Telehealth may have reduced inpatient readmissions for patients with depression and enabled adherence to behavioral healthcare services during the COVID-19 emergency period. Looking forward, however, a more localized approach may be needed to ensure those who could most benefit from telehealth services are not left behind.
These are just a few among many preliminary insights datathon’ers uncovered during the 2020 DHCRC Telehealth Data Challenge, held December 1 – 8, 2020. Led by the Digital Health Cooperative Research Centre (DHCRC), an international research collaborative aiming to develop solutions for urgent healthcare industry needs, the datathon examined the intersection of telehealth, behavioral health and the COVID-19 pandemic in the U.S.
Assessing Telehealth’s Long-Term Viability
The concept of telehealth — or telemedicine — as a technology-enabled practice dates as far back as the late 1950s, when the Nebraska Psychiatric Institute and Norfolk State Hospital established a closed-circuit television connection to perform psychiatric evaluations. As the digital landscape evolved with the advent of the internet and digital health tools, the value proposition of telehealth became increasingly evident. By facilitating greater access to care for underserved populations, telehealth has the potential to improve health outcomes and, in turn, reduce unsustainable healthcare costs.
Despite its long history and promise, adoption of telehealth has historically been slow. Like many facets of healthcare, however, things changed drastically with the emergence of a global pandemic, as protecting vulnerable individuals and preserving healthcare capacity became a priority. Between March and June 2020, the Centers for Medicare and Medicaid Services (CMS) noted a 2,600% increase in telehealth services rendered to Medicaid and Children’s Health Insurance Program (CHIP) beneficiaries compared to the same period in 2019.
Emerging data indicates that, as non-emergency in-person and elective care resumed, telehealth visits began to decline, though still remained higher than pre-pandemic levels. Additionally, HMS data shows that behavioral health claims have made up the majority of telehealth visits during the COVID-19 emergency period — an intriguing phenomenon given the prevalence of cardiovascular disease, diabetes and other chronic physical health conditions in the U.S.
It is this combination of factors that led HMS’ team of data scientists, clinicians, operations staff and executives to select telehealth — and, specifically, its connection to behavioral health in light of the pandemic — as the focus of the datathon research. By better understanding how people utilized telehealth during a period of heightened stress and demand on the healthcare system, we can begin to evolve the existing technology and policy infrastructure to support its long-term viability.
Logistics and Challenges
As the lead corporate sponsor of the DHCRC and the datathon, HMS supplied the limited data set, which comprised Medicaid claims data from four states over the course of four years. HMS receives claims data from several of the healthcare plans and programs it serves, forming a HIPAA-compliant limited research database that, in its entirety, covers more than 20 million unique lives across a broad geographic range and dating back five years. For the purpose of the datathon, HMS pared down its data set to 136 million claims — a moderately manageable, yet nonetheless colossal, amount of information to parse.
Participants had just eight days to understand and validate the data set, formulate a hypothesis and learn and apply the tools provided to advance their research. Of equal weight was participants’ ability to present their findings in a comprehensible, compelling and actionable context for the panel of executive judges — HMS Chairman and CEO Bill Lucia, Chief Growth Officer Maria Perrin and President of Population Health Management Emmet O’Gara.
But the challenges didn’t end there. Most notable, perhaps, was the fact that the DHCRC and the majority of datathon participants are based in Australia, with a few in the UK. Though there are parallels between Australia’s hybrid public-private healthcare model and the U.S. healthcare system, they are vastly different, which presented a significant learning curve. This was further compounded by the need to recognize the Medicaid population as a subset of the general U.S. healthcare population. Furthermore, participants needed to understand enough about HMS as a business to make data-driven recommendations that would advance the company’s mission of making healthcare better for everyone.
Given this context, it is particularly extraordinary what the participants were able to uncover in such a short amount of time — “nothing short of miraculous,” as HMS Chief Analytics Officer and datathon mentor Jason Cooper put it.
“What a couple of these teams have accomplished — wow,” he said. “I can’t wait to share the findings and methodologies with my data science teams. The bar has officially been raised.”
The competition was divided into two main categories — a retrospective challenge to assess the impact of telehealth during the COVID-19 emergency period and a prospective challenge to understand its future implications from a population health standpoint.
- Team 5, winners of the Retrospective Challenge, examined whether telehealth was an effective intervention in reducing inpatient readmissions and curbing the “revolving door” phenomenon of outpatient to inpatient visits for patients with depression. Looking at 145 patients in two states, the team found that patients who used telehealth were more likely to remain out of inpatient care than patients who did not. This finding, they noted, “could result in significant costs savings for Medicaid, along with significant personal and societal impacts.”
- Team 7, winners of the Prospective Challenge, sought to identify groups most at risk of being left behind by telehealth and posed recommendations for healthcare providers and other key stakeholders to reduce disparities in telehealth access, focusing on a more rural state. Noting that healthcare needs and preferences vary by population, the team recommended the pursuit of localized solutions and cross-sector collaboration to enable telehealth access for those who stand to benefit most.
- Team 2, placing second in the Retrospective Challenge, studied whether telehealth was effective in maintaining or improving adherence to behavioral healthcare services during the pandemic. The team analyzed claims data from 3,320 patients from a single state with anxiety and depression who had at least one relevant telehealth appointment during the established COVID-19 timeframe. They found no significant drop in treatment adherence among patients who had had two or more in-person appointments pre-COVID-19 despite the transition to telehealth, indicating telehealth enabled continuity of care for this population.
Analysis to Innovation
Given the evolving maturity and utilization of the telehealth landscape and that the pandemic remains ongoing, the results of the datathon provide a very early, albeit critical, snapshot of the implications of telehealth’s expansion for the nation’s most vulnerable populations.
As innovators of population health management and payment accuracy solutions and advocates for a more sustainable healthcare system, HMS is continuously pursuing new ways to address the challenges of today — and strengthen our systems for the future. The impressive work of the datathon participants provides a strong basis for further analysis within HMS’ broader data sets, including social determinants of health data, to inform our product development and advocacy efforts.
Help Advance HMS & DHCRC Research. HMS’ clients are key to advancing our research with the DHCRC and our mission to make healthcare better for everyone. If you are interested in participating in our free research collaborative and gaining access to our research for your own business purposes, we invite you to learn more about our current projects and ways to get involved here.
Congratulations to the Winners of the 2020 DHCRC Telehealth Data Challenge!
- Winner: Team 5
Michael Rasmussen, Penelope Main, Sunita Bapuji, Sylvia Daravong, Alice Wilkin & Sarah Anderson, all from AHPRA (Australia Health Practitioner Regulation Agency).
- 2nd place: Team 2
Ilana solo (Bendigo Health), Keven Bennett (NSW Health), Joanna Ling (RMIT University), Mamatha Chandra Singh (University of Wollongong) and Anna Evangelista (Western Sydney University).
- Winner: Team 7
Emma Wainwright Marcel Chee Huan Yee Koh & Harvey Jia Wei Koh (Monash University), & Dr Jane Miskovic-Wheatley & Bernard Bucalon (both from University of Sydney).
Social Media Winners