At Long Last Data: Reducing Trainees' Exposure to Marketing Makes a Difference

We can start with this nice summary piece by Kevin O'Reilly in American Medical News:
http://www.ama-assn.org/amednews/2013/02/18/prsb0218.htm#.USHqTMUZI1s.twitter

We've had a problem for some years now. After decades of howling in the wilderness, the pharmascolds started to win battles around 2005-2009. We now have many more policies in place to limit the influence of pharmaceutical marketing on medical education and practice--not the least of which is PhRMA's own code of conduct from 2009, which is why, when I recently asked a mostly-medical-student audience how many of them were carrying pens with drug firm logos, no hands went up. If I had asked that question of a similar audience in 2005, the response would have probably been, on average, half a dozen branded items per person.

So what if anything has actually changed? There have been very few published data to answer that question. At first, the reforms were obviously too new to allow careful study of effects, but as years have gone by, we naturally wonder what the record is.

Now the article above notes two recent publications, one addressing medical school, the other residency.

First, Marissa King and a team headed by Dr. Joseph Ross of Yale:
http://www.bmj.com/content/346/bmj.f264?view=long&pmid=23372175
--did a very clever study of the impact of medical school policies discouraging contact with drug reps on later physician prescribing. They set about doing their study the same way drug reps market drugs--they bought the IMS Health database on practitioners' prescribing and the AMA masterfile that allows those data to be linked to individual physicians. Only instead of name and address of the doc, they wanted to know which med school he/she graduated from and the graduation year. They then looked at three psychotropic drugs that entered the market in 2006-8, that in their opinion represented me-too drugs that had no major advances over older, often cheaper drugs-- lisdexamfetamine (Vyvanse), palperidone (Invega), and desvenlafaxine (Pristiq). They compared prescribing of these drugs between two groups--those attending schools that had instituted policies to exclude drug reps by 2004 and those that had not. Essentially they tried to figure out the lag time between attending a school that had a certain policy, then completing an average residency, and then being out in practice and prescribing drugs independently.

King et al. noted two reasons to predict that they'd find no significant differences. First, the schools that implemented policies back in 2004 often had pretty weak and tentative policies; it was quite normal for these policies to be substantially tightened as the decade progressed. Second, they were unable to control for the residency experience, so for all they knew, when med students who had seen no drug reps as part of their MD education then proceed on to residencies where they were awash in marketing, the effects of the med school policy would be washed out.

Theredfore it is notable that King et al found significant differences for two of the three drugs they studied, where graduates of the schools with strict policies prescribed significantly less of the expensive but hardly better new drugs. The authors admitted that they had no data on whether this would carry over to really useful new drugs--would they prescribe less of those too? But in any event we have some tentative and preliminary data to suggest that med school policies make a difference.

What about residencies? The next study is out of a team headed by Dr. Andrew Epstein at Penn:
http://www.ncbi.nlm.nih.gov/pubmed/23142772 (subscription required). These folks looked specifically at psychiatrists prescribing antidepressants, and asked how often they prescribed three different groups of drugs--heavily promoted; reformulated; and brand name. (Arguably, smart docs would prescribe less of all three groups.) They looked at two variables--what year the psychiatrist graduated from residency, and whether the residency was located within an institution that had a strict, moderate, or loose policy restricting drug marketing (based on the AMSA scorecard). They found that merely graduating at a later date did not substantially effect prescribing, but the degree of policy restriction did.

We have no evidence to show that overall, pharmaceutical marketing is less a part of the overall medical environment today than it was a decade ago.  In the commonly cited statistics of Eric Campbell's group, the number of docs reporting contact with drug industry sources went from 94% in 2004 to 84% in 2009 (see: http://brodyhooked.blogspot.com/2010/11/are-physicians-taking-fewer-bribesgifts.html). That's hardly a major decrease. In such an environment, many would imagine that a med school or residency banning reps and drug lunches is a prely symbolic and ineffective gesture; and the argument is still heard that instead of banning these activities, a smart residency director would encourage them and use the contacts with reps as a teaching opportunity for the "real world." So these new studies, while hardly definitive, are nonetheless major steps toward showing what actually happens when one implements such policy changes.

(Hat tip to good friend and fellow blogger Alice Dreger for the link to the O'Reilly piece.)