Measuring one’s own life has already become part of everyday life for many people – so why would we not also make use of this opportunity in clinical trials? For years, I have encouraged the inclusion of sensors or applications in study designs, at least on an exploratory basis. One would think that by following the everyday life of a study participant, it should at least be possible to see whether life activity increases during treatment with the medicine compared with the control group. After all, digital monitoring turns into objective numbers what used to be asked of the patient, with the recorded information based on memory or opinion.
If only everything were that simple. Over the years, my enthusiasm has been challenged by many highly relevant questions: What does the sensor measure? What factors could produce data that look like what we want to measure and thereby confound the results? Does the sensor function technically as intended? Does it calculate correctly what is being measured? How will the devices be delivered to patients? Can the patient use the device? Will they remember to charge it? Will the network work? How do we analyse the vast amount of collected data and find exactly what we are looking for? What happens to the data when the underlying algorithm is updated? And above all, what do the results mean from the patient’s perspective – are we truly measuring things that matter?
Fortunately, regulators, companies and various expert consortia have also been considering these same challenges for a long time. The need for new objective measures is obvious, and the closer we get to the patient, the more real the results become. Over the years, regulatory guidance has become increasingly clear – although even these documents have not provided simple solutions to the problems; in practice, the requirements have often been impossible to fulfil. For example, how can a new measure be validated against a previous standard if no such standard exists, and the problem has previously been measurable only by asking for the opinion of the treating physician or the patient?
A couple of weeks ago, the U.S. Food and Drug Administration (FDA) published a document (Key Considerations) that places clearer emphasis than before on case-by-case and patient-centred thinking. Validation of a measure is, of course, extremely important both analytically (that is, whether the right thing is being measured in the right way) and clinically (whether an important patient problem is being measured, and whether a change in it meaningfully improves the patient’s life). In validation, it is now possible to use, for example, several other data sources that point in the same direction – a gold standard for comparison is not required. In several places, the document encourages direct discussion with FDA – and new mechanisms for rapid discussion and questions are also available and under development.
Clinical validation is carried out with patients – what truly burdens them in life, and what would they hope medication could affect? Shared discussion leads experts to conclusions about the quantities to be measured. And these, too, can be discussed with the regulator in good time – while there is still time to correct the study plan.
Last week, FDA held an open workshop related to the document, which is also available as a YouTube recording. The excellent presentations and panel discussions, with representatives from FDA, device manufacturers, the pharmaceutical industry and academia, reinforced the feeling that this is finally being taken seriously. Regulators, too, understand the practical challenges and are genuinely trying to help device manufacturers and industry move forward with digital measurement. It was also noted that the greatest benefit of digital measurement lies precisely in areas where adequate measures do not yet exist – and where there is no gold standard. New case-by-case approaches to validation are much needed.
According to FDA’s calculations, only 12% of investigational applications submitted to them include digital measures (DHT, digital health technology). One might expect the figure to be higher, given that guidance on the topic has existed for years. FDA wants the number to increase – not because clinical trials should use modern measurement methods simply because they exist, but because they can capture problems in patients’ lives for which no other good measures are available.
The pharmaceutical industry and research groups need the courage to jump in – validation of a new measure requires parallel research and a completely new way of thinking alongside and instead of the old. In my view, the benefit within the next few years for the operational execution of clinical trials, patient-centredness, and the monitoring of product efficacy and safety will be so substantial that digital measurement should be seriously considered in every clinical programme.
Picture created by OpenAI. Text without use of AI.