Dementia Therapy Badly Needs Big Data and Drug Repurposing

Last Saturday (April 30, 2016) I had the honor to present at an educational session for Austrian neuropsychiatry practitioners who care for dementia patients. Somewhat contrary to the general tone of the event, which was dedicated to the challenges of daily practice and their state-of-the-art therapy, my talk rather was about the developments that are about to change the future of said daily routines in a quite disruptive way.

Dementia therapy could indeed profit from radically new options. Almost fifteen years have come and gone since the last drug approval for any type of major cognitive impairment. It was not for a lack of effort from the pharmaceutical industry. But all we got during these one and a half decades is disappointed hopes. Of course, the fact that we still do not know what actually starts the neurodegenerative process that ends in Alzheimer’s disease or frontotemporal dementia provides a good part of the explanation. However, not even symptomatically effective drugs for these primary dementias have passed the regulatory process. Not even simple chemical modifications of drugs that had been approved around the millennium have reached the market.

In my talk, I proposed that a concerted Big Data effort, based on deployment of ubiquitous personal sensors, crowdsourcing of medical and behavioral observations, and explorative data mining could provide an avenue to both basic dementia science and repurposing of existing drugs with dementia as a new therapeutic indication. Neuropsychiatrists would be the key node in a data collection and interpretation network that extends to general physicians, caregivers, and patients.

The technical basis for this is already here, or at least it is being developed. After all, a dementia consultation in ten or fifteen years from now will be very different from what psychiatrists are doing today, tracing broader changes in the physician-patient relationship that could allow a much more individualized diagnosis and therapy – if it is done right.

Whoever works with dementia patients could contribute to this data-driven and crowdsourced effort — and to a degree, even the patients themselves. Drug side effects that are currently interpreted as idiosyncratic (because they cannot be explained based on the drug’s reported pharmacological mechanisms) could provide valuable leads pointing to unknown drug-target interactions. Alterations in patient mobility and sleep patterns (which are not systematically and quantitatively recorded at present) could be hugely interesting. Dementia care facilities are an almost entirely untapped source of patient data that are simply not recorded, let alone analyzed.

Psychiatrists are deeply unsatisfied with the stagnant spectrum of dementia drugs that offer limited efficacy on the purely symptomatic level. Caregivers – mostly close relatives — are frustrated, and later overwhelmed, by behavioral symptoms for which drugs that worsen cognition are used off-label. With technology and cleverness, we can change that – not in the current decade, but during the 2020s.