scholarly journals Precision Health–Enabled Machine Learning to Identify Need for Wraparound Social Services Using Patient- and Population-Level Data Sets: Algorithm Development and Validation (Preprint)

2019 ◽  
Author(s):  
Suranga N Kasthurirathne ◽  
Shaun Grannis ◽  
Paul K Halverson ◽  
Justin Morea ◽  
Nir Menachemi ◽  
...  
Keyword(s):  
Machine Learning ◽  
Social Determinants Of Health ◽  
Social Services ◽  
Performance Measures ◽  
Social Determinants ◽  
Population Level ◽  
Decision Models ◽  
Data Sets ◽  
Level Data ◽  
Precision Health

BACKGROUND Emerging interest in precision health and the increasing availability of patient- and population-level data sets present considerable potential to enable analytical approaches to identify and mitigate the negative effects of social factors on health. These issues are not satisfactorily addressed in typical medical care encounters, and thus, opportunities to improve health outcomes, reduce costs, and improve coordination of care are not realized. Furthermore, methodological expertise on the use of varied patient- and population-level data sets and machine learning to predict need for supplemental services is limited. OBJECTIVE The objective of this study was to leverage a comprehensive range of clinical, behavioral, social risk, and social determinants of health factors in order to develop decision models capable of identifying patients in need of various wraparound social services. METHODS We used comprehensive patient- and population-level data sets to build decision models capable of predicting need for behavioral health, dietitian, social work, or other social service referrals within a safety-net health system using area under the receiver operating characteristic curve (AUROC), sensitivity, precision, F1 score, and specificity. We also evaluated the value of population-level social determinants of health data sets in improving machine learning performance of the models. RESULTS Decision models for each wraparound service demonstrated performance measures ranging between 59.2%% and 99.3%. These results were statistically superior to the performance measures demonstrated by our previous models which used a limited data set and whose performance measures ranged from 38.2% to 88.3% (behavioural health: F1 score <i>P</i>&lt;.001, AUROC <i>P</i>=.01; social work: F1 score <i>P</i>&lt;.001, AUROC <i>P</i>=.03; dietitian: F1 score <i>P</i>=.001, AUROC <i>P</i>=.001; other: F1 score <i>P</i>=.01, AUROC <i>P</i>=.02); however, inclusion of additional population-level social determinants of health did not contribute to any performance improvements (behavioural health: F1 score <i>P</i>=.08, AUROC <i>P</i>=.09; social work: F1 score <i>P</i>=.16, AUROC <i>P</i>=.09; dietitian: F1 score <i>P</i>=.08, AUROC <i>P</i>=.14; other: F1 score <i>P</i>=.33, AUROC <i>P</i>=.21) in predicting the need for referral in our population of vulnerable patients seeking care at a safety-net provider. CONCLUSIONS Precision health–enabled decision models that leverage a wide range of patient- and population-level data sets and advanced machine learning methods are capable of predicting need for various wraparound social services with good performance.

scholarly journals Precision Health–Enabled Machine Learning to Identify Need for Wraparound Social Services Using Patient- and Population-Level Data Sets: Algorithm Development and Validation

10.2196/16129 ◽  
2020 ◽  
Vol 8 (7) ◽  
pp. e16129 ◽  
Author(s):  
Suranga N Kasthurirathne ◽  
Shaun Grannis ◽  
Paul K Halverson ◽  
Justin Morea ◽  
Nir Menachemi ◽  
...  
Keyword(s):  
Machine Learning ◽  
Social Determinants Of Health ◽  
Social Services ◽  
Performance Measures ◽  
Social Determinants ◽  
Population Level ◽  
Decision Models ◽  
Data Sets ◽  
Level Data ◽  
Precision Health

