Forecasting Years Lived in Poor Health in Kenya: A Comparative Analysis of ARIMA and LSTM Models with Implications for East African Health Policy
Abstract
Kenya's life expectancy increased from 55.5 to 64.7 years (1990-2024), but years lived in poor health (GAP) have not declined proportionally, creating substantial public health planning challenges. Evidence-based forecasting of GAP trends is essential for health system resource allocation, yet no systematic forecasting methodologies exist for East African health systems. This study compared ARIMA and LSTM forecasting models for predicting Kenya's GAP trends and established methodological frameworks for health system planning across East Africa. Comparative time series analysis was conducted using Global Burden of Disease Study 2021 data spanning 1990-2021 for Kenya, Uganda, and Tanzania health systems, with 32 annual observations for each country. ARIMA and LSTM models were developed and validated using identical specifications, with performance evaluated using RMSE, MAE, and Diebold-Mariano statistical tests for significance. ARIMA significantly outperformed LSTM in Kenya (RMSE: 5.67 vs 6.66, p<0.001), reflecting stable health system patterns suitable for systematic planning, while LSTM demonstrated superior performance in Uganda (RMSE: 8.47 vs 15.03) and Tanzania (RMSE: 7.30 vs 10.10), indicating more complex health dynamics requiring sophisticated modelling approaches. Kenya's predictable GAP patterns enable reliable ARIMA-based forecasting for health system planning, while regional variations necessitate context-specific methodological approaches across East African health systems. This study provides the first systematic GAP forecasting framework for East Africa, offering health policy makers evidence-based tools for resource allocation while establishing methodological foundations for public health planning that can strengthen health systems across the region
Downloads
References
Alwago, W. O. (2023). The nexus between health expenditure, life expectancy, and economic growth: ARDL model analysis for Kenya. Regional Science Policy & Practice, 15(5), 1064–1086. https://doi.org/10.1111/rsp3.12588
Cao, X., Hou, Y., Zhang, X., Xu, C., Jia, P., Sun, X., Sun, L., Gao, Y., Yang, H., Cui, Z., Wang, Y., & Wang, Y. (2020). A comparative, correlational analysis and projection of global and regional life expectancy, healthy life expectancy, and their GAP: 1995-2025. Journal of Global Health, 10(2), 020407. https://doi.org/10.7189/jogh.10.020407
Cerqueira, V., Torgo, L., & Mozetič, I. (2020). Evaluating time series forecasting models: An empirical study on performance estimation methods. Machine Learning, 109(11), 1997–2028. https://doi.org/10.1007/s10994-020-05910-7
Elsaraiti, M., & Merabet, A. (2021). A Comparative Analysis of the ARIMA and LSTM Predictive Models and Their Effectiveness for Predicting Wind Speed. Energies, 14(20), 6782. https://doi.org/10.3390/en14206782
Esteban, C., Staeck, O., Baier, S., Yang, Y., & Tresp, V. (2016). Predicting Clinical Events by Combining Static and Dynamic Information Using Recurrent Neural Networks. 2016 IEEE International Conference on Healthcare Informatics (ICHI), 93–101. https://doi.org/10.1109/ICHI.2016.16
Garrido, J., Shang, Y., & Xu, R. (2024). LSTM-Based Coherent Mortality Forecasting for Developing Countries. Risks, 12(2), 27. https://doi.org/10.3390/risks12020027
Hyden, G., & Onyango, G. (2021). Kenya: A Comparative East African Perspective. In G. Onyango & G. Hyden (Eds.), Governing Kenya (pp. 257–277). Springer International Publishing. https://doi.org/10.1007/978-3-030-61784-4_15
