We had a great event last week on clinical trials. A big thanks to our speakers: Carrie Northcott of Pfizer, Melissa Naylor of Takeda, and Wendy R. Sanhai of Deloitte. Also thanks to our host RightPoint Boston and our event partner the Healthcare Businesswomen’s Association.
Here are ten takeaways from the discussion:
- Digital endpoints have many advantage over traditional clinical endpoints in clinical trials. Data can be collected 24/7and can be randomized by time. Measures can be continuous which is more effective at assessing quality of life. Digital endpoints can take away the burden on the patient to make repeated clinic visits which also reduces the trial cost. Digital endpoints can also help make sure that trial participants aren’t downplaying their reported symptoms or practicing for tests in advance of clinic visits.
- Digital endpoints include data collected from a growing variety of devices. Some of the devices discussed include thermal radiography sensors and accelerometers for itching and scratching, mobile apps for measuring cognitive ability, wearables to measure the motions of Parkinson’s, image recognition of facial expressivity, tracking social phone interactions in schizophrenia, and detection of pill taking.
- New digital endpoints need to be validated in comparison to traditional clinical measures. One challenge is that the original clinical measures may not be very useful or accurate in comparison to the new digital measures. Another challenge is that the digital data sets are often huge (4 terabytes of data per study is typical). So the initial phase of a trial exists only to validate the new technology with the targeted population and a control group and won’t include drugs or other treatment.
- Digital measures can be especially valuable for conditions that involve difficult to measure cognitive changes like Alzheimers and Parkinson’s. The collection of large sets of data from multiple sensors and devices that interact with patients give hope for the discovery of more effective ways to measure early stages of the disease.
- When choosing a device for data collection, it’s important to keep in mind that most people don’t live in Boston or Silicon Valley. People may be uncomfortable using technology, may have no wireless connectivity at home, and can’t be relied upon to keep batteries charged. It’s often the best practice to choose a devices with no screen and no connectivity that can be dropped at the clinic at the end of the trial period to extract the data.
- There’s an important trend in clinical trials to become more patient centered. This is increasingly leading to sharing a participant’s data with them at the conclusion of a trial. Sharing data too soon isn’t recommended because it can lead people to change their behavior and skew the trial results.
- Security is a growing area of concern. Many technology companies are new to healthcare and don’t have experience with the rigor of a regulated environment. One example includes a technology company in the midst of a clinical trial that was infiltrated by hackers who threatened the company with ransom. Since the company had no cybersecurity policy, they lacked the means to notify trial participants of the threat to their health.
- Another rising issue is data privacy and ownership. As the big tech companies of the world become involved in healthcare and come into contact with the FDA, there will need to be changes to the way they have previously handled privacy. One way the FDA is addressing this is through their Center of Excellence for Digital pre-certification program which consists of a background check on companies to demonstrate they have a good track record.
- Biotech companies working with technology vendors have a different set of challenges. Many vendors will try to keep their algorithms proprietary as a black box which makes validating the data more difficult. In other cases, vendors will want to publish early training data that shows initial signs of success, which is misleading if the data hasn’t yet been validated. Another area of concern are the statistical impacts of software that is modified during the course of a study or trial, for example an algorithm or the underlying software OS.
- We are at a point in time when the rising costs of clinical trials are at a pace to become unsustainable. Digital endpoints offer an opportunity to bring down the cost of trials while increasing their effectiveness.
Hosted in partnership with the HBA https://www.hbanet.org/