Improving Predictive Outcomes of Social Determinants of Health (SDOH) Parameters with Machine Learning Techniques in Healthcare Ecosystem

Abstract

This dissertation examines transforming the healthcare system in the United
States into a value-based care paradigm, explicitly focusing on the Social Determinants
of Health (SDOH) and the obstacles impeding its implementation. The review of existing
literature reveals on-going endeavors by the government and healthcare agencies to
facilitate this transition by implementing new regulations, educating stakeholders, and
centralizing providers. However, a significant trust deficit exists, which can be attributed
to the need for more harmonization among stakeholders and the data related to SDOH,
thus hindering the smooth adoption of value-based care.
The author emphasizes the significance of collaboration among members,
providers, and payers to successfully implement value-based care. The importance of
transparent disclosures, appropriate pricing of health plans, and objective claims
assessment are identified as essential factors for attaining value-based care objectives. To
address these challenges, the dissertation proposes the integration of community
information as a means of establishing trust. Although specialized companies focusing on
value-based care are emerging, concerns arise regarding patient trust and whether
financial incentives alone are sufficient to transition from a risk-free fee-for-service
model.
Moreover, the literature review highlights the necessity of adopting a humancentric
design approach to implement SDOH. Current models often fall short in
addressing personalized healthcare programs and risk modelling, neglecting to consider
the experiences of clinicians and consumers. To address this, the author introduces a
novel disease-centric classification of SDOH to enhance the effectiveness of these
models in clinical settings and disease management.
In conclusion, the dissertation examines innovative technological initiatives as
potential solutions to the challenges of adopting value-based care. Prognostic risk
modelling, data fragmentation, consumer activation, empowerment of physicians, and
financial modelling tools are identified as key facilitators for adopting value-based care,
underscoring the crucial role of technology in improving mechanisms for data capture
and providing personalized insights.