The time horizons of formal decision analyses

Area of Science:

• Health economics
• Biostatistics
• Clinical decision analysis

Background:

• Markov models are commonly used in lifetime clinical decision analyses.
• These models simulate patient lifespan using discrete time cycles.
• Results are typically presented as life expectancy.

Purpose of the Study:

• To highlight two overlooked limitations of lifetime Markov models.
• To emphasize the impact of constant disease-specific hazards.
• To advocate for alternative methods of result presentation.

Main Methods:

• Review of limitations in lifetime Markov models.
• Discussion of data limitations for disease-specific hazard estimation.
• Exploration of alternative time horizons and outcome measures.

Main Results:

• Lifetime Markov models often assume constant disease-specific hazards due to data scarcity.
• Presenting results solely as life expectancy may obscure important health state transitions.
• Constant hazard assumption can lead to inaccurate long-term predictions.

Conclusions:

• Restricting time horizons and presenting results as health states can improve model accuracy.
• Using time-variant disease-specific hazards derived from long-term data is recommended.
• Accurate disease progression modeling is crucial for reliable clinical decision analysis.