Jun 2026· Discover Public Health· Vol 23· 0 citations· 30 references
TL;DR
These forecasts provide a robust statistical baseline to assist the Somali Ministry of Health in strategic planning for oncology infrastructure, diagnostic capacity expansion, and the formal establishment of a National Cancer Registry to transition from reconstructed estimates to actual patient records.
Abstract
Cancer mortality is rising rapidly across the African continent, yet Somalia lacks localized national forecasting evidence due to decades of instability and the absence of a formal registry. This study aims to develop and compare high-precision time-series forecasting models to project cancer mortality trends in Somalia through 2030, providing a critical statistical baseline for proactive healthcare resource allocation. Annual cancer mortality data (1980–2021) were retrieved from the Global Burden of Disease (GBD) database. A rigorous comparative analysis of 12 representative forecasting configurations was conducted, comprising six single models (including ARIMA, ETS, TBATS, and the Theta method) and six optimized hybrid ensembles integrated with Artificial Intelligence (AI) proxies like Extreme Learning Machines (ELM) and Neural Network Autoregression (NNAR). To stabilize non-stationary trends and mitigate overfitting in a limited sample context (n = 42), second-order differencing (d = 2) and a 10-year rolling-origin cross-validation (tsCV) were implemented. Model performance was benchmarked using RMSE, MAPE, Theil’s U, Willmott’s Index of Agreement (WI), and Skill Scores. All evaluated configurations achieved Theil’s U statistics below 1.0, indicating superior performance relative to the naive benchmark. The parsimonious ARIMA (2,2,1) model emerged as the most accurate configuration for long-term forecasting, achieving the lowest MAPE (1.127%) and RMSE (100.16) in the independent holdout set. The statistical superiority of ARIMA (2,2,1) was confirmed via the Diebold–Mariano test (p < 0.01) against the naive benchmark, outperforming complex hybrid ensembles which showed higher error accumulation over the 8-year horizon. Projections indicate a consistent monotonic upward trend, with annual cancer deaths estimated to reach approximately 9,652 by the year 2030. The findings indicate a significant and ongoing epidemiological transition in Somalia. These forecasts provide a robust statistical baseline to assist the Somali Ministry of Health in strategic planning for oncology infrastructure, diagnostic capacity expansion, and the formal establishment of a National Cancer Registry to transition from reconstructed estimates to actual patient records.
ABSTRACT Background and Aims Dengue fever is a growing menace in Somalia, a climate change prone region with a weak healthcare system. An imperative of public health is effective forecasting models. This paper will provide a detailed comparative analysis comparing mechanistic, statistical, and hybrid models in order to find the best forecasting model to use in this data‐sparse situation of dengue fever. Methods We utilized raw reported annual incidence data of dengue (1990–2021, N = 32) obtained from Our World in Data. Due to the lack of standardized case definitions and high under‐reporting inherent to Somalia's surveillance system, forecasts represent the projected reported burden rather than true infection rates. The data was divided into a training and a testing set (1990–2016 and 2017–2021, respectively). We tested a mechanistic Susceptible‐Infected‐Recovered (SIR) model, 8 single time series models (ARIMA, ETS, TBATS, and NNAR), and 12 hybrids using an averaging ensemble strategy. Findings Although the time series models (individually) offered the basics of predictive power, the hybrid ARIMA‐TBATS model was the most successful and it surpassed the other models. It performed the highest accuracy statistics on the testing data with a Mean Absolute Percentage Error (MAPE) of 6.49% and a Root Mean Squared Error (RMSE) of 663.81 and outperformed the best single model (TBATS) which had a MAPE of 7.14%. The basic reproduction number (R0) calculated by the SIR model was 1.015 which shows that the disease is endemic. Conclusion This paper concludes that the ARIMA‐TBATS hybrid should be the most empirically effective tool out of the ones considered when planning to forecast dengue in Somalia. Real‐World Application: This ARIMA‐TBATS framework is designed to serve as the mathematical engine for a National Dengue Early Warning System (DEWS) in Somalia, allowing public health officials to physically preempt outbreaks by routing medical supplies, mobilizing fumigation teams, and distributing bed nets to high‐risk zones months before peak incidence occurs.
