Data-driven decision support for community health outreach: Prioritizing senior health interventions using county-level public health indicators

Qi Xin

doi:10.65112/tcmis.10088

Authors

Qi Xin Management Information Systems, University of Pittsburgh, 4200 Fifth Ave, Pittsburgh, PA 15260, USA https://orcid.org/0009-0008-9805-6649

DOI:

https://doi.org/10.65112/tcmis.10088

Keywords:

Decision support, community health outreach, senior communities, nonprofit planning, chronic disease intervention

Abstract

Senior-serving nonprofit organizations and public health planners need transparent methods for deciding which counties should receive technology training, health education, chronic disease support, care navigation, and related outreach resources first. This study developed a county-level decision-support framework that converts CDC PLACES public health indicators into outreach priority scores, empirical risk clusters, rule-recovery audits, deterministic cluster-specific outreach plans, uncertainty checks, and a senior-focused resource-allocation scenario. The main analysis used the 2024 CDC PLACES county Open Data release (3,144 county equivalents after excluding the national aggregate), and external validation used the 2025 PLACES county GIS-friendly release (3,143 counties). Forty indicators were harmonized across health outcomes, preventive care, health risk behaviors, disability, health status, and health-related social needs. A 2024 Census county age-sex file supplied the estimated population aged 65 years and older for the optimization module. The 2024 priority score had a median of 49.50, a top-quintile threshold of 69.16, and a top-decile threshold of 76.17. The critical top-decile band contained 315 counties. A five-cluster operational solution separated critical multimorbidity and social-need counties, high chronic-disease and behavioral-risk counties, preventive-care gap counties, moderate mixed-risk counties, and lower-risk monitoring counties. KNN and MICE-style imputations produced ordered cluster mean-score correlations of 0.964 and 0.967. Logistic regression achieved five-fold cross-validation F1=0.979; XGBoost achieved held-out F1=0.955; and XGBoost achieved 2025 external F1=0.954. A PLACES confidence-interval simulation produced a median top-quintile threshold of 68.07 with a 2.5th-97.5th percentile interval of 67.77-68.34; the deterministic threshold was above this simulated interval, so threshold-level agreement was not assumed, although median county priority-band stability was 0.973. Under a 600-unit budget, the mixed-integer optimization selected 158 elevated-or-higher-priority counties and covered approximately 4.61 million residents aged 65 years or older. The framework provides a reproducible county-level planning workflow while preserving the limits of model-based local estimates.

Downloads

Download data is not yet available.

References

[1] Centers for Disease Control and Prevention, "PLACES: Local Data for Better Health," 2024. [Online]. Available: https://www.cdc.gov/places/

[2] Centers for Disease Control and Prevention, "Methodology: PLACES," 2025. [Online]. Available: https://www.cdc.gov/places/methodology/

[3] Centers for Disease Control and Prevention, "PLACES Data Portal," 2024. [Online]. Available: https://www.cdc.gov/places/tools/data-portal.html

[4] U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, "PLACES: Local Data for Better Health, County Data 2024 release," Data.gov, 2024. [Online]. Available: https://catalog.data.gov/dataset/places-local-data-for-better-health-county-data-2024-release-cb149

[5] Centers for Disease Control and Prevention, "PLACES County Data GIS Friendly Format, 2025 release," 2025. [Online]. Available: https://data.cdc.gov/500-Cities-Places/PLACES-County-Data-GIS-Friendly-Format-2025-releas/i46a-9kgh

[6] K. J. Greenlund, H. Lu, Y. Wang, K. A. Matthews, J. M. LeClercq, B. Lee, et al., “PLACES: Local data for better health,” Prev. Chronic Dis., vol. 19, p. E31, Jun. 16 2022.

[7] P. L. Remington, B. B. Catlin, and K. P. Gennuso, “The County Health Rankings: Rationale and methods,” Popul. Health Metr., vol. 13, no. 11, 2015.

