As connectivity accelerates around the world, stakeholders in both clinical and public health see numerous opportunities to collect more comprehensive data and leverage AI to mobilize that data in service of improved health outcomes. In particular, AI and data are poised to deeply impact our understanding of social determinants of health. Social determinants of health (SDOH) – non-medical factors like education, employment, food, transportation, and access to healthcare – are understood to have a massive impact on health outcomes. If data and AI can be used to better understand the social determinants that impact community-specific health outcomes and intervene accordingly, the results could be transformative.
Health leaders are clear-eyed about the difficulties of integrating ballooning data volumes (including clinical data) in a secure system that enables concrete, insightful population-level interventions. At the same time, the imperative to solve these integration problems is undeniable. Once solutions and best practices are identified, they can be quickly scaled across numerous health departments, healthcare organizations, NGOs, and beyond.
Barriers to Data-driven SDOH Interventions
When it comes to leveraging data and AI, health leaders face barriers related to both data and analytics, including:
- Data availability. Data is heterogeneous, and clinical data is restricted to individual enterprises (healthcare systems, insurers, etc.).
- Data integration. Integrating data and architecting a shared yet fully secure data ecosystem may prove to be the most challenging aspect of unleashing AI interventions in the contest of SDOH.
- Data gaps. Geotagged data will be essential to understanding location-specific social determinants of health, but much data (including clinical data) may not be sufficiently geo-specific when it comes to concrete living arrangements and travel patterns. Similarly, SDOH data around race and ethnicity may be challenging to integrate, due to sensitivities around capturing and accessing such data.
- Data incentives. Currently, individuals have few strong incentives to enable data-sharing and self-report. But as health insurance companies explore data monetization, the incentives they develop to promote information sharing may also cascade into the public health sphere, particularly if they see monetary or reputational value in partnering with public health entities.
Opportunities for Data-driven SDOH Interventions
At least three data and AI-led approaches are likely to create value in the context of SDOH: Leveraging the combination of clinical and public health data, integrating the clinical trial ecosystem with public health imperatives, and drawing on SDOH data to drive new forms of policy-relevant research.
- Leveraging Data and Connectivity Across Clinical Care + Public Health
AI can already begin to explore SDOH within the secure confines of clinical settings. Healthcare providers can use AI to identify emerging health threats, enable early interventions, and communicate rapidly and clearly with patients. For urban populations, many of these innovations will be cost-effective and scalable, meaning that they can and will impact health outcomes across the socioeconomic spectrum, from free and urgent care clinics to large hospital systems and at-home care.
At the same time, population-level data can inform public health messaging aimed at improving social determinants of health. AI tools will give public health stakeholders the capability to deliver targeted, segmented, cohort-based behavioral nudges and improve the outcomes of those nudges in an iterative, data-driven fashion. Meanwhile, the translation capabilities of GenAI will complement AI-driven analytics, enabling public health officials to break down language barriers and effectively deliver accurately translated, culturally appropriate messaging to communities.
The technical challenge of integrating these two massive data pipelines is decreasing daily. 5G, for example, will support the seamless transfer of large medical and social datasets while also enabling efficient remote monitoring. However, new regulatory and privacy frameworks will be required to enable the most cutting-edge AI-driven interventions at the intersection of medical and social determinants of health. Public health entities will need to work with private stakeholders (including hospitals and insurers) to build and automate data pipelines while maintaining strong governance and privacy frameworks.
- Integrating the Clinical Trial Ecosystem with Public Health
Clinical trial data is an untapped resource that can also advance our understanding of SDOH, particularly as clinical trials become more geographically distributed and diverse. In the US, the DEPICT Act now mandates greater diversity in clinical trials. Life sciences companies now have an enormous opportunity to contribute to population-level health by diversifying the clinical trials that give birth to new health interventions. From a patient perspective, the metaverse and 5G are emerging as technologies that will help connect a more diverse group of patients to more seamless clinical trials. Behind the scenes, AI will also play a critical role in the emerging clinical trial landscape, particularly as it analyzes the digital biomarker data streaming in from connected medical devices and other sources.
The clinical trial process itself must continue to focus specifically on clinical outcomes. Even so, the data collected by the clinical trial ecosystem (particularly as it embraces connectivity and digital biomarkers) could, with the right regulatory frameworks, be fed back into the larger public health data ecosystem. Clinical trials may well have a role to play in refining our understanding of social determinants of health.
- Determining SDOH Cause-and-Effect
Rich SDOH datasets will increasingly enable massive longitudinal studies of defined populations. Currently, the links between urban planning, zoning, and public health are tenuous. But better data and AI-enabled analytical techniques may allow communities to consider all major projects through a data-driven public health lens. Particularly in the context of rapidly urbanizing economies, such frameworks could be essential to building health-centric mega-cities of the future.
Large development projects will present opportunities to measure natural experiments over the course of decades, for example by using wearables to track the impacts of divergent public transportation policies on neighboring cities. Frameworks for collecting the optimal data and identifying the most applicable areas for study and comparison (and even automating much of that comparative process) will enable granular insights about health determinants, and drive development and policy decisions at the city, state, and even national levels.
Toward Data and AI Strategies that Target Social Determinants of Health
The AI revolution presents an opportunity to revisit how data and technology can enable continuous data-driven health interventions, particularly in the context of social determinants of health. But AI, data, and connectivity alone won’t deliver bold transformation. Integrating this data will require creative approaches, robust data frameworks, and novel collaborations that prioritize a more holistic view of the drivers of health outcomes. If successful, this this integration will revolutionize our understanding of the relationships between medical and social determinants of health, driving hyper-targeted and hyper-intelligent interventions at scale.