Abstract
Predictive analytics in health is a complex, transdisciplinary field requiring collaboration across diverse scientific and stakeholder groups. This study details the pilot implementation of participatory research to foster team science in predictive analytics through a partnered-symposium and funding competition. In total, 85 stakeholders were engaged across diverse translational domains, with a significant increase in perceived importance of early inclusion of patients and communities in research. Participatory research approaches may be an effective model for engaging broad stakeholders in predictive analytics.
Introduction
Predictive analytics in healthcare is inherently a "team sport," necessitating the integration of expertise from computer science, statistics, clinical medicine, and the social sciences. However, these disciplines often operate in silos, hindered by distinct terminologies and methodologies. As research increasingly utilizes sensitive data—such as electronic health records and mobile activity—to predict health outcomes, ethical concerns regarding privacy, ownership, and trust become paramount, particularly in under-resourced communities. Traditional top-down research models often fail to address these nuances. This paper explores how Community-Partnered Participatory Research (CPPR) can intentionally foster "team science" that aligns academic rigor with community priorities.
Methods and Participant Profile
The researchers adapted the CPPR model—traditionally used for community health interventions—to the planning and execution of a scientific symposium. A stakeholder advisory panel from UCLA, USC, and the University of Pittsburgh co-designed the event. The symposium featured "partnered" keynote talks and "lightning" slides to encourage co-learning among the 85 participants.
The participant pool was highly transdisciplinary, representing a wide array of techniques and principles. As shown in Figure 2, while statistics and predictive analytics were dominant, participants also drew from clinical care, health services, and community engagement. Notably, 77% of participants reported using techniques across three or more distinct areas.
To catalyze action, the organizers offered two $15,000 pilot awards prioritized for new interdisciplinary collaborations that included community or patient stakeholders.
Key Insights: Navigating Complexity and Funding
The study identified significant hurdles to effective collaboration in predictive analytics. Participants highlighted the "complexity of data and computational techniques" (61%) as the most challenging aspect of their work. Furthermore, while finding collaborators is often cited as a barrier, the survey revealed that a lack of funding for such collaborations (61%) was the primary obstacle to progress.
The pilot grants served as the necessary "glue" to solidify new interdisciplinary bonds, proving that targeted funding can overcome the inertia of siloed research and address the technical complexity inherent in the field.
Results and Impact
The participatory approach led to a measurable shift in researcher attitudes. There was a significant increase in the perceived importance of involving patients and community organizations before identifying research questions (rising from a mean of 3.67 to 4.15 on a 5-point scale; p < 0.05).
Long-term outcomes over a three-year follow-up period included:
- 8 new pairwise collaborations established across different institutions.
- 3 journal publications and 8 symposium presentations resulting directly from the pilot competition.
- Broad Translational Reach: Participants spanned the entire translational spectrum, from basic discovery (T0) to population health (T4).
Limitations and Future Directions
While successful, the study noted an academic bias, with 86% of participants hailing from research institutions. Future efforts should aim for a higher proportion of non-academic stakeholders—including patients and families—in the planning phase. Additionally, the authors suggest that "standing" participatory advisory boards might be more effective for long-term capacity building than one-off events. As analytic techniques advance, future work must also address the evolving landscape of data privacy and ethics to maintain community trust.
Summary updated