Institute for Health Metrics and Evaluation (IHME). (2021). Global Burden of Disease Study 2021 (GBD 2021) Data Resources [Dataset]. https://ghdx.healthdata.org/gbd-2021
Kim, E., Ha, J., Jeon, Y., & Lee, S. (2004). Ljung-Box Test in Unit Root AR-ARCH Model. Communications for Statistical Applications and Methods, 11(2), 323–327. https://doi.org/10.5351/CKSS.2004.11.2.323
Kontopoulou, V. I., Panagopoulos, A. D., Kakkos, I., & Matsopoulos, G. K. (2023). A Review of ARIMA vs. Machine Learning Approaches for Time Series Forecasting in Data Driven Networks. Future Internet, 15(8), 255. https://doi.org/10.3390/fi15080255
Labbe, J. A. (2010). Health-Adjusted Life Expectancy: Concepts and Estimates. In V. R. Preedy & R. R. Watson (Eds.), Handbook of Disease Burdens and Quality of Life Measures (pp. 417–424). Springer New York. https://doi.org/10.1007/978-0-387-78665-0_23
Li, J. (2024). Bayesian modelling of best-performance healthy life expectancy. Journal of Population Research, 41(2), 8. https://doi.org/10.1007/s12546-024-09330-5
Majidnia, M., Ahmadabadi, Z., Zolfaghari, P., & Khosravi, A. (2023). Time series analysis of cutaneous leishmaniasis incidence in Shahroud based on ARIMA model. BMC Public Health, 23(1), 1190. https://doi.org/10.1186/s12889-023-16121-9
Martinez, R., Morsch, P., Soliz, P., Hommes, C., Ordunez, P., & Vega, E. (2021). Life expectancy, healthy life expectancy, and burden of disease in older people in the Americas, 1990–2019: A population-based study. Revista Panamericana de Salud Pública, 45, 1–14. https://doi.org/10.26633/RPSP.2021.114
Mienye, I. D., Swart, T. G., & Obaido, G. (2024). Recurrent Neural Networks: A Comprehensive Review of Architectures, Variants, and Applications. Information, 15(9), 517. https://doi.org/10.3390/info15090517
Moses, M. W., Korir, J., Zeng, W., Musiega, A., Oyasi, J., Lu, R., Chuma, J., & Di Giorgio, L. (2021). Performance assessment of the county healthcare systems in Kenya: A mixed-methods analysis. BMJ Global Health, 6(6), e004707. https://doi.org/10.1136/bmjgh-2020-004707
Mtintsilana, A., Craig, A., Mapanga, W., Dlamini, S. N., & Norris, S. A. (2023). Association between socio-economic status and non-communicable disease risk in young adults from Kenya, South Africa, and the United Kingdom. Scientific Reports, 13(1), 728. https://doi.org/10.1038/s41598-023-28013-4
Naghavi, M., Ong, K. L., Aali, A., Ababneh, H. S., Abate, Y. H., Abbafati, C., Abbasgholizadeh, R., Abbasian, M., Abbasi-Kangevari, M., Abbastabar, H., Abd ElHafeez, S., Abdelmasseh, M., Abd-Elsalam, S., Abdelwahab, A., Abdollahi, M., Abdollahifar, M.-A., Abdoun, M., Abdulah, D. M., Abdullahi, A., … Murray, C. J. L. (2024). Global burden of 288 causes of death and life expectancy decomposition in 204 countries and territories and 811 subnational locations, 1990–2021: A systematic analysis for the Global Burden of Disease Study 2021. The Lancet, 403(10440), 2100– 2132. https://doi.org/10.1016/S0140-6736(24)00367-2
Neumark, T., & Prince, R. J. (2021). Digital Health in East Africa: Innovation, Experimentation and the Market. Global Policy, 12(S6), 65–74. https://doi.org/10.1111/1758-5899.12990
Njenga, J. K., & Kipchirchir, I. C. (2024). Modelling Mortality in Kenya. Asian Research Journal of Mathematics, 20(1), 1–15. https://doi.org/10.9734/arjom/2024/v20i1777
