From 19th century fiction like Pinocchio to modern science fiction like Star Trek, Artificial Intelligence (AI) has intrigued us with existential dilemmas and life-changing potential. It is unsurprising then that AI’s arrival has impacted Life Sciences so profoundly. AI has driven major advances across the Life Sciences value chain from molecule discovery to clinical trials to manufacturing. However, in an era where technology can enable not only personalized but individualized care, we should re-examine AI’s potential not only in how medications are researched, tested, and produced but also in how they are consumed; AI can help here too.
For decades, the World Health Organization has maintained that “poor adherence is the primary reason for not achieving the full health benefits medicines can provide to patients.” Adherence studies are informative, but responses to COVID-19 mask mandates and parental advice are enough for anyone to conclude that people just don’t like being told what to do. In Richard H. Thaler and Cass R. Sunstein’s “Nudge: Improving Decisions About Health, Wealth, and Happiness,” the authors note that humans are irrational and need a little help making good decisions. However, simply telling people what they need to do won’t work, even if it is in their best interest, as in the case of adherence. “Nudges” (cheap and easy interventions that one can choose to avoid) can provide a path to better adherence and better health outcomes, and AI can provide a path to better, more relevant ones. AI will need some help from consumer technology and there are challenges.
The past three decades have introduced incredible technological innovation from the internet to smartphones to the flavor of the month. As Clouds get bigger, medical devices get smaller, and AI gets “smarter,” possibilities continue to expand. The issues at this point, however, are not with technology. We have the ability to monitor biometrics with injectable devices no larger than a grain of rice, to store those readings on infrastructure of incomprehensible power, to process those massive volumes of data into insights that can help drive desirable adherence and health outcomes, and to engage patients through mobile devices, wearables, and virtual assistants. The technology is there, the trust is not. To drive adherence with AI we must:
Get predictive A startup’s AI identified an interesting correlation between people who let their cell phone batteries drop below 12% and people who default on their loans. This is just one example of how data can help drive predictive intelligence in interesting ways. Using data, we can identify patterns with out-of-compliance patients and apply those learnings to similar audiences. AI-administered adherence programs connected to smart pill boxes, and pharmacies can determine whether patients are failing to comply and can then intervene. Analysis of these patients’ medical and socioeconomic conditions can further inform intervention strategies to get involved even sooner.
Get “phygital” Acknowledging a terrible word and terrific performer, adherence nudges can be decidedly analog, like pillboxes, and decidedly digital, like reminders from Alexa, but these are not mutually exclusive. Augmented Reality remains a promising frontier and the largest User Experience Design challenge of this century. Today patients can receive detailed visual instructions on complex medical devices simply by holding their phones up to them, powered by AI-administered adherence programs adjusting as they receive new data.
Get practical Every day we understand more about AI’s limitations, including inherent bias and really dumb mistakes. Security and privacy also remain a concern— we had problems with Amazon speakers listening for action words, let alone AI predicting medication misses from our social calendars. It will be critical to strike the right balance in the shifting tides of security versus experience and privacy versus efficiency. AI-administered programs sending standard SMS text messages like “You may have forgotten to take your medication today” is one thing; home speakers blaring “I see that you’re home and haven’t taken your prescribed medication yet” is another.
Get personal The impacts of the erosion of trust brought on by misinformation and disinformation and cybercriminals are still being felt today. One of the fundamental drivers of digital has been the need for speed, but how relevant is speed for human health conditions? Failing health brings moments of trauma and in moments of trauma we seek human comfort (or a facsimile, hello teddy bears!). Vulnerable patients will need to speak with fellow humans; we need to acknowledge the limits of AI-administered adherence programs and ensure that there are real people ready to administer care to those most opposed to technology.
Get familiar and friendly Familiarity breeds trust. There’s a reason we feel great when we call our insurance agent of ten years or sit with the server that knows us by name. AI may be able to articulate a great next step for adherence, but it is likely to be unwelcome if delivered through unfamiliar means. Stick with well- accepted channels like email, text messaging, and push notifications to drive adherence, at least until the world is more comfortable speaking to its homes and ordering from kiosks. According to a recent survey, 60% of people 65 and older own a smartphone; this is the population that would appreciate a simple email or text message reminder to “Please remember to take your medication today” from an AI-administered adherence program, with guidance and guardrails as mentioned above.
The world of possibilities AI offers for driving adherence is breathtaking with its unmatched capability to ingest unstructured information like text and images and quickly arrive at relevant insights, enriched via the wealth of information we are capable of gathering from personal devices and clinical settings. But knowledge without understanding is a challenge to our shared human condition and we must continue to promote the practices that serve the basic human needs to understand and to connect with other people while we harness its power. Only then can we elevate the quality of human life—the shared ambition of both adherence and AI.