Abdiftah Mohamud Abdi, S. Nadarajah, Abdisalam Hassan Muse· Health Science Reports· 0 citations
Hypertension is a major global risk factor for cardiovascular disease and premature mortality, with increasing public health implications in India. This study compares ARIMA, LSTM, and Hybrid ARIMA–LSTM models for forecasting India’s hypertension burden using annual data from 1990–2025 across six epidemiological indicators: prevalence, prevalence cases, ASPR, incidence, deaths, and DALYs. After data pre-processing and stationarity testing, ARIMA captured linear trends, LSTM modeled nonlinear patterns, and the hybrid model integrated both approaches. Model performance was evaluated using MAE, RMSE, MAPE, R², and residual diagnostics. The Hybrid ARIMA–LSTM model consistently achieved superior forecasting accuracy, with lower errors and better goodness-of-fit than the standalone models. Forecasts for 2026–2045 suggest that hypertension will remain a significant health challenge, with prevalence reaching 30.87% and affecting approximately 350.11 million people by 2045, along with 13.83 million incident cases, 217.77 thousand deaths, and 5.54 million DALYs. The findings highlight the need for strengthened prevention, early detection, disease management, and evidence-based health policies, while supporting the Hybrid ARIMA–LSTM framework as a robust tool for long-term epidemiological forecasting.
D. Singh· International Journal For Mu...· 0 citations
Background Perinatal mortality is a critical indicator of the quality of maternal and newborn care across sub-Saharan Africa. A recent systematic review and meta-analysis estimated Ghana's pooled perinatal mortality rate at 44.8 per 1,000 births, highlighting ongoing barriers to meeting Sustainable Development Goal targets for neonatal survival. Methods We conducted a retrospective, hospital-based time series analysis of 192 monthly observations from January 2010 through December 2025. Perinatal mortality rate (PMR) was defined as the sum of stillbirths and early neonatal deaths per 1,000 births. Stationarity was evaluated using the Augmented Dickey Fuller (ADF) test. Forecasting performance was compared across four models—ARIMA, backpropagation neural network (BPNN), deep learning neural network (DLNN), and generalized regression neural network (GRNN)—with model validation performed on a 2025 temporal holdout. Results The hospital recorded 46,108 live births and 1,152 perinatal deaths, giving an overall PMR of 24.98 per 1,000 births. The undifferenced monthly PMR series was borderline non-stationary (ADF statistic −2.695; p = 0.075), while the first-differenced series was stationary (ADF statistic −7.118; p < 0.001). The best-performing ARIMA model on the 2025 holdout was ARIMA (3, 0, 0). Test-set RMSE values were 11.74 for ARIMA, 14.98 for BPNN, 12.87 for DLNN, and 13.33 for GRNN. In ecological monthly models, higher ANC coverage was associated with lower PMR, whereas higher hypertension burden was associated with higher PMR. Conclusion Perinatal mortality declined over the long term but remained unstable. Among the evaluated models, ARIMA showed the best out-of-sample accuracy, while GRNN was the strongest neural-network comparator. Forecasts should be interpreted as operational projections rather than causal predictions. What is already known on this topic? Perinatal mortality remains high in many low- and middle-income countries, and stillbirths plus early neonatal deaths continue to contribute substantially to under-5 mortality in Ghana ( 1– 6). What this study adds This study provides a 16-year monthly hospital time series, compares classical time-series forecasting with three neural-network approaches, and demonstrates that better ANC coverage and lower hypertension burden track with lower monthly PMR in this setting. How this study might affect research, practice, or policy Monthly PMR surveillance may help maternity hospitals monitor service quality, and ARIMA-based operational forecasting may assist planning for high-risk periods while service-improvement efforts focus on ANC utilization and maternal complication control.
A. Lartey, D. Yar, Ama Asamaniwa Attua et al.· Frontiers in Reproductive He...· 0 citations
Accurate mortality predictions are important for pension sustainability and life insurance valuation. Existing extensions of the Lee–Carter (LC) model typically use a fixed fitting period and rely on a single forecasting approach to predict the time component. This study proposes a hybrid mortality forecasting framework based on the LC model, with a particular focus on improving the estimation of its time component. The approach integrates an optimized selection of the fitting period with both traditional time-series modeling with auto-regressive integrated moving average (ARIMA) (p,d,q) and machine learning techniques, namely artificial neural networks (ANNs) and random forests (RFs). The objective is to assess whether these enhancements improve forecasting performance. Using 45 years of Malaysian age-specific mortality data (1980–2024), this study compares the predictive performance of the standard LC model with the proposed extensions: LC-ARIMA, LC-ANN, and LC-RF. Results showed that, while the LC ARIMA version minimizes prediction residuals by using fitting periods of 1980–2002 for males and 1980–2004 for females, the LC-ANN version achieves the highest aggregate predictive accuracy when averaged across genders. These findings suggest that integrating neural networks into the LC framework effectively captures the time-component patterns. Our projections through 2038 indicate a continuous decline in mortality rates among Malaysians, with greater improvement among females. Overall, the proposed hybrid framework offers a more accurate and flexible approach to mortality forecasting. These improvements are particularly relevant for applications such as pension planning and population projections, where reliable mortality estimates are essential for long-term policy decisions.
