Why doctors ignore drug warnings

By David Jon Phillips     
February 26, 2015

From: Magazine

Sometimes prescription drugs already on the market are shown, after further study, to pose significant risk of serious or life-threatening adverse effects. In the United States, regulators slap prominent warnings on the labels of these drugs.

But the way that doctors respond to these warnings varies significantly depending on where in the country they are, according to research. That variation could mean some patients are being exposed to potentially unsafe medications.

The research team focused on the type-2-diabetes drug rosiglitazone. Doctors in the US wrote 13 million prescriptions in 2006 for Avandia, GlaxoSmithKline’s brand of the drug, generating more than $2 billion in sales, according to health-care data vendor IMS Health. 

The researchers looked at what happened before, during, and after a landmark 2007 New England Journal of Medicine study that suggested using rosiglitazone was associated with a 43 percent increase in risk of heart attack.

After the results were released, the US Food and Drug Administration issued a “black-box warning,” included in the drug-package insert and formatted in large, bold type framed by a black border. While physicians are free to continue to dispense the drug as they see fit, they are likely to change prescription habits and consider alternatives.

The researchers—including those from Chicago Booth, Southern Methodist University, University of Virginia, Loyola University, and University of Chicago—analyzed rosiglitazone use across the 21 regions  defined by the US Department of Veterans Affairs (VA). They studied data from 2003 to 2008, a window of time that documented differences in clinician response related to four key events:
  • The addition of rosiglitazone to the VA national formulary on November 4, 2003

  • Communication within the VA on the day of the first FDA safety alert, May 21, 2007

  • The FDA’s box warning on August 14, 2007

  • The removal of rosiglitazone from the VA formulary on October 4, 2007

The researchers analyzed data on more than 550,000 VA diabetes patients. The VA provides a good environment for study because all VA facilities prescribe from the same list of approved drugs, adhere to uniform administrative and clinical guideline policies, serve a relatively homogenous population, and have salaried physicians whose prescribing habits are less impacted than others by drug-company-offered incentives.

They find that at the aggregate level, rosiglitazone use peaked just prior to the first FDA safety alert, when it represented 14.8 percent of the total prescriptions for diabetes. VA clinicians responded swiftly to both the first safety alert and the black-box warning, with drug use dropping 80 percent to a 2.9 percent share by the end of the study period.

However, the researchers find that doctors in some regions responded more quickly. The variation factor—the ratio of the highest prescription rate to the lowest one—went from an average of 2.8 in the prewarning periods to an average of 3.9 after the warning. Nine of 21 regions recorded peak rosiglitazone-use levels after FDA warnings were issued.

The differences may be due to local practice patterns, the researchers write. In some areas, some VA clinicians and patients are more likely to practice or receive care outside of the VA system, where clinicians could be more impacted by drug marketers. The differences could also be due to poor
communication about risks, beyond the black-box warnings themselves.   

Understanding the variation in doctors’ reactions “is critical to policymakers and hospital administrators responsible for communicating FDA warnings,” says Chicago Booth’s John R. Birge. The findings suggest that these leaders need to develop effective ways to disseminate information related to FDA warnings, as well as better guidelines for providers.—David Jon Phillips

Vishal Ahuja, Min-Woong Sohn, John R. Birge, Chad Syverson, Elly Budiman-Mak, Nicholas Emanuele, Jennifer Cooper, and Elbert S. Huang, “Geographic Variation in Clinical Prescribing Behavior Surrounding FDA Warnings: Evidence from a Nationally Integrated System,” Working paper, October 2014.

Steven E. Nissen and Kathy Wolski, “Effect of Rosiglitazone on the Risk of Myocardial Infarction and Death from Cardiovascular Causes,” New England Journal of Medicine, June 2007.