How Can Mental Health Join in the Big Data Movement?

Integrated care, like all mental health care, can dramatically benefit from the strategic use of big data science. In this month’s column, Kari Stephens, Ph.D., a clinical psychologist and biomedical informaticist, challenges us to think about how we can best use big data to disseminate evidence-based mental health treatment to improve population health. —Jürgen Unützer, M.D., M.P.H.

Technology is booming with big data science discoveries using new tools and methods that permeate just about every aspect of our daily lives. Big data science is being used not just to help us buy a better car, but to monitor health in our population, from drug safety surveillance to prevent unnecessary deaths to creating robust predictive algorithms to prevent fatal incidences during hospitalizations. Yet when it comes to big data science in mental health care, we live in the dark ages. We have only primitive ways to identify and measure mental health care (that is, number and timing of visits to a mental health provider, self-report-based symptom screeners, medications prescribed), missing out on opportunities to mine and learn from the data using strategies that can create new discoveries we have not yet imagined.

The question worth asking is, how do we catch up and not miss out on the big data movement in mental health? We must mobilize our innovative minds to work alongside data scientists to create solutions to address one of mental health care’s biggest issues—delivering good, evidence-based care to the masses who would benefit most. Mental health care is best understood not just through individual patient data, but through a comprehensive lens of health care delivery.

If we are to use our most precious human-based limited resources wisely, we must understand complex patterns of care hidden within existing big datasets. We should examine risk factors to identify cases in highest need to make sure we are detecting everyone as early as possible. We should examine treatment data to make sure we are reaching patients who are suffering the most. We should evaluate outcomes to make sure patients are improving and look at functional status and patient-centered goals. We should use publicly available, geocoded datasets that allow us to look at social determinants of health (that is, socioeconomic status, environmental factors, access to community resources) to enhance what we learn and strengthen algorithms that can improve care and reach. Deep learning strategies hold promise, but we need to determine where it makes the most sense to use these tools and remember that the tools are only as good as the data themselves.

If we advocate to collect better data, we must make sure that the data offer immediate benefits to care and are feasible to collect within existing workflows. Models of integrated care for behavioral health often support measurement-based care and stepping-up care as needed, providing opportunities to enhance data collection across large populations. Primary care is primed to incorporate these big data strategies into workflow as a means to reach many people who would otherwise miss out on critical mental health treatment. As these integrated care models proliferate, we must make sure that we collect data to address gaps and measure value, positioning mental health for better discovery and success in health care reform.

Existing health care data registries and networks offer a means for systematic data collection that can drive tools that improve quality at the population and individual patient levels. The American Psychiatric Association’s quality improvement registry, PsychPRO, will leverage data collected with current workflows in electronic practice management systems and electronic health records (EHRs) as well as electronic patient-reported screening and outcomes to achieve better care. APA and PsychPRO aim to improve on the quality of data across a broad scope and breadth of mental health services. Registries like PsychPRO and other data networks (for example, PCORnet) are positioned to drive development of data-science-focused technologies that can significantly improve the population’s mental health.

It is time in mental health that we catch up and find ways to partner with technologists to help shape the use of big data to disseminate evidence-based treatment to the masses. APA is stepping out into the future as it envisions partners in technology, currently exploring ways to incorporate relevant health data from many diverse sources. As all of this progress continues to take off, we have a challenge to get educated within mental health about big data science. We must continue to create inroads with our data science colleagues, collaborations that are crucial for effectively joining the big data movement and advocating for those suffering with mental illness. ■