Improving patient compliance to medical device protocols with threat models
To paraphrase Lord Kelvin – “You cannot improve what you cannot measure”.
I have about 10′ before Shabbat and I wanted to offer 2 possible approaches for improving patient compliance to medical device clinical protocols.
One approach considers the patient as an attacker to the study data. This approach considers social, cost, adverse events, personal, technical and privacy aspects as study data vulnerabilities. The idea is to construct a prioritised countermeasure plan during the study and refine it with real-world data
The second approach uses a behavioral model as opposed to a threat model.
It assumes that patient compliance to a protocol in a trial will always be better than in real-life but that at the end of the day – people have various reasons sometimes not clearly known to themselves why the do not comply.
In this approach, a cost-effective strategy for assuring compliance post-marketing in the real-world uses validated machine learning models of what affected patient compliance during the controlled clinical trial. Reinforcement during the trial also reveals to the model what worked and what didn’t.
In order for a medical device company to decide what model works best for them – they must measure the movement and value of their data, and weigh that in terms of their data model.