Background Emerging interest in precision health and the increasing availability of patient- and population-level data sets present considerable potential to enable analytical approaches to identify and mitigate the negative effects of social factors on health. These issues are not satisfactorily addressed in typical medical care encounters, and thus, opportunities to improve health outcomes, reduce costs, and improve coordination of care are not realized. Furthermore, methodological expertise on the use of varied patient- and population-level data sets and machine learning to predict need for supplemental services is limited. Objective The objective of this study was to leverage a comprehensive range of clinical, behavioral, social risk, and social determinants of health factors in order to develop decision models capable of identifying patients in need of various wraparound social services. Methods We used comprehensive patient- and population-level data sets to build decision models capable of predicting need for behavioral health, dietitian, social work, or other social service referrals within a safety-net health system using area under the receiver operating characteristic curve (AUROC), sensitivity, precision, F1 score, and specificity. We also evaluated the value of population-level social determinants of health data sets in improving machine learning performance of the models. Results Decision models for each wraparound service demonstrated performance measures ranging between 59.2%% and 99.3%. These results were statistically superior to the performance measures demonstrated by our previous models which used a limited data set and whose performance measures ranged from 38.2% to 88.3% (behavioural health: F1 score P<.001, AUROC P=.01; social work: F1 score P<.001, AUROC P=.03; dietitian: F1 score P=.001, AUROC P=.001; other: F1 score P=.01, AUROC P=.02); however, inclusion of additional population-level social determinants of health did not contribute to any performance improvements (behavioural health: F1 score P=.08, AUROC P=.09; social work: F1 score P=.16, AUROC P=.09; dietitian: F1 score P=.08, AUROC P=.14; other: F1 score P=.33, AUROC P=.21) in predicting the need for referral in our population of vulnerable patients seeking care at a safety-net provider. Conclusions Precision health–enabled decision models that leverage a wide range of patient- and population-level data sets and advanced machine learning methods are capable of predicting need for various wraparound social services with good performance.

scholarly journals Influence of social determinants of health and county vaccination rates on machine learning models to predict COVID-19 case growth in Tennessee

2021 ◽  
Vol 28 (1) ◽  
pp. e100439
Author(s):  
Lukasz S Wylezinski ◽  
Coleman R Harris ◽  
Cody N Heiser ◽  
Jamieson D Gray ◽  
Charles F Spurlock
Keyword(s):  
Machine Learning ◽  
Risk Factors ◽  
Social Determinants Of Health ◽  
Social Determinants ◽  
Determinants Of Health ◽  
Learning Models ◽  
Vaccination Rates ◽  
Data Framework ◽  
The Impact ◽  
Machine Learning Models

IntroductionThe SARS-CoV-2 (COVID-19) pandemic has exposed health disparities throughout the USA, particularly among racial and ethnic minorities. As a result, there is a need for data-driven approaches to pinpoint the unique constellation of clinical and social determinants of health (SDOH) risk factors that give rise to poor patient outcomes following infection in US communities.MethodsWe combined county-level COVID-19 testing data, COVID-19 vaccination rates and SDOH information in Tennessee. Between February and May 2021, we trained machine learning models on a semimonthly basis using these datasets to predict COVID-19 incidence in Tennessee counties. We then analyzed SDOH data features at each time point to rank the impact of each feature on model performance.ResultsOur results indicate that COVID-19 vaccination rates play a crucial role in determining future COVID-19 disease risk. Beginning in mid-March 2021, higher vaccination rates significantly correlated with lower COVID-19 case growth predictions. Further, as the relative importance of COVID-19 vaccination data features grew, demographic SDOH features such as age, race and ethnicity decreased while the impact of socioeconomic and environmental factors, including access to healthcare and transportation, increased.ConclusionIncorporating a data framework to track the evolving patterns of community-level SDOH risk factors could provide policy-makers with additional data resources to improve health equity and resilience to future public health emergencies.

scholarly journals Addressing Social Needs in Health Care Settings: Evidence, Challenges, and Opportunities for Public Health

2020 ◽  
Vol 42 (1) ◽  
Author(s):  
Matthew W. Kreuter ◽  
Tess Thompson ◽  
Amy McQueen ◽  
Rachel Garg
Keyword(s):  
Public Health ◽  
Health Care ◽  
Social Determinants Of Health ◽  
Social Services ◽  
Social Determinants ◽  
Safety Net ◽  
Social Needs ◽  
Determinants Of Health ◽  
Health Improvement ◽  
Health Care Settings

There has been an explosion of interest in addressing social needs in health care settings. Some efforts, such as screening patients for social needs and connecting them to needed social services, are already in widespread practice. These and other major investments from the health care sector hint at the potential for new multisector collaborations to address social determinants of health and individual social needs. This article discusses the rapidly growing body of research describing the links between social needs and health and the impact of social needs interventions on health improvement, utilization, and costs. We also identify gaps in the knowledge base and implementation challenges to be overcome. We conclude that complementary partnerships among the health care, public health, and social services sectors can build on current momentum to strengthen social safety net policies, modernize social services, and reshape resource allocation to address social determinants of health. Expected final online publication date for the Annual Review of Public Health, Volume 42 is April 2021. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

scholarly journals Differences in COVID-19 Risk by Race and County-Level Social Determinants of Health among Veterans

