AI Readiness for Health Systems as a Policy Capacity Problem in Ghana and LMICs.

• In Event: AI in Health Systems and Healthcare Innovation

Abstract

Artificial intelligence (AI) is increasingly positioned as a solution to health system constraints in low- and middle-income countries (LMICs), yet national readiness for AI adoption is often assessed using narrow technical indicators or high-level governance checklists that provide limited guidance for policy sequencing. As a result, governments face pressure to pursue AI-enabled reforms without clear evidence of whether underlying institutional conditions can support implementation at scale. This paper treats readiness not as a technological threshold but as a structured policy capacity problem shaped by financing systems, workforce availability, regulatory coordination, and delivery infrastructure.

The study develops a multidimensional readiness framework that integrates seven domains relevant to national adoption: digital infrastructure, data governance arrangements, workforce capability, health service delivery capacity, procurement and financing systems, regulatory and ethical oversight, and domestic AI ecosystem development. Ghana is examined as a focal case and situated within a broader comparative assessment of Sub-Saharan African LMICs and selected peer countries between 2020 and 2026 using harmonised secondary indicators. Domain-level benchmarking and cluster analysis are used to identify patterns of institutional readiness rather than rank countries along a single technological trajectory.

Results show that Ghana’s readiness profile reflects relatively strong policy coordination and regulatory positioning alongside persistent constraints in workforce availability, financing structures, and service-delivery infrastructure required for operational deployment. Comparative clustering further indicates that LMIC health systems exhibit distinct readiness configurations that imply different policy pathways for integrating AI, rather than converging toward a common sequence of adoption stages.

These findings suggest that treating AI readiness as a linear infrastructure problem risks encouraging premature deployment strategies. A capacity-orientated approach instead supports more realistic prioritisation by identifying which institutional investments are prerequisites for sustainable implementation. The analysis provides a structured basis for policymakers to sequence AI adoption in ways that align with existing system capabilities and governance arrangements in resource-constrained health systems.

Author

• Kwabena Ampong Kutin-Buah, University College London
  Presenting Author