ISSN (Online): 3108-1584
Volume 2, Issue 1 - July 2026
Kenyan government agencies routinely collect large volumes of health data, yet much of this information remains underutilized because of weak analytical infrastructure, resulting in disease surveillance that is largely reactive rather than proactive. Kajiado County, a vast pastoralist arid and semi-arid land (ASAL) region, exemplifies this challenge. To address this gap, a cross-sectional mixed-methods study was conducted across all five sub-counties of Kajiado between January 2025 and April 2026. The study analysed four complementary datasets comprising 450 records each: routine health facility disease records, demographic household survey data, mobile network connectivity data, and socioeconomic profiles. Statistical analyses, including descriptive statistics, independent t-tests, one-way ANOVA, Pearson correlation, and chi-square tests, were performed using Python 3.11. The findings identified malaria as the leading disease burden, with a mean of 4.62 cases per facility-month (SD = 2.51; total = 2,077 cases) and a statistically significant seasonal peak in April compared with July (t = 2.251, p = 0.028). County-wide childhood vaccination coverage averaged 83.38%, remaining below the national target of 90%, while Kajiado North recorded the lowest sub-county coverage at 82.79%. Low-income households accounted for 41.8% of the sample, and no significant relationship was observed between household income and consumption (F = 0.497, p = 0.609). Anomaly detection thresholds established at the 95th percentile were 9 cases per month for malaria, 6 for pneumonia, and 7.5 for diarrhea. These findings demonstrate that an AI-driven framework integrating Isolation Forest anomaly detection, LSTM time-series forecasting, and K-means spatial clustering within the existing DHIS2 infrastructure could transform routinely collected health data into actionable decision intelligence. With a projected return on investment of 42–58% over 36 months, the proposed framework presents a strong economic case for a 12-month pilot in Kajiado County and offers a scalable model for strengthening health surveillance across other ASAL health systems in Sub-Saharan Africa.
Artificial Intelligence, Mobile Data Analytics, Public Health Monitoring, Predictive Analytics, Arid and Semi-Arid Lands, Kenya
Jeff Wafubwa, Nelson Mapema, Okwiri Saad, Bernard Wesonga, "Leveraging Artificial Intelligence and Mobile Data Analytics for Enhancing Public Health Monitoring and Service Delivery in Kajiado County, Kenya", Cosmo Research & Science International Journal, vol. Jul-25, no. 1, pp. 348-357, 2026.
Jeff Wafubwa, Nelson Mapema, Okwiri Saad, Bernard Wesonga (2026). Leveraging Artificial Intelligence and Mobile Data Analytics for Enhancing Public Health Monitoring and Service Delivery in Kajiado County, Kenya. Cosmo Research & Science International Journal, Jul-25(1), 348-357.
Jeff Wafubwa, Nelson Mapema, Okwiri Saad, Bernard Wesonga. "Leveraging Artificial Intelligence and Mobile Data Analytics for Enhancing Public Health Monitoring and Service Delivery in Kajiado County, Kenya." Cosmo Research & Science International Journal, vol. Jul-25, no. 1, 2026, pp. 348-357.
@article{CRSIJ26000248,
author = {Jeff Wafubwa, Nelson Mapema, Okwiri Saad, Bernard Wesonga},
title = {Leveraging Artificial Intelligence and Mobile Data Analytics for Enhancing Public Health Monitoring and Service Delivery in Kajiado County, Kenya},
journal = {Cosmo Research and Science International Journal},
year = {2025},
volume = {2},
number = {1},
pages = {348-357},
issn = {3108-1584},
url = {https://cosmorsij.com/published/CRSIJ26000248.pdf},
abstract = {Kenyan government agencies routinely collect large volumes of health data, yet much of this information remains underutilized because of weak analytical infrastructure, resulting in disease surveillance that is largely reactive rather than proactive. Kajiado County, a vast pastoralist arid and semi-arid land (ASAL) region, exemplifies this challenge. To address this gap, a cross-sectional mixed-methods study was conducted across all five sub-counties of Kajiado between January 2025 and April 2026. The study analysed four complementary datasets comprising 450 records each: routine health facility disease records, demographic household survey data, mobile network connectivity data, and socioeconomic profiles. Statistical analyses, including descriptive statistics, independent t-tests, one-way ANOVA, Pearson correlation, and chi-square tests, were performed using Python 3.11. The findings identified malaria as the leading disease burden, with a mean of 4.62 cases per facility-month (SD = 2.51; total = 2,077 cases) and a statistically significant seasonal peak in April compared with July (t = 2.251, p = 0.028). County-wide childhood vaccination coverage averaged 83.38%, remaining below the national target of 90%, while Kajiado North recorded the lowest sub-county coverage at 82.79%. Low-income households accounted for 41.8% of the sample, and no significant relationship was observed between household income and consumption (F = 0.497, p = 0.609). Anomaly detection thresholds established at the 95th percentile were 9 cases per month for malaria, 6 for pneumonia, and 7.5 for diarrhea. These findings demonstrate that an AI-driven framework integrating Isolation Forest anomaly detection, LSTM time-series forecasting, and K-means spatial clustering within the existing DHIS2 infrastructure could transform routinely collected health data into actionable decision intelligence. With a projected return on investment of 42–58% over 36 months, the proposed framework presents a strong economic case for a 12-month pilot in Kajiado County and offers a scalable model for strengthening health surveillance across other ASAL health systems in Sub-Saharan Africa.},
keywords = {Artificial Intelligence, Mobile Data Analytics, Public Health Monitoring, Predictive Analytics, Arid and Semi-Arid Lands, Kenya},
month = {July}
}