A SECURITY-GOVERNED, AUDITABLE SYSTEM DYNAMICS MODEL FOR LUNG CANCER CASE LOAD MANAGEMENT IN KENYA: INTEGRATING PATTERN ANALYSIS WITH COMPLIANCE CONTROL

Abstract

Lung cancer remains one of the leading contributors to cancer mortality globally and in Kenya, accounting for approximately 18% of cancer-related deaths as reported by the Kenya National Cancer Registry and GLOBOCAN 2024. This study developed and evaluated a security-governed, auditable System Dynamics Model (SDM) for lung cancer caseload management, conceptualised as the coordinated control of patient volumes across healthcare levels, facilities, and referral pathways. Guided by four objectives, the study assessed the structural configuration, facility distribution, and reporting patterns influencing caseload management; examined ICT integration and secure data architecture within existing systems; designed and simulated a System Dynamics Model incorporating referral delays, facility capacity, and feedback structures; and evaluated the model‘s forecasting accuracy and decision-support capability. A mixed-methods design was employed. Quantitative data were obtained from the Kenya National Cancer Registry, the Kenya Health Information System, and GLOBOCAN datasets, while qualitative insights were drawn from a Delphi panel comprising oncologists, ICT managers, and healthcare policymakers. Analytical procedures integrated pattern analysis techniques using Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN) models for referral delay and trend prediction, complemented by Vensim for model construction, simulation, and sensitivity analysis. Security validation applied the Security Maturity Index (SMI) and a probability-of-breach metric consistent with the Kenya Data Protection Act (2019/2022). The findings indicated that more than 65% of patients were diagnosed at advanced stages, with diagnostic and treatment capacity heavily concentrated in national referral hospitals and reporting processes remaining inconsistent in lower-level facilities. The SDM achieved a prediction accuracy of 98.1% (MAPE = 1.9%) and demonstrated that strengthening referral linkages, enhancing ICT integration, and adopting secure data practices significantly improves caseload coordination, forecasting reliability, and data integrity. The study concludes that combining pattern analysis with System Dynamics provides a practical, secure, and evidence-driven decision-support tool for healthcare managers and policymakers, enabling more efficient resource planning and strengthened caseload management within Kenya‘s cancer control framework.