S. N. Shair, Norazliani Md Lazam, Nur Ezyan et al.· Journal of Artificial Intell...· 0 citations
Time series forecasting is a pivotal tool across multiple disciplines, particularly in epidemiology, where precise predictions can guide resource allocation and inform public health strategies. This study explores the application of Seasonal Autoregressive Integrated Moving Average (SARIMA) models for short-term forecasting of time series data, using epidemiological data from Kazakhstan related to COVID-19 as a practical case study. The dataset encompasses total confirmed cases, daily ambulatory care patients, and daily hospitalized patients, spanning from May 30, 2022, to December 14, 2022, with forecasts extending 10 days forward to December 24, 2022. The analysis leverages SARIMA’s ability to capture both seasonal and non-stationary patterns, making it highly effective for modeling complex epidemiological dynamics.
The methodology involves several key steps: data preprocessing through natural-log transformation to stabilize variance, stationarity testing using autocorrelation function (ACF) plots, and differencing to address non-stationarity. The optimal SARIMA models identified were (5,3,1)(1,1,0)_7 for total cases, (1,2,2)(0,1,1)_7 for ambulatory care patients, and (2,2,1)(2,1,1)_7 for hospitalized patients, with forecasting accuracies of 99%, 97%, and 91%, respectively. These high accuracies, measured via Mean Absolute Percentage Error (MAPE), underscore SARIMA’s robustness in short-term predictions. Residual analysis, including Shapiro-Wilk and Ljung-Box tests, confirmed that the models’ residuals were normally distributed and independent, validating their suitability for forecasting.
This study highlights SARIMA’s versatility in capturing weekly seasonal patterns, as observed in the 7-day cycles within the COVID-19 data, which reflect reporting or behavioral trends. While the example focuses on epidemiological data, the methodology is broadly applicable to other domains, such as ecological monitoring or economic forecasting, where seasonal time series are prevalent. The findings demonstrate that SARIMA models provide a reliable framework for short-term forecasting, offering actionable insights for public health interventions and resource planning, with the COVID-19 dataset serving as an illustrative example of the approach’s efficacy and adaptability.
M. Sorokina, I. Korshukov, N. Omarbekova et al.· Medicine and ecology· 0 citations
Objective: This study aimed to predict and analyze air pollution-related mortality rates in Bangkok, Thailand, using a comprehensive neural network (NN) model. Objectives included analyzing temporal dynamics, evaluating model effectiveness, and identifying the influential factors. Material and Methods: Daily air quality and mortality data from 2016 to 2020 were used. We employed recurrent neural networks (RNNs), long short-term memory (LSTM), and gated recurrent units (GRU) models to capture the complex relationships between six key pollutants and mortality. The best model was selected based on the lowest R2 and mean difference from the Bland–Altman analysis. SHAP feature importance and subgroup analyses for age (0–5, 5–60, and 60+ years) and cause of death (respiratory, circulatory, and other) were conducted. Results: Analysis included 170,612 mortality cases over 1,828 days, with a median=36 daily premature deaths. An LSTM model with a 23-day time lag demonstrated the highest predictive accuracy for overall mortality. Subgroup analysis identified different optimal models; an RNN model was best for the “Older Adult” subgroup, and an LSTM model was best for the “Respiratory” subgroup. For feature importance, relative humidity, particulate Matter (PM2.5), and ozone (O3) were the most influential for overall mortality. The most influential factors for older adults were PM10, PM2.5, and carbon monoxide (CO); for respiratory, PM2.5, nitrogen dioxide (NO2), and PM10 were the most influential. Conclusion: This study demonstrates NN’s potential in predicting air pollution-related mortality rates in Bangkok. Findings highlight the importance of considering temporal dynamics, subgroup-specific characteristics, and the key environmental factors in model development. These data-driven insights can inform public health policies and facilitate targeted interventions to mitigate the health impacts of urban air pollution.
Kanakorn Horsiritham, Natthaya Bunplod, P. Sirinara et al.· Journal of Health Science an...· 0 citations