NATIONAL Cancer Control Plans (NCCPs) demonstrated considerable variation in comprehensiveness across 50 countries, with financing, outcome measurement, and contextual analysis emerging as the weakest areas, according to a new AI-supported assessment. The findings highlight opportunities to strengthen cancer policy development through standardised evaluation and scalable analytical approaches.
Published findings from the assessment used the policy reasoning integrated sequential model (PRISM), a validated multiagent AI system, to evaluate the content and comprehensiveness of national cancer strategies. The approach aimed to address limitations associated with traditional expert-led reviews, which can be resource intensive and restrict comparisons between countries.
AI Supports a Standardised Assessment Approach
Researchers identified the most recent eligible NCCP for each of 50 countries, drawing on international repositories and government websites. They developed an analytical framework comprising 12 elements and 76 sub-elements, informed by international cancer control guidance. The framework was validated through a modified two-stage Delphi process involving an online survey of 67 experts and a workshop attended by 31 participants.
PRISM combined document ingestion, AI-supported policy analysis using retrieval-augmented generation, and semantic quality assurance. Its performance was assessed against independent dual-expert ratings of six selected cancer control plans before being applied across the full dataset.
Financing and Outcome Measurement Lag Behind
The median overall comprehensiveness score was 2.78 on a five-point scale, indicating moderate performance overall. European countries achieved the highest regional median (2.96), while the Western Pacific recorded the lowest (2.21). Among income groups, lower-middle-income countries had the highest median score (2.97), compared with 2.05 in low-income countries.
At the individual element level, implementation achieved the highest median score (4.17), followed by health services (3.89) and governance and organisation (3.67). Conversely, financing (1.88), outcomes (1.83), and threats (2.19) recorded the lowest scores. Financing represented a particularly substantial gap, with five of its 11 sub-elements receiving median scores of 1.0 or below.
Opportunities to Strengthen Global Cancer Policy
The findings suggest that greater attention to sustainable financing, measurable outcomes, and contextual factors could improve the comprehensiveness of national cancer strategies.
The researchers propose that shared health systems frameworks and scalable AI assessment tools, such as PRISM, could support more consistent comparisons between countries. Rapid appraisal may also help policymakers develop timely, achievable, evidence-informed, and context-specific cancer control plans that incorporate the latest available data.
Reference
Kumar CJ et al. Assessment of comprehensiveness of national cancer control plans using a multiagent artificial intelligence model: a global analysis. The Lancet Oncology. 2026;27(11): p1393-402.