Social Determinants in Machine Learning Cardiovascular Disease Prediction Models: A Systematic Review

Zhao, Yuan; Wood, Erica P.; Mirin, Nicholas; Cook, Stephanie; Chunara, Rumi

doi:10.1016/j.amepre.2021.04.016

Cited by 46 publications

(35 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thirdly, the insight gathered by such community-level as well as individual-level data may by beneficial at a system level by enhancing the performance of our predictive models. A recent systematic review that examined the use of SDOH in machine learning cardiovascular risk prediction models showed increased performance in the models that included social determinants (Zhao et al, 2021). In turn, risk stratification could lead to improved clinical decision support.…”

Section: Discussionmentioning

confidence: 99%

Integration of Social Determinants of Health Data into the Largest, Not-for-Profit Health System in South Florida

Rojas¹,

Vincent²,

Parris³

2022

Journal of Data Science

View full text Add to dashboard Cite

Social determinants of health (SDOH) are the conditions in which people are born, grow, work, and live. Although evidence suggests that SDOH influence a range of health outcomes, health systems lack the infrastructure to access and act upon this information. The purpose of this manuscript is to explain the methodology that a health system used to: 1) identify and integrate publicly available SDOH data into the health systems’ Data Warehouse, 2) integrate a HIPAA compliant geocoding software (via DeGAUSS), and 3) visualize data to inform SDOH projects (via Tableau). First, authors engaged key stakeholders across the health system to convey the implications of SDOH data for our patient population and identify variables of interest. As a result, fourteen publicly available data sets, accounting for >30,800 variables representing national, state, county, and census tract information over 2016–2019, were cleaned and integrated into our Data Warehouse. To pilot the data visualization, we created county and census tract level maps for our service areas and plotted common SDOH metrics (e.g., income, education, insurance status, etc.). This practical, methodological integration of SDOH data at a large health system demonstrated feasibility. Ultimately, we will repeat this process system wide to further understand the risk burden in our patient population and improve our prediction models – allowing us to become better partners with our community.

show abstract

Section: Discussionmentioning

confidence: 99%

Integration of Social Determinants of Health Data into the Largest, Not-for-Profit Health System in South Florida

Rojas¹,

Vincent²,

Parris³

2022

Journal of Data Science

View full text Add to dashboard Cite

show abstract

“…This misestimation has been reported to occur both overall [68][69][70][71] and for subgroups defined on the basis of race/ethnicity [72][73][74], sex [68,69,75], socioeconomic status [41], or for patients with comorbidities which influence ASCVD risk or the expected benefit and harms of statin therapy, including diabetes [71,74], chronic kidney disease (CKD) [74,76,77], and rheumatoid arthritis (RA) [78,79]. Approaches undertaken to address these issues include the development of new risk estimators from large, diverse observational cohorts using modern machine learning methods [10,67,[80][81][82], revisions to guidelines to encourage follow-up testing when the benefits of statin therapy are unclear and shared patient-clinician decision-making to incorporate patient preferences and other context [41], and the incorporation of fairness constraints into the model development process [9,10,15].…”

Section: Algorithmic Fairness Training Objectivesmentioning

confidence: 99%

Net benefit, calibration, threshold selection, and training objectives for algorithmic fairness in healthcare

Pfohl¹,

Xu²,

Foryciarz³

et al. 2022

Preprint

View full text Add to dashboard Cite

A growing body of work uses the paradigm of algorithmic fairness to frame the development of techniques to anticipate and proactively mitigate the introduction or exacerbation of health inequities that may follow from the use of model-guided decision-making. We evaluate the interplay between measures of model performance, fairness, and the expected utility of decision-making to offer practical recommendations for the operationalization of algorithmic fairness principles for the development and evaluation of predictive models in healthcare. We conduct an empirical case-study via development of models to estimate the ten-year risk of atherosclerotic cardiovascular disease to inform statin initiation in accordance with clinical practice guidelines. We demonstrate that approaches that incorporate fairness considerations into the model training objective typically do not improve model performance or confer greater net benefit for any of the studied patient populations compared to the use of standard learning paradigms followed by threshold selection concordant with patient preferences, evidence of intervention effectiveness, and model calibration. These results hold when the measured outcomes are not subject to differential measurement error across patient populations and threshold selection is unconstrained, regardless of whether differences in model performance metrics, such as in true and false positive error rates, are present. In closing, we argue for focusing model development efforts on developing calibrated models that predict outcomes well for all patient populations while emphasizing that such efforts are complementary to transparent reporting, participatory design, and reasoning about the impact of model-informed interventions in context.

show abstract

“…38 While provider-level decisions may be due to inferring certain characteristics or prioritizing information needed for them to direct care, completeness rates are important for decision support tool performance, which can improve with social determinants of health information. 39,40 When assessing data quality and completeness, which is emphasized by machine learning for EHR guidelines, 2,4,19,41,42 the implications of pursuing LHS initiatives at different levels should also be considered. For example, a subset of CHCs capture self-reported measures of health, which are valuable research outcomes.…”

Section: Sociodemographic Characteristicsmentioning

confidence: 99%

“…These findings align with a framework to assess selection bias in EHR data that suggests multiple mechanisms are usually responsible for missingness so the focus should be on "what data are observed [instead of missing] and why?" [38] While provider-level decisions may be due to inferring certain characteristics or prioritizing information needed for them to direct care, completeness rates are important for decision support tool performance, which can improve with social determinants of health information [39,40]. When assessing data quality and completeness, which is emphasized by machine learning for EHR guidelines [2,4,19,41,42], the implications of pursuing LHS initiatives at different levels should also be considered.…”

Section: Sociodemographic Characteristicsmentioning

confidence: 99%

Describing a complex primary health care population in a learning health system to support future decision support and artificial intelligence initiatives

Kueper

Rayner

Zwarenstein

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: Learning health systems (LHS) use data to improve care. Descriptive epidemiology to reveal health states and needs of the LHS population is essential for informing LHS initiatives. To properly characterize complex populations, both simple statistical and artificial intelligence techniques, applied with an epidemiological lens, can be useful. We present the first large-scale description of the population served by one of the first primary care LHS in Canada. Methods: We use electronic health record data from 2009-2019 to describe sociodemographic, clinical, and health care use characteristics of adult primary care clients served by the Alliance for Healthier Communities. In addition to simple summary statistics, we apply unsupervised leaning techniques to explore patterns of common condition co-occurrence, care provider teams, and care frequency. Results: There are 221 047 eligible clients. Clients at community health centres that primarily serve those most at risk (homeless, mental health, addictions) and clients with multimorbidity tended to have more social determinants of health and chronic conditions. Most care is provided by physician and nursing providers, with heterogeneous combinations of other provider types. A subset of clients have many issues addressed within single-visits and there is notable within- and between-client variability in care frequency. Conclusions: This population-level overview of clients served by the Alliance provides a foundation for future LHS initiatives. In addition to substantive findings, we demonstrate the use of methods from statistics and artificial intelligence to describe a complex primary care population. We discuss implications for future initiatives, including development of decision support tools.

show abstract

Social Determinants in Machine Learning Cardiovascular Disease Prediction Models: A Systematic Review

Cited by 46 publications

References 54 publications

Integration of Social Determinants of Health Data into the Largest, Not-for-Profit Health System in South Florida

Integration of Social Determinants of Health Data into the Largest, Not-for-Profit Health System in South Florida

Net benefit, calibration, threshold selection, and training objectives for algorithmic fairness in healthcare

Describing a complex primary health care population in a learning health system to support future decision support and artificial intelligence initiatives

Contact Info

Product

Resources

About