Does the ‘Healthy Adherer’ Effect Muddy the Waters in Medication Adherence?

It’s quite clear that across medical conditions, higher levels of medication adherence are associated with better health outcomes. You may or may not have noticed my careful word choice: “associated with.”

Certainly, it’s widely believed that most approved medications have a strong hand in causing positive health outcomes, with better adherence causally linked to better outcomes. But there are a few wrinkles here. Medicine is not quite so neat and tidy. Consider these interesting findings:

  • A meta-analysis published in the British Medical Journal showed that better adherence to placebos are correlated with better outcomes, specifically lower mortality rates.1This was across a number of conditions, including coronary artery disease, HIV, and diabetes.
  • A study published in Circulation demonstrated that high statin adherence is associated with a significantly decreased risk of motor vehicle and workplace accidents.2
  • In a double-blinded, controlled, randomized clinical trial of placebo vs. medication in congestive heart failure, published in The Lancet, higher adherence was associated with better outcomes—35% lower mortality—for both arms of the study.3

What’s happening here? The “healthy adherer effect” is a key consideration, which is the concept that patients who are careful about adhering to their prescription medications are the same sorts of patients who are likely to adhere to other healthy behaviors that can influence outcomes, like eating right, exercising, avoiding smoking, limiting alcohol, attending regular doctor visits, undergoing screening tests and getting flu shots. I might as well add a few other critical behaviors also linked to decreased mortality: wearing a seatbelt, driving at the speed limit, and avoiding drinking and driving. Probably the same patients here, too.

As an aside, it’s also worth acknowledging that some medical studies, when later repeated by different research groups, reveal completely different results, casting uncertainty on the original studies. This could potentially be the case, for example, with that third study of CHF patients. See this fascinating piece in The Economist…but I digress.

You might even extend the healthy adherer concept to other behaviors. Consider the ridiculous notion of studying the “effect” of statin adherence on voting compliance during election years. I’m almost positive that higher statin adherence would be correlated with a statistically significant higher rate of voting.

Such a study would never make it into the medical journals, but other studies with questionable underpinnings most certainly do, particularly in the lesser journals. The healthy adherer bias may be particularly strong in the study of vitamins and supplements, leading to unwarranted claims, as patients who go out of their way to purchase them out-of-pocket are particularly concerned about their own health and well-being and likely to follow a host of helpful behaviors.

In any medication adherence vs. outcomes study, then, it can be a serious challenge to isolate the specific role of the medication. Most studies, actually, don’t even try.4,5After all, consider how difficult it would be to accurately assess every other healthy behavior in a population, or even to figure out which behaviors mattered most. You certainly wouldn’t want to rely simply upon self-reported questionnaires of adherence to exercise, diet, smoking, seat-belt use, etc. Not to mention factors beyond the behavioral that might also affect outcomes: air quality, income, intelligence, social support, and so on. The promise of big data, with big sample sizes, diverse data sets, and longer timelines, may help clarify things going forward, assuming that we have the right data and data of adequate quality.

It’s reasonable to assume that disentangling the healthy adherer effect from the effect of medication adherence alone is trickier in certain conditions as compared to others. The influence of diet and exercise in diabetes, for example, is arguably stronger than it is for a condition like Gaucher’s disease, where adherence to medication is likely—far and away—the most critical form of adherence.

Regardless, the point here is not to despair that when it comes to medication adherence, nothing can be known for certain. Disentangling behaviors is simply too hard and health outcomes are maddeningly multi-factorial. Instead, the point is to always keep the healthy adherer effect in mind when assessing medication adherence studies, to see if the authors at least tried to account for it, and to maintain a healthy dose of skepticism (but not cynicism), like all good scientists do.

Shrank et al conclude: “Failing to account for behaviors that correlate with medication adherence will lead researchers to conclude that preventive medication use and adherence to preventive medications are more strongly associated with outcomes than is the case.”4

Dormuth et al conclude, in their statin and car accident study: “Our study contributes compelling evidence that patients who adhere to statins are systematically more health seeking than comparable patients who do not remain adherent. Caution is warranted when interpreting analyses that attribute surprising protective effects to preventive medications.”2

I conclude that yes, the healthy adherer effect does muddy the waters when it comes to interpreting the magnitude of the role of medication adherence in health outcomes. But I would also conclude that in an effort to improve outcomes, we need to promote adherence to as many healthy behaviors as possible, including taking necessary prescription medications. Isn’t that what the modern “holistic” approach is all about?


  1. Simpson SH, Eurich DT, Majumdar SR, et al. A meta-analysis of the association between adherence to drug therapy and mortality. BMJ 2006 July 1; 333(7557):15.
  2. Dormuth CR, Patrick AR, Shrank WH, et al. Statin adherence and risk of accidents: a cautionary tale. Circulation 2009;119(5):2051-7.
  3. Granger BB, Swedberg K, Ekman I, et al. Adherence to candesartan and placebo and outcomes in chronic heart failure in the CHARM programme: double-blind, randomized, controlled clinical trial. The Lancet 2005;366(9502):2005-11.
  4. Shrank WH, Patrick AR, Brookhart MA. Healthy User and Related Biases in Observational Studies of Preventive Interventions: A Primer for Physicians. J Gen Intern Med 2011;26(5):546-50.
  5. Bitton A, Choudhry NK, Matlin OS, et al. The Impact of Medication Adherence on Coronary Artery Disease Costs and Outcomes: A Systematic Review. Am J Med 2013;126:357.e7-357.e27.

Questions or comments? You can share you thoughts with our audience in the comments section below, alternatively you can email the author directly at

eyeforpharma Barcelona 2014

Mar 18, 2014 - Mar 20, 2014, Barcelona, Spain

The Future of Pharma