Leveraging Artificial Intelligence to Optimize Vaccine Delivery and Cold Chain Logistics in Remote Areas: A Scoping Review
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background: Vaccine delivery in remote and low-resource areas remains constrained by fragile cold chain systems, fragmented data infrastructures, and logistical inefficiencies. Artificial intelligence (AI) offers potential solutions through predictive analytics, Internet of Things (IoT)-enabled monitoring, and autonomous delivery systems. However, evidence on the real-world use of AI in vaccine logistics remains fragmented, with few syntheses capturing its scope, challenges, and equity implications. Objective: This scoping review aimed to map the global evidence on AI applications in vaccine delivery and cold chain logistics, focusing on their implementation in remote and low-resource contexts, and to identify the barriers and enablers influencing the sustainable deployment of these applications. Methods: Following the PRISMA-ScR framework, literature published between 2015 and 2025 was systematically searched across PubMed, Scopus, IEEE Xplore, Web of Science, WHO IRIS, and UNICEF repositories. Eligible studies included peer-reviewed and grey literature explicitly describing AI-enabled tools for vaccine storage, routing, or delivery. Data were charted for the AI method, logistical function, outcomes, and implementation context, and synthesized narratively. The protocol registered in PROSPERO, with registration ID: 1230853. Results: Fifty-six studies met the inclusion criteria. AI was applied across four main domains: (1) AI + IoT for real-time cold chain monitoring, (2) AI-assisted drone delivery for last-mile transport, (3) predictive analytics for demand forecasting and routing, and (4) frugal AI-enhanced refrigeration systems. Reported outcomes included improved temperature compliance, reduced wastage, and enhanced delivery efficiency, but limited prospective validation and context-specific adaptation. Most pilots originated in high-income settings, while African and other LMIC implementations remained small-scale. Conclusions: AI-driven innovations are reshaping the vaccine supply chain, but their deployment remains uneven and under-evaluated. To achieve equitable impact, integration must align with local infrastructure, governance, and workforce capacities. Building regional data ecosystems, supporting open-source AI tools, and embedding participatory design can transform AI from a technical add-on into a driver of immunization equity and resilience.