Thursday, February 02, 2012

Scoping it out

I have often expressed my annoyance at technical medical journals that restrict access to articles of public interest.  I understand their need for revenue, but I don't see that they would lose income if they had a policy of publishing a few articles each month that might have a broad audience.  The New England Journal of Medicine, for example, does this, with no loss of readership.  Indeed, you can argue that its status is enhanced by the practice.

But beyond the convenience factor, when articles only appear in abstract form, they can raise more questions than they answer.  Here's an example from the journal Forensic Science, Medicine, and Pathology.  I reprint the entire abstract:

Fiberoptic endoscopy is utilized to diagnose and treat a wide variety of gastrointestinal tract diseases and is currently one of the most commonly performed invasive medical procedures. Though generally considered to be safe, the procedure may be associated with serious complications including death. Herein, we correlate the clinical history with the autopsy findings in gastrointestinal endoscopy (GIE) related fatalities that were reported to the New York City Medical Examiner Office over a 5 year period. The complication rate is higher in therapeutic procedures when compared to diagnostic procedures. Though fatalities following iatrogenic complications are four times more common, emphasis is placed on patients who died following a cardiorespiratory event during or immediately following the procedure. Of 41 upper and 36 lower GIE related fatalities, 10 and 9 respectively, fit this category. The comorbidities frequently associated with fatal cardiorespiratory events were: hypertension and obesity with associated cardiac hypertrophy, diabetes mellitus and coronary artery atherosclerosis. For colonoscopies, the approximate incidence of perforation was 0.00625% and the procedure-related fatality rate was 0.014%. The literature on anesthesia for GIE is also discussed. 

The article seems to be saying that doctors should be quite careful in using therapeutic endoscopies when patients have hypertension and those other conditions.  But it is hard to judge from the abstract how important this all is.  We often hear that the most important adverse event associated with colonoscopies is perforation of the colon, but here we learn that procedure-related fatalities occur twice as often, with some percentage associated with the patients' baseline conditions.  Yes, that is still a very small percentage of the time.  But does the figure nonetheless suggest that best practice by GI doctors should consist of screening for hypertension and obesity with associated cardiac hypertrophy, diabetes mellitus and coronary artery atherosclerosis?  Does is suggest that the patient consent process for people with those conditions should include a special warning?

Or, is this much ado about nothing, with a number so small as to be insignificant?  I bet that GI doctors don't regularly receive this journal.  If they are not subscribers, how would they be able to evaluate the article for application to their own practices?


Anonymous said...

The GI docs at our hospital are very much aware of the high risks associated with endoscopic procedures in patients with diabetes, hypertension, obesity, CHF etc. We routinely screen for these co-morbidities, and if present we often perform the procedures with the help of an anesthesiologist who evaluates the patients pre-procedure and delivers the sedation. Our goal is to make our procedures as safe as possible.

The medical literature, particularly the GI journals like Gastroenterology, Amer J of Gastro, J of GI Endoscopy and many more report on this topic frequently so we have good access to studies and guidelines from our professional societies.By implementing protocols and guidelines here, we have achieved endoscopic safety and outcomes far better than the national averages for each procedure.

Anonymous said...

I'm reasonably comfortable that at a competitive hospital the above is true. But we shouldn't assume so in most environments, and should not have to assume at all. It is testable. What is the patient mix being brought into the procedure? What is their outcome? Is it reported and acted upon as learning?

The most critical question is: What do GI patients know about their risks? While consenting for risks is increasingly (with scientific progress) a moving target, it is an administrative and training issue for providers, but a potentially life-or-death one for patients.

Where is strong oversight of risk knowledge and consenting in quality improvement? What board committee or external regulator (or malpractice insurer) assesses this progress?

There is a gap between results presented in this paper and results as experienced by patients harmed. If you are actually an expected poor outcome based on provider metrics (e.g. it isn't harm because you were diabetic, after all), does that make your poor outcome any easier to experience? Did you know that as a diabetic/obese/smoker/lame-exerciser/sometime-consumer-of-vegetables that your prognosis is not as good?

Perhaps this is conversation where patient education can finally meet the traction of provider accountability.