Evaluating dementia risk prediction in mild cognitive impairment: an early health technology assessment of the AI-Mind tool
This study evaluated the cost-effectiveness of an AI-based prediction tool (AI-Mind) for assessing dementia conversion risk in individuals with mild cognitive impairment (MCI). Findings suggest that without highly effective disease-modifying treatments, the tool leads to lower quality-adjusted life years (QALYs) and higher costs due to the negative impact of perceived dementia risk on quality of life. However, when combined with targeted, cost-effective treatment strategies and specific prediction characteristics, AI-Mind shows potential for cost-effectiveness, especially with shorter prediction horizons and high sensitivity/specificity.
This study aimed to evaluate the potential cost-effectiveness of implementing a prediction tool for estimating mild cognitive impairment (MCI) to dementia conversion risk. A decision-analytic model was developed to compare the costs and effects of current practice for subjects with MCI to a situation in which the risk of dementia is estimated using a prediction tool. Different scenarios in terms of prediction horizons, prediction characteristics (e.g. sensitivity and specificity), and treatment availability were evaluated. The model was applied to the AI-Mind tool, which is currently under development for predicting MCI to dementia risk. In a clinical situation, with no widely applicable and highly effective disease-modifying treatment available, implementing a dementia risk prediction tool leads to lower QALYs and higher costs compared to current practice without such a prediction tool (9.32 vs 9.36 QALYs and €115,837 vs €115,032 for the analyses in this paper). This loss in QALYs was caused by the impact on quality of life associated with predicted dementia conversion risk. Risk prediction followed by efficient treatment strategies based on the predicted risk could lead to a cost-effective alternative in case of specific treatment characteristics. These findings suggest that standalone (i.e. without highly effective treatment options) use of a dementia risk prediction tool may not be cost-effective, but it could result in a cost-effective alternative in combination with a treatment with favourable efficacy and cost profile.
Executive Impact: Key Metrics at a Glance
Understanding the immediate and quantifiable effects on your enterprise.
0
Increased Costs (AI-Mind vs. Current Practice)
0
Lower QALYs (AI-Mind vs. Current Practice)
Deep Analysis & Enterprise Applications
Select a topic to dive deeper, then explore the specific findings from the research, rebuilt as interactive, enterprise-focused modules.
Enterprise Process Flow
Current Practice: Diagnostic + Prognostic Tests
→
Prognostic Information Shared
→
Treatment + Follow-up
→
Personnel Involved
→
Dementia Risk Prediction Tool: Decision Model Input (Costs, QoL, Survival, Prediction Performance)
→
Decision Model Output (Costs, Effects, Difference)
→
Results: Lower QALYs, Higher Costs (without effective treatment)
The decision-analytic model compares current practice with the AI-Mind tool, evaluating costs and effects across different stages, from diagnostic testing to long-term follow-up and outcomes in QALYs.
| Comparison Point | Scenario: 2-Year Horizon, 95% Sensitivity/Specificity | Scenario: 2-Year Horizon, 80% Sensitivity/Specificity |
|---|---|---|
| QALYs (AI-Mind vs. Current Practice) |
|
|
| ICER (€/QALY) |
|
|
A shorter prediction horizon (2 years) with high sensitivity/specificity (95%) shows AI-Mind approaching cost-effectiveness, while lower performance (80%) makes current practice dominant.
| Comparison Point | Scenario: 5-Year Horizon, 95% Sensitivity/Specificity, 8% Eligible for Treatment | Scenario: 5-Year Horizon, 95% Sensitivity/Specificity, Universal Eligibility, Reduced Cost |
|---|---|---|
| Cost-Effectiveness Outcome |
|
|
Targeted treatment with lecanemab under strict eligibility criteria doesn't make AI-Mind cost-effective. However, with universal eligibility and reduced treatment costs, AI-Mind becomes dominant.
Disutility of High-Risk Prediction
One of the key findings is that individuals receiving a high-risk prediction experience a negative impact on quality of life (disutility of -0.18). This often outweighs the utility gained from a low-risk prediction (+0.06), especially when no effective treatment is available. This disutility contributes significantly to the tool's overall impact on QALYs and its cost-effectiveness.
Key Learnings: The psychological impact of prognostic information is a critical factor influencing QALYs and must be carefully considered in health technology assessments for prediction tools.
Calculate Your Potential AI-Driven ROI
Estimate the financial and operational benefits of integrating AI into your enterprise workflows.
Your AI Implementation Roadmap
A structured approach to integrating advanced AI capabilities into your enterprise.
Phase 1: Discovery & Strategy Alignment
Initial consultations to understand your current infrastructure, business goals, and identify high-impact AI opportunities. Develop a tailored strategy roadmap.
Phase 2: Data Preparation & Model Development
Cleanse, transform, and prepare your enterprise data. Develop custom AI models using state-of-the-art techniques, ensuring optimal performance and relevance.
Phase 3: Integration & Pilot Deployment
Seamlessly integrate AI solutions into existing systems. Conduct pilot programs to test performance, gather feedback, and iterate in a controlled environment.
Phase 4: Full-Scale Rollout & Optimization
Deploy AI solutions across your enterprise. Continuously monitor, evaluate, and optimize performance for maximum ROI and sustained competitive advantage.
Ready to Transform Your Enterprise with AI?
Connect with our AI specialists to discuss how these insights can be applied to your unique business challenges and opportunities.
Ready to Get Started?
Book Your Free Consultation.
Let's Discuss Your AI Strategy!
Lets Discuss Your Needs
Select a Date
Sun
Mon
Tue
Wed
Thu
Fri
Sat