duct: I was trained to place two clips on it, while a resident. Since residency, if there was room on the cystic duct, I might even place 4 clips (I figured that nobody ever died of stainless steel or titanium poisoning.)
However, SS, your point is very well taken, that bile leak may well be a non-preventable complication. Anybody who says otherwise, has only operated on cadavers or dolls.

artiger said...

Scalpel, I'm going to guess that the reviewers called it preventable, no matter what the cause of the leak actually was.

Skeptical Scalpel said...

Anon, I agree with you that some by a leaks are preventable and some, such as the one you mentioned from a duct of Luschka, are not.

That's why I chose that example.

It's an interesting theory you have that perhaps blunt dissection is less likely to cause such a leak. It would be difficult to study because duct of Luschka leaks are so uncommon.

Skeptical Scalpel said...

Artiger, of course you are correct.

Anonymous said...

SS, when using only cautery on difficult gallbladders, I guess I found myself getting out of the proper plane of dissection more often.
I guess that is what makes surgery so hard to study: the surgeon is an important (and unwieldy) variable.

Anonymous said...

I have a question. How would any study handle adverse events and poor outcomes that come with higher risk patients? My mother was diabetic, prone to tia's, and had heart disease. She needed bypass surgery - I worked at a hospital [office staff] and our "best" heart surgeon refused to even consult - given her medical history. She had the surgery done at a different hospital - it was successful and she did well for 20 more years. But a realistic look at her possible outcomes was not the best. One doctor looked and decided his stats were more important [confirmed by a 2nd party], one doctor decided to operate.
The pressure to keep M & M stats positive has to impact both physicians and hospitals. Seems like there is no good way to differentiate statistically between a patient with a greater expectation of good outcome over a patient with a greater expectation of poor outcome.

Wordy and verbose, but I hope you understood my question.

Skeptical Scalpel said...

Anon from 1:28 PM-you need to do what works best for you. If blunt dissection is easier, I see no problem doing it that way.

Anon from 10:46 PM-reports of complications and deaths are supposed to be adjusted for risks. However, many studies use administrative (billing) data not clinical data to report complications and death rates. Admin data are not as precise.

It is well-known that some surgeons react to publicly published complication rates by cherry picking less sick patients. It may very well be that the best surgeons are the ones who take on the sickest patients. That's one of the fallacies of the so-called "Surgeon Scorecard" I have blogged about.

William Reichert said...

It is 100% preventable by not operating, of course. A few years ago, I was told I had gall stones, found during an echocardiogram.The tech said "while were here let's check".
I never had had symptoms. A surgeon met me in the hall and
asked when I wanted by gall bladder out. A GI guy said I better get it out in case I have an attack while camping out in Idaho.
Wary of duct leaks, which I have seen and can be very serious,
and to my mind not 100% preventable, I passed on the surgery.
That was 30 years ago. Still doing fine. Thank you.
Obviously these leaks are not preventable. If they were, then
the word would get out that there was one method which prevents them. There is no such method. No one knows what causes these leaks. The supposition is that the surgeon did something wrong. That is not provable in court. We can say that the surgeon is responsible for the leak. In which case the surgeon
should increase his fees to pay for the costs incurred by the leak. This is what is known as malpractice insurance. This is a misnomer. It should be called adverse outcome insurance.
Getting the jury to call it preventable assists in larger payouts.
It is absurd for a committee to say an adverse outcome is preventable unless the committee can point to a method scientifically proven to prevent a certain adverse outcome.
This will be impossible in surgery due to the multiple actions the surgeon makes during operation.

Skeptical Scalpel said...

William, very well said.

Anonymous said...

We live in a world where we cannot guarantee every car rolling off the assembly line will work flawlessly 100% of the time. This is something we created. How much more complex is the human body? There is so much we don't even know. To think we can prevent every complication or death is just narcissistic pride the likes of which would make Satan envious.

Skeptical Scalpel said...

And many in medicine would like us to adopt "Lean" methods because it has worked so well at Toyota which has recalled about 20 million vehicles in the last 10 years. I'm diving a car with airbags that can kill me and they can't be replaced until July because the parts are not available.

William Reichert said...

The Toyota story is illustrating a very important point.,
When big government or big business decides that a certain method should be universally applied, the rationale is that the use of this specific method will reduce errors which when multiplied by large numbers will result in a better product over all. However, when the method is in fact flawed,this results in
massive defects everywhere.
Today we see the widespread acceptance of central planning
in medicine based on the notion that this will improve care for
the whole world.However, when the method turns out to be flawed ( eg: use of the 1-10 pain scale and a pain hotline ordained by the Joint Commission back in the 90's ) we see a massive problem created: the opioid epidemic. Will the insistence on all the "quality indicators" chosen by the central
medical authorities result in the realization in ten years that
the power thus used inflicted terrible damage in the heath of
Americans and the "health care system" itself. Unchecked bureaucratic power has the potential to do enormous damage.
The history of the Russian planned economy after the Revolution
which destroyed agriculture and resulted in widespread famine
demonstrates the risks in top down management by bureaucrats that made tragic errors in the name of progress. Are we witnessing the same phenomenon in medicine in the USA today?

Skeptical Scalpel said...

William, great points. We may have already lost the battle though.

Skeptical Scalpel said...

Of course, the reviewers of the bile leak after laparoscopic cholecystectomy said it was a preventable complication. As noted in the comments above, most postop bile leaks are NOT preventable.

Post a Comment

Note: Only a member of this blog may post a comment.