Monday, January 28, 2013

Coffee Drinking, Mortality, and Prespecified Falsification Endpoints

A few months back, the NEJM published this letter in response to an article by Freedman et al in the May 17, 2012 NEJM reporting an association between coffee drinking and reduced mortality found in a large observational dataset.  In a nutshell, the letter said that there was no biological plausibility for mortality reductions resulting from coffee drinking so the results were probably due to residual confounding, and that reductions in mortality in almost all categories (see Figure 1 of the index article) including accidents and injuries made the results dubious at best.  The positive result in the accidents and injuries category was in essence a failed negative control in the observational study.

Last week in the January 16th issue of JAMA Prasad and Jena operationally formalized this idea of negative controls for observational studies, especially in light of Ioannidis' call for a registry of observational studies.  They recommend that investigators mining databases establish a priori hypotheses that ought to turn out negative because they are biologically implausible.  These hypotheses can therefore serve as negative controls for the observational associations of interest, the ones that the authors want to be positive.  In essence, they recommend that the approach to observational data become more scientific.  At the most rudimentary end of the dataset analysis spectrum, investigators just mine the data to see what interesting associations they can find.  In the middle of the spectrum, investigators have a specific question that they wish to answer (usually in the affirmative), and they leverage a database to try to answer that question.  Prasad and Jena are suggesting going a step further towards the ideal end of the spectrum:  to specify both positive and negative associations that should be expected in a more holistic assessment of the ability of the dataset to answer the question of interest.  (If an investigator were looking to rule out an association rather than to find one, s/he could use a positive control rather than a negative one [a falsification end point] to establish the database's ability to confirm expected differences.)

I think that they are correct in noting that the burgeoning availability of large databases (of almost anything) and the ease with which they can be analyzed poses some problems for interpretation of results.  Registering observational studies and assigning prespecified falsification end points should go a long way towards reducing incorrect causal inferences and false associations.

I wish I had thought of that.

Added 3/3/2013 - I just realized that another recent study of dubious veracity had some inadvertent unspecified falsification endpoints, which nonetheless cast doubt on the results.  I blogged about it here:  Multivitamins caused epistaxis and reduced hematuria in male physicians.

Sunday, January 27, 2013

Therapeutic Agnosticism: Stochastic Dominance of the Null Hypothesis

Here are some more thoughts on the epistemology of medical science and practice that were stimulated by reading three articles this week relating to monitoring interventions:  monitoring respiratory muscle function in the ICU (AJRCCM, January 1, 2013); monitoring intracranial pressure in traumatic brain injury (NEJM, December 27, 2013); and monitoring of gastric residual volume in the ICU (JAMA, January 16th, 2013).

In my last post about transfusion thresholds, I mused that overconfidence in their understanding of complex pathophysiological phenomena (did I say arrogance?) leads investigators and practitioners to overestimate their ability to discern the value and efficacy of a therapy in medicine.  Take, for instance, the vascular biologist studying pulmonary hypertension who, rounding in the ICU, elects to give sildenafil to a patient with acute right heart failure, and who proffers a plethora of complex physiological explanations for this selection.  Is there really any way for anyone to know the effects of sildenafil in this scenario?

Monday, January 14, 2013

Death by 1000 Needlesticks: The Nocebo effects of Hospitalization

I couldn't decide if this belonged on Status Iatrogenicus or the Medical Evidence Blog.  Since it has relevance to both, I'll post a link here:

Hemoglobin In Limbo: How Low Can [should] It Go?

In this post about transfusion thresholds in elderly patients undergoing surgery for hip fracture, I indulged in a rant about the irresistible but dodgy lure of transfusing hospitalized patients with anemia (which I attributed to the normalization heuristic) and the wastefullness and potential harms it entails.  But I also hedged my bets, stating that I could get by with transfusing only one unit of blood a month in non-acutely bleeding patients, while noting in a comment that a Cochrane review of this population was equivocal and the authors suggested an RCT of transfusion in acute upper gastrointestinal hemorrhage.  Little did I know at the time that just such a trial was nearing completion, and that 12 units of PRBCs could probably get me by for a year in just about all the patients I see.

In this article by Villanueva in the January 3, 2013 issue of the NEJM, Spanish investigators report the results of a trial of transfusion thresholds in patients with acute upper gastrointestinal hemorrhage.  After receiving one unit of PRBCs for initial stabalization, such patients were randomized to receive transfusions at a hemoglobin threshold of 7 versus 9 mg/dL.  And lo! - the probability of transfusion was reduced 35%, survival increased by 4%, rebleeding decreased by 4%, and adverse events decreased by 8% in the lower threshold group - all significant!  So it is becoming increasingly clear that the data belie the sophomoric logic of transfusion.