Value in Health最新文献

{"title":"Mapping methodologies for economic evaluation of digital health technologies: a scoping review.","authors":"Ana Rita Santos, Francisco von Hafe, Filipa Sampaio, Ana Rita Londral, Julian Perelman","doi":"10.1016/j.jval.2025.09.012","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.012","url":null,"abstract":"<p><strong>Objectives: </strong>Digital Health Technologies (DHTs) are reshaping healthcare delivery, yet their diverse functionalities, dynamic nature, and non-traditional impact pathways challenge conventional economic evaluation methods. This scoping review aimed to systematically map existing frameworks for the economic evaluation of DHTs, assess their methodological components, and identify gaps to inform more robust, standardised approaches.</p><p><strong>Methods: </strong>Following PRISMA-ScR guidelines, we conducted a comprehensive literature search across academic databases (PubMed, Scopus, Web of Science, and Business Source Complete) and grey literature. Records published since 2015 were included if they described frameworks incorporating economic evaluation methods for DHTs. Data were extracted across key methodological dimensions: evaluation type, study design, comparator, time horizon, perspective, effectiveness measures, cost components, and uncertainty analysis.</p><p><strong>Results: </strong>We identified 26 frameworks, and the analysis revealed pronounced heterogeneity across methodological domains. While 50% (n=13) included both full and partial evaluations, core components were often missing: 81% (n=21) did not define a time horizon, and in 73% (n=19) the evaluation perspective was absent. Cost-utility analysis and budget impact analysis were the most frequently cited methods, yet few frameworks justified their choice or linked it to the maturity of the technology. Some addressed adaptive study designs or aligned evaluation strategies with DHTs lifecycle stages. Cost inclusions varied substantially, with limited attention to productivity losses costs.</p><p><strong>Conclusions: </strong>Current frameworks lack standardisation and are not fully adapted to the characteristics of DHTs. Future development should prioritise flexible, lifecycle-aligned evaluation models and standardised guidance to support evidence-based digital health decision-making.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145182196","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Preferences for adopting artificial intelligence in radiation therapy treatment: A discrete choice experiment.","authors":"Milena Lewandowska, Deborah Street, Jackie Yim, Scott Jones, Rosalie Viney","doi":"10.1016/j.jval.2025.09.013","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.013","url":null,"abstract":"<p><strong>Objectives: </strong>The integration of artificial intelligence (AI) in radiation therapy offers significant potential to enhance cancer care by improving diagnostic accuracy, streamlining workflows, and reducing treatment delays. However, the adoption of AI in clinical settings depends heavily on its acceptability, shaped by perceptions of accuracy, cost, efficiency, and ethical considerations. This study explores the preferences of the Australian general population regarding the features of AI systems in radiation therapy.</p><p><strong>Methods: </strong>A discrete choice experiment (DCE) was conducted with 533 respondents, who were representative of the Australian population. Participants were presented with hypothetical scenarios comparing AI systems described by attributes including accuracy, decision-making autonomy, impact on out-of-pocket costs, treatment timelines, and data privacy. Preferences were analysed using mixed logit and latent class models to evaluate heterogeneity and willingness-to-pay for AI system attributes.</p><p><strong>Results: </strong>Respondents preferred AI systems with enhanced accuracy and reduced treatment delays. Systems less likely to misclassify tissues were highly valued, while fully autonomous AI systems were less favoured compared to assistive systems requiring clinician oversight. Data privacy concerns varied, with some participants prioritizing consent-based data usage. Heterogeneity analysis revealed four distinct preference classes, highlighting trade-offs between cost, speed, and ethical considerations. Willingness-to-pay estimates showed the respondents were willing to pay features such as enhanced oversight, reduced time burden, and AI-driven data use.</p><p><strong>Conclusions: </strong>This study provides important information about Australian general public preferences for AI in treatment planning and can be used to inform future research on economic evaluations and implementation of AI-driven technologies in radiation therapy.