Are the Digital Medicine Products Moving Too Fast?
(Thursday, October 21, 2021)
The rapid increase in digital medicine products is leading to flawed products that could hurt patients using them, according to a report from the Brookings Institute this week. The report states that the core technology behind such products, “affective computing” or “emotionally intelligent computing”, remains in its infancy and is being introduced improperly “without scientific validation nor public consent”. The report poses serious questions to the assumptions behind algorithms and training of the software intended to treat or manage complex clinical therapies. Digital health products rely on recognizing, expressing, and adapting to human emotions, all of which are too complex to be simplified as is needed for software design. The software behind digital products uses wearable devices and patient inputs such as physical cues, written text, and/or physiological signals to detect emotional changes. Facial expressions, speech, gait, heartbeats, and even eye blinks are the sources of data used to draw conclusions about emotional and neurological health. However, these data are often not enough to capture the diversity of human emotional experiences. On top the underlying algorithms are “often embedded with the programmer’s own cultural bias”. The generalizations of software driven outcomes ignore the variability of one population from the other. The Brookings report lists several concerns about digital health products available currently for patient health management. These include limited data extrapolated to wider population, lack of consensus among therapists and subject matter experts regarding acceptability of the emotional and physiological cues forming the basis of digital health outcomes, and low long-term use by patients owing to unreliable results, which in turn leads to absence of data on long-term effectiveness and the consequences of mental health self-treatment through digital tools. The concerns are exacerbated by the lax regulation of these products by the FDA. While this report is pretty damning on the utility of currently available and under development digital health products, it could be very helpful to developers in defining the primary concerns for these product’s acceptability by patients and medical community which in turn could help creating better programs to address such concerns for digital products with long-term potential. It should, at the minimum, be useful to create a risk-benefit profile for such products for both patients and regulators.
Dr. Mukesh Kumar
Founder & CEO, FDAMap
Linkedin: Mukesh Kumar, PhD, RAC