2021 ◽  
Vol 18 (24) ◽  
pp. 13140
Author(s):  
Hoda S. Abdel Magid ◽  
Jacqueline M. Ferguson ◽  
Raymond Van Cleve ◽  
Amanda L. Purnell ◽  
Thomas F. Osborne
Keyword(s):  
Social Determinants Of Health ◽  
Social Determinants ◽  
Linear Models ◽  
Determinants Of Health ◽  
Health Concern ◽  
County Level ◽  
Department Of Veterans Affairs ◽  
Optimal Care ◽  
Level Data ◽  
Significant Public Health

COVID-19 disparities by area-level social determinants of health (SDH) have been a significant public health concern and may also be impacting U.S. Veterans. This retrospective analysis was designed to inform optimal care and prevention strategies at the U.S. Department of Veterans Affairs (VA) and utilized COVID-19 data from the VAs EHR and geographically linked county-level data from 18 area-based socioeconomic measures. The risk of testing positive with Veterans’ county-level SDHs, adjusting for demographics, comorbidities, and facility characteristics, was calculated using generalized linear models. We found an exposure–response relationship whereby individual COVID-19 infection risk increased with each increasing quartile of adverse county-level SDH, such as the percentage of residents in a county without a college degree, eligible for Medicaid, and living in crowded housing.

scholarly journals Using ICD-10-based Social Determinants of Health Categories to Assess Patients Risk for Acute Care Utilization

2021 ◽  
Author(s):  
Peter H. Nguyen ◽  
James Wang ◽  
Pamela Garcia-Filion ◽  
Deborah Dominick ◽  
Hamed Abbaszadegan ◽  
...  
Keyword(s):  
Social Determinants Of Health ◽  
Social Services ◽  
Acute Care ◽  
Social Determinants ◽  
Hospital Admissions ◽  
Medical Center ◽  
Determinants Of Health ◽  
Care Utilization ◽  
Icd 10 ◽  
Acute Care Utilization

ABSTRACTObjectiveSocial determinants of health (SDoH) play a pivotal role in health care utilization and adverse health outcomes. However, the optimal method to identify SDoH remains debatable. We ascertained SDoH based on International Classification of Disease 10 (ICD-10) codes in patient electronic health records (EHR) to assess the correlation with acute care utilization, and determine if social services interventions reduced care utilization.MethodsWe analyzed retrospective data for active patients at a Department of Veterans Affairs Medical Center (VAMC) from 2015-2017. Eleven categories of SDoH were developed based on existing literature of the social determinants; the relevant ICD-10 codes were divided among these categories. Emergency Room (ER) visits, hospital admissions, and social work visits were determined for each patient in the cohort.ResultsIn a cohort of 44,401 patients, the presence of ICD-10 codes within the EHR in the 11 SDoH categories was positively correlated with increased acute care utilization. Veterans with at least one SDoH risk factor were 71% (95%CI: 68% - 75%) more likely to use the ED and 71% (95%CI: 65%-77%) more likely to be admitted to the hospital. Utilization decreased with social service interventions.ConclusionThis project demonstrates a potentially meaningful method to capture patient social risk profiles through existing EHR data in the form of ICD-10 codes, which can be used to identify the highest risk patients for intervention with the understanding that not all SDoH codes are uniformly used and some SDoHs may not be captured.

Association of area social determinants of health with prostate cancer mortality.

2020 ◽  
Vol 38 (6_suppl) ◽  
pp. 49-49
Author(s):  
John Clark Henegan
Keyword(s):  
Prostate Cancer ◽  
Linear Regression ◽  
Social Determinants Of Health ◽  
Social Determinants ◽  
Cancer Mortality ◽  
County Level ◽  
Clinical Factors ◽  
Independent Variables ◽  
Prostate Cancer Mortality ◽  
Level Data