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145182194","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Derivation of Meaningful Change Thresholds for EORTC QLQ-C30 and QLQ-LC13 in Patients with Small Cell Lung Cancer.","authors":"Franziska Dirnberger, Jessie Wang, Bellinda King-Kallimanis, Kim Cocks, George Skingley, Nathan Clarke, Shuang Huang, Sujoy Mukherjee, Michael Koller","doi":"10.1016/j.jval.2025.09.011","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.011","url":null,"abstract":"<p><strong>Objectives: </strong>Small-cell lung cancer (SCLC) is associated with poor survival and is accompanied by symptoms that impact patients' quality of life. Although quality of life can be assessed by the EORTC 30-item core questionnaire (QLQ-C30) and 13-item lung cancer (QLQ-LC13) module, interpretation of meaningful improvement or deterioration within and between individuals is challenging. The objective of this study was to derive thresholds for interpretation of changes over time for the key scales of Coughing, Chest Pain, Dyspnea, Physical Functioning and Global Health Status (GHS).</p><p><strong>Methods: </strong>Anchor- and distribution-based analyses were employed to derive meaningful within-patient change (MWPC) and meaningful between-group difference (MBGD) thresholds using data from the phase II DeLLphi-301 trial in patients with previously treated SCLC. Methods included estimating mean change scores, mixed models for repeated measures, classification statistics, receiver operating characteristic curves, alongside calculating half standard deviation and standard error of measurement.</p><p><strong>Results: </strong>MWPC recommendations were ±33.3 for single item scales QLQ-LC13 Chest Pain and Coughing, ±16.7 for both QLQ-LC13 Dyspnea and QLQ-C30 GHS, and ±20.0 for QLQ-C30 Physical Functioning. Estimated MBGD thresholds ranged from ±6 for GHS to ±22 for Coughing.</p><p><strong>Conclusion: </strong>This study is the first to estimate MWPC and MBGD thresholds specifically in patients with relapsed SCLC. Estimated threshold ranges are considered robust given the comprehensive considerations and triangulation of various methods and can help clinicians, researchers, and healthcare decision makers to evaluate the clinical meaningfulness to patients of changes in key SCLC scales over time.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145182237","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Means Can Be Deceiving: Comparing and Contrasting Risk-Tolerance Estimates from a Discrete-Choice Experiment and a Threshold Technique Exercise.","authors":"Jessie Sutphin, Matthew J Wallace, Shelby D Reed","doi":"10.1016/j.jval.2025.09.006","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.006","url":null,"abstract":"<p><strong>Objective: </strong>This study compared quantitative measures of risk tolerance between two preference-elicitation methods; a discrete-choice experiment (DCE) and a probabilistic threshold technique (TT) exercise.</p><p><strong>Methods: </strong>A survey offered benefit-risk tradeoffs pertaining to devices used in revascularization procedures for peripheral artery disease. Survey-design features included alternating the sequence of DCE and TT exercises, testing two risk-communication approaches, and using two DCE experimental designs. The risk-tolerance metric was the maximum-acceptable risk (MAR) increase in 5-year mortality above 8% that patients would accept to choose a device offering lower repeat-procedure risks. DCE data were analyzed with mixed-logit models, and TT data were analyzed with interval regressions. Sensitivity analyses were conducted to examine the impacts of survey-design features.</p><p><strong>Results: </strong>For the full sample (N=249), MARs from the DCE and TT differed by <1 percentage-point (DCE: 13.0%; TT: 13.8%). MAR estimates were not significantly influenced by the risk-communication approach or experimental-design; however, order of DCE and TT mattered. After removing the influence of DCE and TT order, the mean MARs still differed by <1 percentage-point (MAR among those completing DCE first (n=123): 14.3%; MAR among those completing TT first (n=126): 13.7%). Although convergent validity was found at the sample level, discordance (> 2 percentage points) between individual-level DCE MAR and TT MAR were observed for approximately half (48.2%) of the sample.</p><p><strong>Conclusions: </strong>While there was concordance between DCE and TT mean MAR estimates at the sample level, at the individual level, only half of the respondents had an absolute difference of 2 percentage-points or less.