Nyangena, J., Rajgopal, R., Ombech, E. A., Oloo, E., Luchetu, H., Wambugu, S., Kamau, O., Nzioka, C., Gwer, S., & Ndiritu Ndirangu, M. (2021). Maturity assessment of Kenya’s health information system interoperability readiness. BMJ Health & Care Informatics, 28(1), e100241. https://doi.org/10.1136/bmjhci-2020-100241
Nyawira, L., Njuguna, R. G., Tsofa, B., Musiega, A., Munywoki, J., Hanson, K., Mulwa, A., Molyneux, S., Maina, I., Normand, C., Jemutai, J., & Barasa, E. (2023). Examining the influence of health sector coordination on the efficiency of county health systems in Kenya. BMC Health Services Research, 23(1), 355. https://doi.org/10.1186/s12913-023-09344-4
Odeny, B. M., Njoroge, A., Gloyd, S., Hughes, J. P., Wagenaar, B. H., Odhiambo, J., Nyagah, L. M., Manya, A., Oghera, O. W., & Puttkammer, N. (2023). Development of novel composite data quality scores to evaluate facility-level data quality in electronic data in Kenya: A nationwide retrospective cohort study. BMC Health Services Research, 23(1), 1139. https://doi.org/10.1186/s12913-023-10133-2
Permanyer, I., & Bramajo, O. (2023). The Race between Mortality and Morbidity: Implications for the Global Distribution of Health. Population and Development Review, 49(4), 909–937. https://doi.org/10.1111/padr.12582
Robine, J. M., Romieu, I., & Cambois, E. (1999). Health expectancy indicators. Bulletin of the World Health Organization, 77(2), 181–185.
Sherstinsky, A. (2020). Fundamentals of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) network. Physica D: Nonlinear Phenomena, 404, 132306. https://doi.org/10.1016/j.physd.2019.132306
Siami-Namini, S., Tavakoli, N., & Siami Namin, A. (2018). A Comparison of ARIMA and LSTM in Forecasting Time Series. 2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA), 1394–1401. https://doi.org/10.1109/ICMLA.2018.00227
Sirisha, U. M., Belavagi, M. C., & Attigeri, G. (2022). Profit Prediction Using ARIMA, SARIMA, and LSTM Models in Time Series Forecasting: A Comparison. IEEE Access, 10, 124715– 124727. https://doi.org/10.1109/ACCESS.2022.3224938
Tokudome, S., Hashimoto, S., & Igata, A. (2016). Life expectancy and healthy life expectancy of Japan: The fastest graying society in the world. BMC Research Notes, 9(1), 482. https://doi.org/10.1186/s13104-016-2281-2
Tsantekidis, A., Passalis, N., & Tefas, A. (2022). Recurrent neural networks. In Deep Learning for Robot Perception and Cognition (pp. 101–115). Elsevier. https://doi.org/10.1016/B978-0-32-385787-1.00010-5
Wang, T., Tian, Y., & Qiu, R. G. (2020). Long Short-Term Memory Recurrent Neural Networks for Multiple Diseases Risk Prediction by Leveraging Longitudinal Medical Records. IEEE Journal of Biomedical and Health Informatics, 24(8), 2337–2346. https://doi.org/10.1109/JBHI.2019.2962366
Wang, X., Kang, Y., Hyndman, R. J., & Li, F. (2023). Distributed ARIMA models for ultra-long time series. International Journal of Forecasting, 39(3), 1163–1184. https://doi.org/10.1016/j.ijforecast.2022.05.001
World Health Organization (WHO). (2023). Health data overview for the Republic of Kenya [Dataset]. https://data.who.int/countries/404
Zhang, R., Song, H., Chen, Q., Wang, Y., Wang, S., & Li, Y. (2022). Comparison of ARIMA and LSTM for prediction of hemorrhagic fever at different time scales in China. PLOS ONE, 17(1), e0262009. https://doi.org/10.1371/journal.pone.0262009
Zhou, J., Li, H., & Zhong, W. (2021). A modified Diebold–Mariano test for equal forecast accuracy with clustered dependence. Economics Letters, 207, 110029. https://doi.org/10.1016/j.econlet.2021.110029
Copyright (c) 2025 Andrew Karani, Ren Dongxiao
This work is licensed under a Creative Commons Attribution 4.0 International License.
