Geospatial Economic Evaluation of Integrated Community-Based Collaborative Care Models for Long-Term Care in Rural Thailand

Abstract

Thailand’s public health system is facing rising socioeconomic pressures, especially among its rapidly aging society and the fiscally constrained rural regions of this lower-to-middle income country (LMIC). Identifying cost-efficient Long-Term Care (LTC) public health designs is vital to addressing this challenge across rural geographical regions. Our research set out to find synergies between elderly mental health support and LTC public health protocols, through the design integrated community-based collaborative care (I-CBCC) intervention models. The geospatial decision support tool utilizes discrete event simulation models, measured by 3 quantitative metrics: their operational workload, service delivery quality and cost-efficiency metrics, and compares proposed designs to national protocol baseline designs. Experimental model designs build from the current national public health protocols; the proposed I-CBCC model adds four key interventions, which include home-based case initialization, long-term care giver integration and digital system integration. Ablation analysis studies each (intervention) design component. Parameter sensitivity analyses were conducted to verify robustness of each model. Large geospatial datasets over four rural provinces form the data sources of the evaluations, including epidemiology (106,068 elderly participants; collected 2020-2024) and public health facilities (1,032). A geospatial analysis layer (subdistrict choropleth mapping, spatial cluster statistics, and a multi-criteria priority ranking) localizes the simulated impacts for policy rollout planning. The proposed I-CBCC model significantly improves performance in the three metrics via the I-CBCC interventions model. Cost efficiency metric found 42% total cost savings per average-patient case, over the baseline. Service delivery quality metric retained and improved upon the level of the national protocol, with positive net effect: .31 Cohen's D. Operational workload metrics reduced in areas of screenings, follow-ups and travel, by up to 44%. Geospatial analysis of the per-patient outputs shows the benefits are region-wide (positive in all 366 analyzed subdistricts), while unmet need is significantly spatially clustered (Moran's I = 0.394, p = 0.001), yielding a district priority ranking for a staggered policy rollout. The intervention synergies of the I-CBCC specialized protocol are preferable to the national protocol in this setting. These findings serve as a scalable framework for improving the national public health protocol, with a geospatial multi-criteria prioritization index identifying the districts for phased policy rollout.