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145182265","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparing Psychometric Properties of Six 5-Level and 3-Level EQ-5D Bolt-Ons in a Large, Multinational, Longitudinal General Population Sample.","authors":"Sarah Dewilde, Mathieu F Janssen, Glenn Phillips, Brendan Mulhern, Fanni Rencz","doi":"10.1016/j.jval.2025.09.009","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.009","url":null,"abstract":"<p><strong>Objective: </strong>This study systematically compared the psychometric properties of six 5-level and 3-level EQ-5D bolt-ons (vision, breathing, tiredness, sleep, social relationships, and self-confidence).</p><p><strong>Methods: </strong>POPUP is a longitudinal survey collecting general population data from US, Canada, UK, Italy, Spain, Germany, Netherlands, and Belgium. Respondents entered demographics, health conditions data, and completed the EQ-5D-5L/3L with bolt-ons, MG-ADL, and HUI-3. Psychometric properties included ceiling/floor, informativity, divergent and convergent validity, reliability, known-groups validity, and responsiveness.</p><p><strong>Results: </strong>9,758 respondents completed the survey in 2021, and 4,839 respondents in 2023. The 5-level bolt-ons reduced ceiling effects by 35% and floor effects by 55% compared to 3-level, with the largest reductions for vision and sleep (42%/57%) and the smallest for breathing (29%/44%). Informativity was higher for the 5-level bolt-ons (3%-11%) than 3-level (2%-9%), except for breathing (-2%). Most bolt-ons showed weak to moderate correlations with EQ-5D dimensions, except for social relationships, which correlated strongly with anxiety/depression (r=0.61). Both measures showed strong convergent validity for breathing (5-level r=0.68; 3-level r=0.66) and lower for vision (5-level r=0.37; 3-level r=0.32). EQ-5D+5-level and EQ-5D-5L+3-level bolt-ons had excellent reliability (Intraclass Correlation Coefficients=0.92-0.97). Adding bolt-ons did not significantly improve known-groups validity, except for breathing (relative efficiency EQ-5D+5-level=1.12; EQ-5D-5L+3-level=1.10). The 5-level bolt-ons improved responsiveness, detecting 47.1% more improvements and 55.7% more worsening over time.</p><p><strong>Conclusions: </strong>The six 5-level EQ-5D bolt-ons outperformed the 3-level bolt-ons by reducing ceiling and floor, enhancing informativity, demonstrating divergent validity, and improving responsiveness.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145178828","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Patient Preference-Weighted Quantitative Benefit-Risk Analysis of Ritlecitinib for Alopecia Areata to Inform Medical Decision Making.","authors":"Jonathan Mauer, Chiara Whichello, Brett Hauber, Nicolas Krucien, Ernest H Law, Myrto Trapali, Maryanne Senna, Samuel H Zwillich, Robert Wolk, Tommi Tervonen","doi":"10.1016/j.jval.2025.09.007","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.007","url":null,"abstract":"<p><strong>Objectives: </strong>Inhibition of JAK-mediated signaling pathways leads to hair regrowth and improved quality of life in alopecia areata (AA) but may be associated with potential risks. This study assessed the benefit-risk profile of ritlecitinib, an oral selective JAK3/TEC family kinase inhibitor (JAK3/TEC), from the perspective of adult patients with AA to inform regulatory decisions.</p><p><strong>Methods: </strong>We combined patient preference data from a discrete choice experiment (DCE) involving 201 adults with AA in the US and Europe, efficacy data for ritlecitinib 50 mg once daily (QD) and placebo from the ALLEGRO phase 2b/3 trial (NCT03732807) of ritlecitinib efficacy and safety, and safety data from an integrated safety analysis of four studies in AA. The DCE included 6 treatment attributes: 3 benefits (chance of achieving ≥80% scalp hair coverage, moderate or normal eyebrows, and moderate or normal eyelashes) and 3 risks (3-year risks of serious infections, blood clots, and cancer). Choice probabilities were calculated using the estimated utility function. Weighted net benefit scores were calculated using MCDA. Stochastic multicriteria acceptability analysis (SMAA) assessed the impact of uncertainty and preference heterogeneity on net benefit conclusions.</p><p><strong>Results: </strong>Ritlecitinib 50 mg had a higher weighted net benefit score than placebo with a predicted choice probability of 70.9% (95% CI, 61.2%-77.5%), indicating that on average patients would likely evaluate benefits to outweigh risks. The probability of positive weighted net benefit based on the SMAA was 78%.</p><p><strong>Conclusions: </strong>Ritlecitinib 50 mg QD demonstrated a positive benefit-risk profile compared with placebo in adult patients with AA.