Adaptive model of uncertainty in healthcare technology management

Abstract

Healthcare Technology Management (HTM) operates under significant uncertainty arising from technical, managerial, and macro-environmental factors. Existing risk assessment models—including recent data-driven and machine learning–based approaches—focus predominantly on internal maintenance variables and fail to account for broader sources of uncertainty or the dynamic influence of equipment utilization. Addressing this gap, this thesis develops an Adaptive Model of Uncertainty in HTM by integrating expert judgment, fuzzy multi-criteria weighting, and nonlinear utilization–risk modelling. An initial pool of 53 HTM risk factors was systematically screened through expert relevance scoring, after which 41 significant factors were prioritized using a Fuzzy Analytic Hierarchy Process (Fuzzy AHP), yielding stable global weights that reflect region-specific HTM priorities in Saudi Arabia and the Gulf. Building on this weighted risk structure, the thesis formulates the Synergistic Quadratic–Resonant Risk Model (SQRRM) as a nonlinear system-identification law that couples utilization intensity with fuzzy-weighted readiness gaps to adaptively modulate risk magnitude. The proposed model was identified using a five-year hospital imaging dataset and evaluated on the aggregated 2024 cross-section, where it demonstrated superior explanatory performance (R² = 0.943) relative to six competing baseline formulations. To assess the potential cross-domain applicability of the proposed modelling framework, SQRRM was embedded within the BiLSTM and Adaptive Dynamic Graph Neural Network (ADGNN) architectures and evaluated on the NASA C-MAPSS FD001 turbofan degradation dataset. Under the adopted experimental protocol, the SQRRM-enhanced ADGNN achieved the strongest predictive performance among the selected comparator models (RMSE = 12.80, R² = 0.905). These findings provide additional empirical evidence supporting the robustness of the proposed nonlinear modelling framework and its potential applicability beyond the healthcare domain. The contributions of this thesis include: (i) the first region-specific, expert-validated hierarchy of HTM risks; (ii) an integrated risk-quantification framework that unifies readiness gaps, utilization, and operational burden; (iii) the derivation of the SQRRM model as a nonlinear, resonance-modulated risk function; and (iv) successful cross-domain validation through deep learning architectures. Recommendations emphasize the need for improved Computerized Maintenance Management System (CMMS) data governance, evaluation of current mitigation strategies, and expansion of contextual readiness indicators to enhance predictive-risk analytics in HTM. Overall, this thesis proposes and initially validates an integrated Healthcare Technology Management (HTM) risk-modelling framework within the healthcare context of Saudi Arabia and the Gulf region. It establishes a scientifically grounded and operationally implementable foundation for risk-aware, utilization-driven decision making in healthcare technology management, while the cross-domain evaluation provides initial evidence of the framework's potential applicability to other high-reliability engineering systems, including aviation and energy.