49 Background: Validated clinical nomograms prognosticate survival for men with metastatic castrate resistant prostate cancer. In addition to clinical factors, significant disparities in survival in advanced prostate cancer are associated with social determinants of health (SDH). We performed a study of county-level data to evaluate SDH associated with prostate cancer mortality. Methods: County-level data regarding prostate cancer crude mortality rate (dependent variable) and crude incidence rate was extracted from the ten states with information at www.cancer-rates.info for the years 2006-2015. The 2014 Robert Wood Johnson County Health Rankings dataset was used for information regarding county demographics, clinical information, and available SDH(independent variables). A correlation analysis was performed between independent variables and only one variable was analyzed if a criterion of a correlation of ≥ 0.65 was present. All counties with ≤ 30 death in the specified time period or with incomplete independent variable information were excluded from analysis. A univariate linear regression analysis was performed. All individual independent variables with a statistically significant (p ≤ 0.05) association with the dependent variable were included in a multivariate linear regression model. Results: 275 counties were available for analysis. On multivariate analysis, independent variables statistically significantly associated with the crude rate of prostate cancer mortality included household income and the percentage of: a population ≥ 65 years of age, the population < 18 years of age, African Americans, the population classified as rural, women reporting a screening mammography, smokers, and uninsured patients. The multivariate model was associated with a R-squared of 0.62. Conclusions: In addition to differences in baseline clinical factors, prostate cancer quality metrics that evaluate survival should account SDH. Information from this study, if validated in subsequent work, could be used to develop a stepwise model to adjust for both clinical factors and SDH when measuring survival as a quality metric.

First Nations Community Well-Being Research and Large Data Sets: A Respectful Caution

2017 ◽  
Vol 12 (2) ◽  
pp. 15-24 ◽  
Author(s):  
Alexandra S. Drawson ◽  
Aislin R. Mushquash ◽  
Christopher J. Mushquash
Keyword(s):  
First Nations ◽  
Social Determinants ◽  
Large Data ◽  
Well Being ◽  
Large Data Sets ◽  
Data Sets ◽  
Nations People ◽  
Level Data ◽  
Traditional Language ◽  
First Nations People

Health researchers are increasingly encouraged to use large, community-level data sets to examine factors that promote or diminish health, including social determinants. First Nations people in Canada experience disparity in a range of social determinants of health that result in relatively low community well-being scores, when compared to non-First Nations people. However, First Nations people also possess unique protective factors that enhance well-being, such as traditional language usage. Large data sets offer First Nations a new avenue for advocating for supports and services to decrease health inequity while developing culture-based evidence. However, care must be taken to ensure that these data are interpreted appropriately. In this paper, we respectfully offer a cautionary note on the importance of understanding culture and context when conducting First Nations health research with large data sets. We have framed this caution through a narrative presentation of a simple and concrete example. We then outline some approaches to research that can ensure appropriate development of research questions and interpretation of research findings.

scholarly journals Population-Level Data on Child Development at School Entry Reflecting Social Determinants of Health: A Narrative Review of Studies Using the Early Development Instrument

2021 ◽  
Vol 18 (7) ◽  
pp. 3397
Author(s):  
Magdalena Janus ◽  
Caroline Reid-Westoby ◽  
Noam Raiter ◽  
Barry Forer ◽  
Martin Guhn
Keyword(s):  
Early Childhood ◽  
Social Determinants Of Health ◽  
Social Determinants ◽  
Early Childhood Development ◽  
Developmental Outcomes ◽  
Population Level ◽  
Determinants Of Health ◽  
School Entry ◽  
Childhood Development ◽  
Developmental Health

Background: The Early Development Instrument (EDI) was developed as a population-level assessment of children’s developmental health at school entry. EDI data collection has created unprecedented opportunities for population-level studies on children’s developmental outcomes. The goal of this narrative review was to synthesize research using the EDI to describe how it contributes to expanding the understanding of the impacts of social determinants on child development and how it applies to special populations. Methods: Select studies published in peer-reviewed scientific journals between 2015 and 2020 and incorporating the social determinants of health perspectives were chosen to highlight the capability of the EDI to monitor children’s developmental health and contribute knowledge in the area of early childhood development. Results: A number of studies have examined the association between several social determinants of health and children’s developmental outcomes, including hard-to-reach and low-frequency populations of children. The EDI has also been used to evaluate programs and interventions in different countries. Conclusions: The ability of the EDI to monitor children’s developmental outcomes in various populations has been consistently demonstrated. The EDI, by virtue of its comprehensive breadth and census-like collection, widens the scope of research relating to early childhood development and its social determinants of health.

Sign in / Sign up

Export Citation Format

Share Document