[8] C. M. Hood, K. P. Gennuso, G. R. Swain, and B. B. Catlin, “County Health Rankings: Relationships between determinant factors and health outcomes,” Am. J. Prev. Med., vol. 50, no. 2, pp. 129–135, 2016.

[9] Office of Disease Prevention and Health Promotion, "Social determinants of health," Healthy People 2030. [Online]. Available: https://health.gov/healthypeople/priority-areas/social-determinants-health

[10] Centers for Disease Control and Prevention, "The CDC Clear Communication Index," 2024. [Online]. Available: https://www.cdc.gov/ccindex/

[11] M. Ramezani, et al., “Artificial intelligence and machine learning in health policy: A bibliometric analysis,” BMC Health Serv. Res., vol. 23, art. 1416, 2023.

[12] T. Y. C. Tam, S. Sivarajkumar, S. Kapoor, A. V. Stolyar, K. Polanska, K. R. McCarthy, et al., “A framework for human evaluation of large language models in healthcare derived from literature review,” NPJ Digit. Med., vol. 7, no. 1, art. 258, 2024.

[13] L. Breiman, “Random forests,” Mach. Learn., vol. 45, no. 1, pp. 5–32, 2001.

[14] T. Chen and C. Guestrin, "XGBoost: a scalable tree boosting system," in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining, 2016, pp. 785-794.

[15] S. Lloyd, “Least squares quantization in PCM,” IEEE Trans. Inf. Theory, vol. 28, no. 2, pp. 129–137, 1982., doi:

[16] J. H. Ward, Jr., “Hierarchical grouping to optimize an objective function,” J. Am. Stat. Assoc., vol. 58, no. 301, pp. 236–244, 1963.

[17] F. Murtagh and P. Legendre, “Ward’s hierarchical agglomerative clustering method: Which algorithms implement Ward’s criterion?” J. Classif., vol. 31, no. 3, pp. 274–295, 2014.

[18] F. Pedregosa, et al., “Scikit-learn: Machine learning in Python,” J. Mach. Learn. Res., vol. 12, pp. 2825–2830, 2011.

[19] S. M. Lundberg and S.-I. Lee, "A unified approach to interpreting model predictions," in Advances in Neural Information Processing Systems, vol. 30, 2017.

[20] R. Flesch, “A new readability yardstick,” J. Appl. Psychol., vol. 32, no. 3, pp. 221–233, 1948.

[21] J. P. Kincaid, R. P. Fishburne, L. Rogers, and B. Chissom, "Derivation of new readability formulas for Navy enlisted personnel," Naval Technical Training Command, Research Branch Report 8-75, 1975.

[22] G. H. McLaughlin, “SMOG grading: A new readability formula,” J. Read., vol. 12, no. 8, pp. 639–646, 1969.

[23] U.S. Census Bureau, "County Population by Characteristics: 2020-2024," 2025. [Online]. Available: https://www.census.gov/data/tables/time-series/demo/popest/2020s-counties-detail.html

[24] U.S. Census Bureau, "2024 Gazetteer Files," 2025. [Online]. Available: https://www.census.gov/geographies/reference-files/time-series/geo/gazetteer-files.html

[25] S. van Buuren and K. Groothuis-Oudshoorn, “mice: Multivariate imputation by chained equations in R,” J. Stat. Softw., vol. 45, no. 3, pp. 1–67, 2011.

[26] O. Troyanskaya, M. Cantor, G. Sherlock, P. Brown, T. Hastie, R. Tibshirani, et al., “Missing value estimation methods for DNA microarrays,” Bioinformatics, vol. 17, no. 6, pp. 520–525, Jun. 2001.

[27] G. B. Dantzig, Linear Programming and Extensions. Princeton, NJ: Princeton University Press, 1963.

[28] P. Virtanen, R. Gommers, T. E. Oliphant, M. Haberland, T. Reddy, D. Cournapeau, et al.; SciPy 1.0 Contributors, “SciPy 1.0: Fundamental algorithms for scientific computing in Python,” Nat. Methods, vol. 17, no. 3, pp. 261–272, 2020.