دوره 14، شماره 3 - ( 5-1405 )
جلد 14 شماره 3 صفحات 0-0 |
برگشت به فهرست نسخه ها
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Rajaee R, Behzadi A, Vali L, Wakde T, Rahatlou M B, Najafi M. Mapping Artificial Intelligence in Health Policymaking: An Evidence-Mapping Review of Applications, Opportunities, Challenges, and Evidence Gaps. Iran J Health Sci 2026; 14 (3)
URL: http://jhs.mazums.ac.ir/article-1-1173-fa.html
URL: http://jhs.mazums.ac.ir/article-1-1173-fa.html
Mapping Artificial Intelligence in Health Policymaking: An Evidence-Mapping Review of Applications, Opportunities, Challenges, and Evidence Gaps. علوم بهداشتی ایران. 1405; 14 (3)
چکیده: (17 مشاهده)
Background: Artificial intelligence (AI) is increasingly used to analyse population-level data, forecast health needs, support resource allocation and automate parts of health-system decision-making. These opportunities are accompanied by risks related to biased data, privacy, accountability, weak regulation and uneven digital capacity, which are particularly consequential when AI informs population-level policy rather than individual clinical decisions.
Aim: To map and summarize how AI is used across stages of the health policy cycle, identify reported benefits and challenges, and highlight evidence gaps relevant to equity, governance and resource-constrained health systems.
Methods: We conducted an evidence-mapping review, informed by PRISMA-ScR reporting principles, to characterize the distribution and nature of evidence on AI across the health policy cycle. PubMed, Scopus, and Web of Science were systematically searched for English-language publications from 2015 to August 2026, with Google Scholar used as a supplementary source for citation searching. Sources were eligible when they examined an AI application, implementation approach, or governance issue with an explicit link to health-policy agenda setting, formulation/design, implementation, monitoring, or evaluation. Data were charted by AI application, policy-cycle stage, evidence type, benefits, risks, and implementation barriers. Findings were summarized descriptively and presented as an evidence map. No formal quality appraisal was undertaken because the objective was to characterize the distribution and content of the evidence rather than estimate intervention effects.
Results: The searches identified 2,579 records, of which 46 sources met the eligibility criteria. Evidence was concentrated in implementation (n=42) and formulation/design (n=41), while monitoring/evaluation (n=18) and agenda setting (n=11) were less frequently represented. Because sources could address multiple policy-cycle stages, these counts were non-mutually exclusive. Reported opportunities included decision support, forecasting, improved access, workflow support and policy monitoring. Recurrent concerns involved bias, privacy and security, infrastructure, workforce capacity, accountability and regulatory uncertainty. Prospective, longitudinal and comparative implementation evidence remained limited.
Conclusion: The evidence map shows a rapidly expanding but unevenly developed literature on AI in health policymaking. Current evidence is concentrated in policy formulation, implementation, governance, and organizational readiness, with comparatively less evidence on agenda setting, sustained monitoring and evaluation, empirical equity outcomes, and implementation in resource-constrained settings. The findings describe the distribution of available evidence rather than the effectiveness of AI. Future research should prioritize prospective and longitudinal evaluation of AI-supported policy applications under routine health-system conditions.
Aim: To map and summarize how AI is used across stages of the health policy cycle, identify reported benefits and challenges, and highlight evidence gaps relevant to equity, governance and resource-constrained health systems.
Methods: We conducted an evidence-mapping review, informed by PRISMA-ScR reporting principles, to characterize the distribution and nature of evidence on AI across the health policy cycle. PubMed, Scopus, and Web of Science were systematically searched for English-language publications from 2015 to August 2026, with Google Scholar used as a supplementary source for citation searching. Sources were eligible when they examined an AI application, implementation approach, or governance issue with an explicit link to health-policy agenda setting, formulation/design, implementation, monitoring, or evaluation. Data were charted by AI application, policy-cycle stage, evidence type, benefits, risks, and implementation barriers. Findings were summarized descriptively and presented as an evidence map. No formal quality appraisal was undertaken because the objective was to characterize the distribution and content of the evidence rather than estimate intervention effects.
Results: The searches identified 2,579 records, of which 46 sources met the eligibility criteria. Evidence was concentrated in implementation (n=42) and formulation/design (n=41), while monitoring/evaluation (n=18) and agenda setting (n=11) were less frequently represented. Because sources could address multiple policy-cycle stages, these counts were non-mutually exclusive. Reported opportunities included decision support, forecasting, improved access, workflow support and policy monitoring. Recurrent concerns involved bias, privacy and security, infrastructure, workforce capacity, accountability and regulatory uncertainty. Prospective, longitudinal and comparative implementation evidence remained limited.
Conclusion: The evidence map shows a rapidly expanding but unevenly developed literature on AI in health policymaking. Current evidence is concentrated in policy formulation, implementation, governance, and organizational readiness, with comparatively less evidence on agenda setting, sustained monitoring and evaluation, empirical equity outcomes, and implementation in resource-constrained settings. The findings describe the distribution of available evidence rather than the effectiveness of AI. Future research should prioritize prospective and longitudinal evaluation of AI-supported policy applications under routine health-system conditions.
نوع مطالعه: مروری |
موضوع مقاله:
مدیریت خدمات بهداشتی درمانی
| بازنشر اطلاعات | |
 |
این مقاله تحت شرایط Creative Commons Attribution-NonCommercial 4.0 International License قابل بازنشر است. |
