In this article, Dr. Bernard Vrijens discusses the importance of medication adherence data in clinical trials and why medication adherence should be a calculation, not an estimation.
The world is today collecting a mind-blowing amount of data. Indeed, IDC calculates that about two zettabytes (ZB) of digital information was generated in 2010 , a figure dwarfed by estimates that in 2020 that number catapulted to 44ZB .
When it comes to clinical trials, the volume of data generated is equally staggering. In fact, one study reports that phase 3 clinical trials generate an average of 3.6 million data points, roughly three times the volume of data collected by late-stage studies a decade ago .
Almost every aspect of our health can now be captured in dots and dashes – from monitoring glucose levels to adherence to medication via connected drug delivery devices. And yet, when it comes to an individual’s adherence to a drug regimen during clinical trials, while there is an appetite for collecting and interpreting meaningful information, the adoption of tools to effectively monitor adherence has been slow.
Interpreting Medication Adherence Data
As we have discussed before, there are plenty of legacy reasons why electronic adherence monitoring has been slow to be adopted. Perhaps one of the objections is the perceived burden of managing and interpreting the medication adherence data.
Harnessing a robust analytics solution, like MEMS AS® Medication Adherence Software, addresses those concerns. Sponsors and trial managers don’t need to do the data science bit – it’s all done for them, with the medication adherence data interrogated already in a system specifically designed for the particular data sets that enhance clinical trial assessments.
Imagine if you had, for example, a panel of 100 patients on a trial for 100 days with just a single daily dose – even then, you end up with 10,000 data points. That would require a considerable amount of time and resources to process what is a comparatively small amount of data using a standard tool. Quite simply, the complexity would be too great because the data isn’t standard, and as we know, round pegs don’t fit in square holes. What is really required is an initiative, simple, and fit-for-purpose solution, such as MEMS AS®.
With MEMS AS®, the data is nicely formatted, enabling all stakeholders to visualize data at a granular level of patient, study, and site to quickly and intuitively gain accurate data-led insight rather than making assumptive decisions.
With so many data points, clients can utilize the MEMS AS® platform to compute a key features summary for each patient – a profile that can be built up over the multiple years of a study’s duration. It also means that investigators can identify patients with anomalous or even dangerous behaviors, such as overdosing, and deliver help quickly.
But of course, data is only helpful if it can be understood and interpreted effectively. The specific types of data and visualizations delivered are different from any other kinds of day-to-day data investigators are used to, and so some assistance with interpretation is often helpful, particularly as users can select how they wish to visualize their medication adherence data.
Collecting data is fine, but then what?
“Without data, you’re just another person with an opinion’, W. Edwards Deming”
In all types of businesses, a great deal of time and energy is invested in collecting data, but very often, much less effort is put into deciding what to do with it. When it comes to adherence in clinical trials, there are several perspectives on its utility, given that electronic measurement is not yet mandatory.
Each function within a trial with has a different requirement for any intelligence, but first and foremost, it must be to enable and encourage patients to better adhere to their medication. As we have discussed before, medication adherence data doesn’t lie, and very often, the presentation of facts can be the catalyst for change over any emotive arguments, even though its fair to say from time to time, we can all reject facts because they don’t fit with our views of reality.
Or perhaps we may, sometimes, manipulate the truth. Do people really tell their doctor the truth about the number of units of alcohol they consume?
The reality is, of course, that to make truly insightful decisions, you must base them on facts, not intuition or gut feeling. Indeed, even in a clinical setting, it has been found that physicians’ ability to predict their patient’s adherence is at best limited, with a least 50% of their predictions being incorrect .
By using medication adherence data effectively, we can mitigate against some of the reasons or excuses for poor adherence and start pre-empting some common challenges with trials. For example, we know interest and motivation decline during the course of the study, so we use adherence insights to avoid interruption or discontinuation of dosing.
And, as the trial continues, review meetings can be truly personalized, highlighting the successes and areas for improvement in adherence for every patient.
A Trial is a Formal Examination of Evidence, not an Irrational Reaction to Beliefs
The data derived from clinical trial adherence studies can and should be used to enhance the validity of the outcomes, supporting both investigator and sponsor by ensuring the patient has the best possible chance of success by briefing them with insight.
We can monitor their performance and encourage them to ‘stick or twist’ based on their behavior. And we deliver robust, data-led results to enable successful regulatory submission.