UNCERTAINTY PROBABILITY THRESHOLDS AND INSUFFICIENT ITERATIONS OF PROBABILISTIC SENSITIVITY ANALYSIS IN ECONOMIC EVALUATION
Author(s)
Yu Xin, Master1, Yilin Deng, Master2, Jun Hao, PhD3, Changjin Wu, PhD4, Yuanyi Cai, PhD5, ling Zuo, PhD6, Wen Hui, PhD1.
1Department of Science and Technology, West China Hospital of Sichuan University, Chengdu, China, 2Department of Health Behavior and Social Medicine, West China School of Public Health and West China, Sichuan University, Chengdu, China, 3Clinical Trial Research Center, The First Affiliated Hospital of Nanchang University, Nanchang, China, 4College of Public Health, Chongqing Medical University, Chongqing, China, 5Department of Health Service Management, School of Health Management, China Medical University, Shenyang, China, 6Integrated Care Management Centre, Outpatient Department, West China Hospital of Sichuan University, Chengdu, China.
1Department of Science and Technology, West China Hospital of Sichuan University, Chengdu, China, 2Department of Health Behavior and Social Medicine, West China School of Public Health and West China, Sichuan University, Chengdu, China, 3Clinical Trial Research Center, The First Affiliated Hospital of Nanchang University, Nanchang, China, 4College of Public Health, Chongqing Medical University, Chongqing, China, 5Department of Health Service Management, School of Health Management, China Medical University, Shenyang, China, 6Integrated Care Management Centre, Outpatient Department, West China Hospital of Sichuan University, Chengdu, China.
OBJECTIVES: This cross-sectional study investigated the application and reporting of probabilistic sensitivity analysis (PSA) in recent economic evaluations, with a focus on the consistency between reported probability thresholds and stated cost-effectiveness conclusions.
METHODS: We searched PubMed for full economic evaluations published between January 1, 2023, to March 18, 2025, in journals ranked within the top 5% by Journal Impact Factor across 51 Clinical Medicine subcategories according to the Journal Citation Reports. Two independent reviewers extracted data on study design, perspective, research domain, outcome type, specific PSA methods, iteration details, and corresponding conclusions.
RESULTS: From 591 initially retrieved studies, 363 met the inclusion criteria. Of these, 302 studies (83.20%) employed PSA. However, over 30% did not specify the PSA method used, and approximately 15% did not report the probability. The majority of studies performed 1,000 (35.10%, 106/302) or 10,000 (33.77%, 102/302) iterations, though nearly 3% used fewer than 1,000 iterations with 80% of these preliminarily confirmed as insufficient. Among studies using health-adjusted life years (HALYs) with a predefined cost-effectiveness threshold, 80.53% of studies concluded the intervention was “cost-effective,” while 19.91% deemed it “not cost-effective.” Importantly, the median minimum probability in “cost-effective” studies was 86.95% (IQR: 66.40%-99.00%), whereas the median maximum probability in “not cost-effective” studies was 20.00% (IQR: 4.30%-37.00%). However, notable inconsistencies were observed: 38.45% of “cost-effective” studies reported probabilities below 75%, 13.72% below 60%, and 6.04% below 50%. Conversely, 11.11% of “not cost-effective” studies reported probabilities above 50%.
CONCLUSIONS: This study demonstrated that similar reported probability can lead to divergent interpretations of economic conclusions due to the absence of a consensus on PSA probability cutoffs. Furthermore, insufficient iteration numbers may compromise the methodological rigor and reliability of economic evaluations.
METHODS: We searched PubMed for full economic evaluations published between January 1, 2023, to March 18, 2025, in journals ranked within the top 5% by Journal Impact Factor across 51 Clinical Medicine subcategories according to the Journal Citation Reports. Two independent reviewers extracted data on study design, perspective, research domain, outcome type, specific PSA methods, iteration details, and corresponding conclusions.
RESULTS: From 591 initially retrieved studies, 363 met the inclusion criteria. Of these, 302 studies (83.20%) employed PSA. However, over 30% did not specify the PSA method used, and approximately 15% did not report the probability. The majority of studies performed 1,000 (35.10%, 106/302) or 10,000 (33.77%, 102/302) iterations, though nearly 3% used fewer than 1,000 iterations with 80% of these preliminarily confirmed as insufficient. Among studies using health-adjusted life years (HALYs) with a predefined cost-effectiveness threshold, 80.53% of studies concluded the intervention was “cost-effective,” while 19.91% deemed it “not cost-effective.” Importantly, the median minimum probability in “cost-effective” studies was 86.95% (IQR: 66.40%-99.00%), whereas the median maximum probability in “not cost-effective” studies was 20.00% (IQR: 4.30%-37.00%). However, notable inconsistencies were observed: 38.45% of “cost-effective” studies reported probabilities below 75%, 13.72% below 60%, and 6.04% below 50%. Conversely, 11.11% of “not cost-effective” studies reported probabilities above 50%.
CONCLUSIONS: This study demonstrated that similar reported probability can lead to divergent interpretations of economic conclusions due to the absence of a consensus on PSA probability cutoffs. Furthermore, insufficient iteration numbers may compromise the methodological rigor and reliability of economic evaluations.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
P37
Topic
Methodological & Statistical Research
Disease
No Additional Disease & Conditions/Specialized Treatment Areas