Medisafe’s Latest Study: Significant Adherence and Health Impacts

Saturday May 5th, 2018

Dr. Danny Sands
Chief Medical Officer, Medisafe

Today, Medisafe shared preliminary results of a study performed by the leading global provider of information, innovative technology solutions and contract research services focused on healthcare data. The purpose was to understand the real impact of Medisafe on users’ medication adherence.

It goes without saying that patients not taking medications as prescribed contributes to poor health outcomes and great financial cost.

Background and Methods

Medisafe is the leading personalized medication management platform with almost four and a half million users, addressing all major causes of non adherence. In order to evaluate the adherence of Medisafe users with hypertension, we compared patients’ drug adherence before and during the use of the Medisafe app using an industry standard metric of adherence called the Medication Possession Ratio (MPR).

Although there have been millions of Medisafe users, we could only do this comparison in the subset of users for whom we could match their medication claims data, which we needed to calculate the MPR. All matching was done to HIPAA-compliant privacy standards, without revealing individual patient identities. As a company, we pride ourselves in stringently maintaining the confidentiality of our users’ data, and this procedure upheld this principle.

The MPR is a ratio of days’ supply of medications obtained by the patient in a specific period of time to the number of days in that period of time. So a patient with an MPR of 1.0 (100%) would be obtaining (and presumably taking) their medication all the time as prescribed. While imperfect (more on that in a future post), MPR has become an industry standard because it is a convenient metric. And the reality is there is no perfect measurement of drug adherence.

For this study we included users ≥ 18 years-old for whom we had six months of data before starting Medisafe and while using Medisafe who were taking at least one drug to treat hypertension. Over 4000 users met criteria to be included in our analysis.

We hypothesized that the impact of Medisafe on users may vary, depending on their baseline level of adherence. We broke users down into deciles of adherence based on their MPR. So users with MPRs of 0-0.1 started out with the lowest adherence rate and those with MPRs 0.9-1.0 had the greatest adherence.


The results were striking. The absolute MPR across all users increased an average of 8 points. Even more remarkable was the changes of adherence when broken down by their adherence rates before using Medisafe.

If we say that a good adherence rate in hypertension is at least 80% (MPR ≥ 0.8), the percentage of users who had acceptable adherence increased from 71% to 86%, an absolute increase of 15% (and a statistically—and clinically—significant change).

74% of users who started with a low adherence rate of under 50% increased their adherence by 5 deciles while using Medisafe. For example, a user starting out with a 30-40% adherence rate would have increased her adherence to 80-90%.


This study shows that the use of Medisafe has a significant effect on drug adherence. The strengths of the study are the large number of diverse users included and the length of the study period.

This was a self-selected population in a retrospective study, meaning that we did not randomize people to use or not use Medisafe, so it is possible that they were already motivated to change.

But no intervention works for everyone and some patients will remain non-adherence regardless of what we try to get them to do. That said, Medisafe is working to create even greater personalization of medication adherence approaches.  

One analysis suggests that controlling blood pressure of just 10% more people with hypertension would save 14,000 lives in the US alone.

Our analysis to date only includes users with hypertension. Hypertension is generally an asymptomatic disease; patients are treated to reduce the risk of future problems, such as heart attacks, congestive heart failure, stroke, and kidney failure, but we know that people discount future probabilities. Therefore adherence rates for hypertension treatments are notoriously low.

That Medisafe could improve adherence so greatly in this population is remarkable. One analysis suggests that controlling blood pressure of just 10% more people with hypertension would save 14,000 lives in the US alone.

Improving medication adherence is one way to do that. We would anticipate an even greater improvement in adherence in other conditions, which we will be studying in the future. And higher levels of medication adherence in hypertension has been shown to reduce overall medical costs and hospitalizations.

This study provides compelling evidence that tools to improve patient medication adherence do have an impact. We will continue to study the impact of the Medisafe platform on user medication behaviors, even as we work to further personalize our approach to medication adherence.