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145178842","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cost-effectiveness of RADAR: An Innovative Model to Organize Diabetes Care in First Nations Communities in Canada.","authors":"Priyanka Ghule, Ishfaq Rashid, Jasjeet K Minhas-Sandhu, Lynden Crowshoe, Jeffrey A Johnson, Salim Samanani, Cerina Dubois, Carl V Asche, Dean T Eurich","doi":"10.1016/j.jval.2025.09.010","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.010","url":null,"abstract":"<p><strong>Introduction: </strong>The Reorganizing the Approach to Diabetes through the Application of Registries (RADAR) is an innovative care model incorporating a disease registry and electronic health record tailored for First Nations (FN) people in Alberta, Canada. Given its success in improving health outcomes, this study evaluates RADAR's cost-effectiveness compared to standard diabetes care in FN populations with type 2 diabetes (T2DM).</p><p><strong>Methods: </strong>The United Kingdom Prospective Diabetes Study Outcomes Model 2 was used to project clinical and cost outcomes over 25 years. Demographics and clinical parameters were used as input parameters. Costs (for RADAR and complications) and utility estimates were derived from the literature. Outcome measures included life years, quality-adjusted life years (QALY's), total and complication costs. Base case analysis was conducted for eligible T2DM patients ≥ 18 years using a Canadian payer perspective with a 25-year time horizon. Future costs and effects were discounted at 3% annually.</p><p><strong>Results: </strong>RADAR clients experienced higher QALYs (12.69 vs. 12.47) than those receiving baseline care. The total cost for RADAR clients (n=516) was $3,971 more per client compared to baseline care. Over the modeled time horizon, the total discounted costs (client management and complications) were higher for RADAR clients than for those receiving baseline care ($16,165 vs. $14,463), yielding an ICER of $8,105 per QALY for RADAR users relative to baseline.</p><p><strong>Conclusion: </strong>Our analyses showed that RADAR is potentially cost-effective if the willingness to pay a threshold of $50,000 per QALY is considered as per the CADTH guidelines, supporting its value in FN care.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145178770","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Validating Loon Lens 1.0 for Autonomous Abstract Screening and Confidence-Guided Human-in-the-Loop Workflows in Systematic Reviews.","authors":"Ghayath Janoudi, Mara Uzun, Tim Disher, Mia Jurdana, Ena Fuzul, Josip Ivkovic, Brian Hutton","doi":"10.1016/j.jval.2025.09.008","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.008","url":null,"abstract":"<p><strong>Objectives: </strong>Title and Abstract (TiAb) screening is a labour-intensive step in systematic literature reviews (SLR). We examine the performance of Loon Lens 1.0, an agentic AI platform for autonomous TiAb screening and test whether its confidence scores can target minimal human oversight.</p><p><strong>Methods: </strong>Eight SLRs by Canada's Drug Agency were re-screened through dual-human reviewers and adjudicated process (3,796 citations, 287 includes, 7.6%) and separately by Loon Lens, based on predefined eligibility criteria. Accuracy, sensitivity, precision, and specificity were measured and bootstrapped to generate 95% confidence intervals. Logistic regression with (i) confidence alone and (ii) confidence + Include/Exclude decision predicted errors and informed simulated human-in-the-loop (HITL) strategies.</p><p><strong>Results: </strong>Loon Lens achieved 95.5% accuracy (95% CI 94.8-96.1), 98.9% sensitivity (97.6-100), 95.2% specificity (94.5-95.9) and 63.0% precision (58.4-67.3). Errors clustered in Low-Medium-confidence Includes. The extended logistic regression model (confidence + decision; C-index 0.98) estimated a 75% error probability for Low-confidence Includes versus <0.1% for Very-High-confidence Excludes. Simulated HITL review of Low + Medium-confidence Includes only (145 citations, 3.8%), lifted precision to 81.4% and overall accuracy to 98.2% while preserving sensitivity (99.0%). Adding High-confidence Includes (221 citations, 5.8%) pushed precision to 89.9% and accuracy to 99.0%.</p><p><strong>Conclusions: </strong>Across eight SLRs (3,796 citations), Loon Lens 1.0 reproduced adjudicated human screening with 98.9% sensitivity and 95.2% specificity. In simulation, restricting human-in-the-loop review to ≤5.8% of citations, by prioritising low- and medium-confidence Include calls, reduced false positives and increased precision to 89.9% while maintaining sensitivity and raising overall accuracy to 99.0%. These findings indicate that confidence-guided oversight can concentrate reviewer effort on a small subset of records.