In a democratic society, elections are held up as powerful indicators of the will of the people.
But, strictly speaking, that it is not the case. In fact, elections indicate the will of some of the people – namely, those who turned up at the polling station to cast their vote. And, as recent national elections have shown, that turnout figure can vary significantly, from the lowly 59.7% registered in the UK in 2024 to the far more robust figure of 87.4% recorded in Belgium the same year.[1]
Although from an official perspective the voices of non-voters go unheard, that is not to say their thoughts lack validity or value. Indeed, it could be said that election outcomes paint an incomplete picture of the true situation at hand.
Parallels can be drawn here with clinical trials when it comes to the measurement of pharmacokinetic (PK) data. There is an underlying assumption that this well-established metric will provide a reliable and complete picture of the body’s effect on a drug across the entire trial population. And yet the reality is that trials frequently only reflect the PK data elicited from a proportion of the participants, while valuable data from a selection of others is routinely dismissed.
At the heart of this troubling issue are legacy approaches to the measurement of dosing. Without acknowledging, addressing and mitigating against this risk, sponsors face the prospect of failed trials and lost investment as well as delays or even possibly the abandonment of drug development programmes.
Let us start with some context. Currently, in typical trial environments, PK measurements are taken from candidates at scheduled intervals in a hospital setting when they start receiving doses of a drug in development. This phase establishes a benchmark understanding of blood-drug concentration levels over time, effectively setting the shape of the patient’s exposure curve.
After leaving hospital, candidates then provide blood samples at regular, designated visits. This population PK analysis establishes whether levels remain in the ‘corridor’ of expected values, as per the model. Measurements in this phase are taken at the blood-concentration trough to ensure the potential for variability is reduced.
However, it is not uncommon for some unmonitored candidates – who claim to be faithfully taking the trial medication in compliance with the prescribed dosing regimen – to register extremely low blood concentration levels. Indeed, in some cases, the PK measurements indicate that no drug is present in the blood plasma at all.
In this scenario, logic dictates that such data must be flagged as invalid, after all, the zero PK score is directly at odds with what is believed to be adherent behaviour. As an extension, the data must be excluded from the study, since fitting a non-linear PK model will not converge in the presence of non-conforming outlying points of data, which are deemed to be extreme discrepancies. In effect, these data points are sacrificed to safeguard the integrity of the PK model, with the resulting analysis focused on compliant exposure data from the remaining trial candidates.
But to discard this non-conforming data is to disregard the truth of the situation. In reality, the mismatch between weak drug exposure (scientifically observed) and strong medication compliance (either self-reported or inferred) can only come down to one thing: poor adherence. And failure to expose this reality comes down to a misplaced trust in archaic methods of adherence monitoring.
Research has illuminated the extent of this issue. Looking at multiple studies for psychiatry treatments across a ten-year period, researchers found that, on average, 29% of trial subjects registered a sample with PK levels below the limit of quantification for the study drug, indicating widespread non-adherence. But despite this evidence, pill count data across the trials suggested that compliance was extremely high.[2] This led the authors to conclude that “pill counts and self-report greatly overestimate medication adherence relative to PK sampling”.
In a separate study, across three high-recruiting sites, just over half (51%) of participants in a trial showed no detectable concentrations of the test drug.[3]
In these cases, any decision to remove conflicted PK data points from the analysis certainly eliminates doubt over participant exposure to the drug in question. However, when non-adherent patients are excluded from the analysis, valuable opportunities are lost. Each candidate, having been recruited at significant cost and effort, will disappear as a source of information on pharmacokinetics and pharmacodynamics. Without them, key aspects of the overall picture being created remain obscured and the true story of the trial is unlikely to emerge.
If failure ensues, this can prompt reflection and subsequent interrogation of non-adherence as a major contributing cause. However, this is likely to bring little solace in the face of sunk investments and damaging programme delays.
An entirely different scenario plays out when co-ordinators are able to trust in an independent, digital measure of adherence. Here, the ability to continually track patient compliance behaviours with far higher levels of reliability provides the necessary platform to elicit evidence of exposure with far greater levels of accuracy. In turn, this provides the means to spotlight, record and explain outlying or non-existent PK values, and to take appropriate intervention measures as the trial unfolds. The net result is a more complete, more robust and more valuable dataset, which feeds through into superior analysis while avoiding the flaws and issues that can trigger trial failures.
This is where the earlier parallel with elections has the potential to diverge. While there is unlikely ever to be a mechanism for counting the valuable views of non-voters, technologies such as MEMS from AARDEX offer a practical, accessible and impactful mechanism for addressing non-adherence in clinical trials today. Digital monitoring ensures all candidates participate to their full potential, data collection and analysis is fully optimised, and the therapy under investigation is understood to its fullest extent.
Establishing hard evidence for your trial…
AARDEX has a best practice methodology, independent of any device package or software platform. Utilising our expertise and experience in medication adherence and patient compliance we acquire, monitor, analyse, guide and interpret data to deliver absolute clarity and bring confidence to sponsors, trialists, and ultimately, patients.
AARDEX is the only mature, robust and proven adherence solution on the market today, one that maximises the reward, mitigates the risk and delivers resolution for your clinical trial. All delivered with the clarity, integrity and certainty you need to proceed with complete confidence in the exposure-response.
[1] https://ourworldindata.org/grapher/voter-turnout-of-registered-voters
