F034 How are AI and wearables disrupting clinical trials?
ClinicalTrials.gov currently lists 302,091 clinical studies in the US. It is impossible for patients and their doctors to be aware of all clinical trials an individual might be eligible for. The data issues surrounding clinical trials don’t end there: how can we rely on results of trials when studies with negative results often go unpublished? How can access to the execution of trials be expanded from those with established connections to younger professionals?
Various companies are creating platforms and solutions to address these issues. German Viomedo provides better access to clinical trials for patients, Medicinisto is working on improving the way the healthcare industry is communicating with physicians, to name just two.
TriNetX is a global health research network that optimizes clinical research by connecting to hospital EHRs, anonymizes patient data and creates cohorts of specific patients, making it easier for the pharmaceutical industry to find appropriate patients for their trials. Yet, explains dr. Sam Volchenboum — a board-certified pediatric hematologist and oncologist, who is also the director of the Center for Research Informatics at the University of Chicago, data collection in clinical trials is done in a very old fashion manner.
While one would expect the trials to be run and supported by sophisticated software, the reality is often far from that expectation. The awareness of trials depends on individuals.
Patients often come to doctors inquiring about trials doctors might not even have been aware of. Trials data is managed manually on many levels, describes dr. Sam Volchenboum. Clinical trials requirements are written in a word or pdf files. Once clinical trials are running, the data are often manually collected and abstracted from clinical trials to homegrown solutions in each institution participating in a trial. Different pharma companies use different software systems and data then needs to be re-converted again to another format for FDA submissions. “This describes well, how the whole process has no standardization across the enterprise,” says dr. Volchenboum, who is also co-founder of Litmus Health — a data science platform for early-stage clinical trials, which uses data collected at the point of experience from wearables, smart devices, and home sensors to guide management and to inform trial endpoints. Apart from the obvious need for data standards, Volchenboum also believes data from trials should be more widely available to researchers, to make it easier to identify already tested ideas, hypothesis and build further assumption for treatment improvements.
The problem of manual data collection
In 2016 the University of Chicago launched a study with IBD patients, to uncover is Fitbits could help better understand of the relationship between sleep, activity and improve prediction of flares in Crohn’s and Colitis patients. Up until the study, these data points were collected through survey and manual diaries patients had to write. “I would see patients fill out their diaries in the waiting room because they did not write down data along the way. This is a terribly inefficient system producing essentially bad data,” says dr. Sam Volchenboum. Wearables might improve that and bring new metrics regarding the quality of life indicators.
Subjectivity vs. objectivity of clinical data
Getting more specific quantitative data on wellbeing might not mean patients would necessarily be treated differently, if, for example, someone would assess their pain with 7 on a scale of 1 to 10, and the data would show his pain to be 4, compared to other patients. Dr. Volchenboum believes subjectivity will stay an essential part of the quality of life assessment, however, measuring the change in performance with wearables and its comparison with disease status could bring interesting new insights. While wearables for consumers tend to be abandoned after a few months of use, their value could be seen in hospital applications and in monitoring or speeding up transition home.
The need for a systemic and regulatory push
Clearly, there are plenty of opportunities to improve clinical trials with new technologies. However, data silos need to become an archaic feature first.
Dr. Volchenboum believes three things will need to happen to support this evolution:
The pharmaceutical industry will need to acknowledge that better data collection will come from automated methods using standardized data,
Regulatory agencies will need to mandate standardization (FDA is already doing that to a certain extent),
Research organizations and foundations that support research will need to include requirements for data standardization in their funding announcements.
With rapid consumerism of healthcare, the demand for data access and control over access and sharing of data will increase from patients' side. Not all doctors will become informaticians, but everybody will need to have a basic level of data literacy, understanding of what is good data stewardship, data quality and how data is used. Dr. Volchemboum also believes EHR vendors will ultimately have to be incentivized for a better way to move data. And the system that will not interoperate, will fall behind.
Hear more in episode 34 of Faces of digital health, which you can listen to in iTunes, Podbean or Stitcher.
Some questions addressed:
Clinical trials are in dire need of reinventions. There is a gap between patients eligible for clinical trials and awareness or accessibility of those patients to participate in trials. Could you walk us through your research of the area, the competition on the market and the most innovative solutions you came across?
How promising is blockchain technology for clinical trials since it provides opportunities to improve data control, data access, and auditing, which would potentially give patients a trail of how their data is used after the trial? ATM patients can sign that they do not wish their data to be used to anything but a specific study, but they don’t really know if someone is doing something with that data or not.
TriNetX is a company the global health research network that optimizes clinical research and one of the things they do is anonymise patient data in healthcare institution giving pharma the opportunity to find potential new study subjects faster. How do healthcare institutions approach these kinds of partnerships?
In 2016 you took part in the study at the University of Chicago where 5000 patients with IBD — inflammatory bowel disease were using Fitbits to better understand the relationship between sleep, activity and flares in Crohn’s and Colitis patients. Any results you can discuss?
How far are ideas about virtual clinical trials?
In one of your presentations you mentioned that consumers try to share data with the doctor — but often doctors don’t know what to do with it — what’s the problem — lack of statistics knowledge or time…? Is it too soon to do anything with that data?
What role will wearables play in the regular clinical trials — aiming at proving the efficacy of a specific new drug, and pragmatic clinical trial (PCT), sometimes called a practical clinical trial (PCT), which refer to the treatment outcomes in the real-world? To which extent do you think we could see disruption here?
What’s your view on Apples ECG feature? Some doctors complain they get unnecessary calls from concerned healthy individuals, which takes away their time for patients.
Wearables and sensors are adding additional more objective measures to patient-reported outcomes such as pain, diet, sleep activity. What’s the future: are we going to stop rating pain on a scale of 1 to 10?