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145151085","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The cost-effectiveness of community health workers in primary health care: a systematic review.","authors":"Tijs Van Iseghem, Laura Vroonen, Emilie Op de Beeck, Annick Meertens, Caroline Masquillier, Edwin Wouters, Nick Verhaeghe","doi":"10.1016/j.jval.2025.09.004","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.004","url":null,"abstract":"<p><strong>Background: </strong>Global interest in community health worker (CHW) programs in primary health care (PHC) is rising due to their potential to advance universal health coverage and other global health goals. This systematic review examines the evidence on the cost-effectiveness of CHW interventions worldwide, with a focus on vulnerable populations in PHC settings.</p><p><strong>Methods: </strong>A systematic review was conducted using the PubMed, Embase, Web of Science Core Collection, SCOPUS and EconLit databases. The search was last updated on May 13, 2025. Two reviewers independently selected articles, rated their quality and extracted relevant data. Included articles had to be full economic evaluations comparing CHW interventions to usual care without CHWs, focussing on vulnerable populations in PHC. A standardized data extraction template was used and the reporting quality was assessed using the CHEERS checklist.</p><p><strong>Results: </strong>Fifty articles were included, originating from 25 countries. All but one of the economic evaluations were disease-specific, focusing mainly on maternal, newborn and child health, type 2 diabetes mellitus, tuberculosis, cardiovascular disease, mental health, and HIV. Most articles (n = 35, 70%) indicated that CHW interventions were (potentially) cost-effective. Fourteen out of these 35 articles can substantiate their findings with a probabilistic sensitivity analysis (PSA).</p><p><strong>Conclusion: </strong>This review found that CHWs can be cost-effective across multiple health domains, both in LMICs and HICs. The lack of PSA, together with heterogeneity in contexts, interventions and methods used to assess the cost-effectiveness of the CHW interventions makes it difficult to draw general conclusions about the value-for-money of CHWs in PHC.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145126097","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Long-Term Outcomes and Cost-Effectiveness of Artificial Intelligence for Breast Cancer Screening: A Modeling Study.","authors":"Matthew Andersen, Ilana B Richman, Natalia Kunst","doi":"10.1016/j.jval.2025.09.005","DOIUrl":"https://doi.org/10.1016/j.jval.2025.09.005","url":null,"abstract":"<p><strong>Background: </strong>Artificial intelligence (AI)-assisted breast cancer screening may improve diagnostic accuracy, but the long-term health outcomes and cost-effectiveness of AI assisted screening is unknown. We estimated benefits, harms, and cost-effectiveness of incorporating an AI product, Saige-DX, into standard screening with digital breast tomosynthesis (DBT).</p><p><strong>Methods: </strong>We developed a microsimulation model using nationally representative data from SEER (Surveillance, Epidemiology, and End Results), the Breast Cancer Surveillance Consortium, and published data on AI performance. The model compared biennial screening for women ages 40-74 using DBT to DBT plus AI. We estimated false positive and false negative screens, breast cancer cases by stage, and breast cancer deaths per 1000 women screened over a lifetime. We also estimated quality-adjusted life years (QALYs), costs, and the incremental cost effectiveness ratio (ICER).</p><p><strong>Results: </strong>In a cohort of 1000 women screened from ages 40-74, AI-assisted screening reduced false negative screens by 2.1 and false positives by 50 resulting in 0.33 fewer advanced breast cancer cases (regional or metastatic cancer) at diagnosis and 0.13 fewer breast cancer deaths compared to DBT alone. Screening with AI resulted in 3.09 additional QALYs and an increase in lifetime costs of $936,430 per 1000 women, yielding an ICER of $303,279 per QALY. In 98% of simulations, AI was not cost-effective at a $100,000/QALY willingness-to-pay threshold. Findings were not sensitive to changes in test characteristics likely to be observed in routine practice.</p><p><strong>Conclusions: </strong>AI-assisted breast cancer screening yielded modest reductions in breast cancer mortality, but at current pricing, AI is not cost-effective.</p>","PeriodicalId":23508,"journal":{"name":"Value in Health","volume":" ","pages":""},"PeriodicalIF":6.0,"publicationDate":"2025-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145126152","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}