It’s time to change the way we measure medication adherence

Solutions to complex issues often appear to be obvious in hindsight, and such could be the case with medication adherence — whether or not patients take their medication as prescribed.

Medication adherence is critical for one simple reason, as succinctly put by our former Surgeon General C. Everett Koop in 1985: “Drugs don’t work in patients who don’t take them (1).” Actual adherence can only be measured on a dose-by-dose basis.

One of the biggest challenges facing our healthcare system is the fact that people generally do not take much of their medication. Healthcare costs related to nonadherence exceed $300 billion annually (2), but because current measurements of adherence are rough approximations at best, this cost is likely an underestimate.

A solution to the problem has existed for decades, practiced daily by public health experts all over the country: Directly Observed Therapy or DOT, where a healthcare professional watches the patient take every dose of medication. The confusing semantics around the measurement of adherence blind our most sophisticated health economists and policymakers to DOT’s unparalleled success. Most importantly, research shows that in underfunded and overlooked health departments all over the country, technology has made DOT affordable and scalable while achieving medication adherence rates greater than 90 percent.

Healthcare researchers, clinicians, payers, and institutions have come up with multiple confusing and indirect measures of “adherence” over the years, many of which describe proxies for whether patients actually take medication. These various methods — medication possession ratios, pill counts, electronic pill bottles — are all based on the notion that we are not able to truly measure dose-by-dose adherence.

“Primary adherence,” which is whether a person possesses the prescription, is the most pervasive of these indirect measures. Our reliance on the standard may also be the most costly in terms of financial and healthcare outcomes. Primary adherence is measured through analyzing pharmacy fill and refill data. Of course, the fill or the refill of medication is tied to a financial transaction — billing for medication — and therefore an entire industry has developed with incentives for manufacturers, PBMs, pharmacies, ACOs, payers, and even nonprofits, to make sure that a person’s prescription is filled in a timely manner. This has lead to sharply rising medication costs particularly in Medicaid and Medicare.

Every single player in the supply chain, except the patient, benefits financially from the filling of a prescription — even if the patient does not take their medication. With approximately 50 percent of medication taken as prescribed (3), why worry about rising drug costs if we are not going to make sure that people actually take these effective, innovative, and expensive therapies?

DOT is the standard of care for tuberculosis (TB) treatment and has been wildly successful, helping cut the incidence of TB from more than 30,000 in the 1970’s to under 10,000 the last three years. The practice is much more than just watching someone take medication — it is a system which involves case management, coordination of medication delivery, and compassionate coaching and support by healthcare professionals who keep people motivated to stay on track.

DOT’s reputation is a barrier to widespread adoption. It is perceived as inconvenient, expensive, and impossible to scale. In fact, traveling to a patient or having the patient travel to a clinic in order to observe medication being taken is quite burdensome for everyone involved, despite the effectiveness of the method. Traditional DOT is certainly not patient-centric, especially where daily travel to the clinic is required. While DOT has been proven to secure greater than 90 percent adherence, it is not a surprise that it has been limited to a handful of serious conditions.

Technology has changed DOT so that it can be scaled. Thousands of people video themselves taking their medication every day all over the country and — in combination with regular face-to-face meetings with clinicians — achieve extremely high rates of medication adherence at a fraction of the cost of in-person DOT.

Now that DOT is less expensive, and less burdensome, we can measure adherence directly, quantifying the number of doses actually taken by a patient. With more than $350 billion spent on medication each year, it is an urgent matter to clearly understand whether people are taking their medication regularly, and help those who are not doing so for various reasons.

Our vocabulary around medication adherence and our preconceptions about the feasibility of dose-by-dose measurement of adherence need to catch up to our technological readiness. HEDIS measures give triple weighting to primary adherence, while PBMs use the same measure in claiming that they have solved medication adherence.

Public health departments implemented DOT long ago to stop a killer in its tracks. Now opioid overdose kills more Americans each year than TB ever did, and more than half of adult Americans have more than one chronic condition. It’s time to change the way we measure medication adherence to make sure we can effectively address today’s major health crises.


  1. Koop CE. Keynote Address: Council NP, ed. Improving Medication Compliance: Proceedings of a Symposium. Reston, VA; 1985.
  2. Aurel Luga, Maura McGuire, Adherence and Healthcare Costs, Risk Management and Healthcare Policy, 2014; 7: 35-44
  3. Marie Brown, MD, Jennifer Bussell, MD, Medication Adherence: WHO Cares?, Mayo Clinic Proceedings, 2011 April; 86(4): 304-314