Scholarlink

<p>Atrial fibrillation (AF) significantly contributes to the incidence of strokes. Screening for AF enhances its detection and effective management. However, universal AF screening in rural areas poses a challenge. This study evaluates the cost-effectiveness of artificial intelligence-enabled 12-lead electrocardiography (AI-ECG) model for AF screening in rural communities.</p><p>This cost-effectiveness analysis targeted individuals aged 65 or older, employing a lifelong decision analytic Markov model. AI-ECG model, trained and validated at three Taiwanese hospitals with 285,108 patients, achieved sensitivities of 97.8% and specificities of 99.1%. The study incorporated costs and efficacy of anticoagulant treatments, health status utilities, and clinical variables, derived from literature and Taiwan’s epidemiological data. Outcomes were expressed in US dollars per quality-adjusted life year (QALY). The base-case analysis contrasted AI-ECG screening performed by nurses and physician evaluations using standard 12-lead ECGs against no screening, incorporating uncertainty through probabilistic sensitivity analysis. Results were compared with one GDP per capita in Taiwan (≈$32,327 per QALY), a commonly cited willingness-to-pay (WTP) benchmark.</p><p>Both AI-ECG and physician-led screenings were costlier yet more effective compared with no screening. Although both methods showed comparable effectiveness in detecting AF and in QALYs gained, AI-ECG screening was less expensive ($141 versus $196). Based on 5,000 Monte Carlo simulations, AI-based screening is more cost-effective at lower thresholds ($4,349 to $6,132 per QALY), while physician-led screening becomes preferable beyond $6,132 per QALY. Both strategies remained cost-effective relative to the WTP benchmark. Sensitivity analyses further identified the referral rate following a positive AI-ECG screening as a critical determinant of its cost-effectiveness.</p><p>AI-ECG screening for AF is a cost-effective alternative, particularly suitable for areas with limited medical resources.</p> Graphical Abstract <p>Illustration of the results of the cost-effectiveness analysis comparing AF screening methods: AI-ECG, physician-led screening, and no screening.</p> <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Universal Atrial Fibrillation Screening Using Electrocardiographic Artificial Intelligence: A Cost-Effective Approach in Rural Communities

  • 期刊论文

  • 发表日期：

    2025年10月27日

查看全文
  • Wei-Ting Liu,
  • Chin-Sheng Lin,
  • Chin Lin,
  • Tsung-Kun Lin,
  • Wen-Yu Lin,
  • Chiao-Chin Lee,
  • Chiao-Hsiang Chang,
  • Chien-Sung Tsai,
  • Yi-Jen Hung,
  • Ping-Hsuan Hsieh

摘要

Atrial fibrillation (AF) significantly contributes to the incidence of strokes. Screening for AF enhances its detection and effective management. However, universal AF screening in rural areas poses a challenge. This study evaluates the cost-effectiveness of artificial intelligence-enabled 12-lead electrocardiography (AI-ECG) model for AF screening in rural communities.

This cost-effectiveness analysis targeted individuals aged 65 or older, employing a lifelong decision analytic Markov model. AI-ECG model, trained and validated at three Taiwanese hospitals with 285,108 patients, achieved sensitivities of 97.8% and specificities of 99.1%. The study incorporated costs and efficacy of anticoagulant treatments, health status utilities, and clinical variables, derived from literature and Taiwan’s epidemiological data. Outcomes were expressed in US dollars per quality-adjusted life year (QALY). The base-case analysis contrasted AI-ECG screening performed by nurses and physician evaluations using standard 12-lead ECGs against no screening, incorporating uncertainty through probabilistic sensitivity analysis. Results were compared with one GDP per capita in Taiwan (≈$32,327 per QALY), a commonly cited willingness-to-pay (WTP) benchmark.

Both AI-ECG and physician-led screenings were costlier yet more effective compared with no screening. Although both methods showed comparable effectiveness in detecting AF and in QALYs gained, AI-ECG screening was less expensive ($141 versus $196). Based on 5,000 Monte Carlo simulations, AI-based screening is more cost-effective at lower thresholds ($4,349 to $6,132 per QALY), while physician-led screening becomes preferable beyond $6,132 per QALY. Both strategies remained cost-effective relative to the WTP benchmark. Sensitivity analyses further identified the referral rate following a positive AI-ECG screening as a critical determinant of its cost-effectiveness.

AI-ECG screening for AF is a cost-effective alternative, particularly suitable for areas with limited medical resources.

Graphical Abstract

Illustration of the results of the cost-effectiveness analysis comparing AF screening methods: AI-ECG, physician-led screening, and no screening.