An economic evaluation of a new diabetes drug uses a 5-year time horizon, even though diabetes is a lifelong condition. Why might this underestimate the drug's value, and what would be a more appropriate approach?
A5 years is always sufficient for chronic diseases
BA 5-year horizon misses long-term benefits (reduced cardiovascular events, preserved kidney function) and long-term cost offsets that accrue over decades. A lifetime Markov model with annual health state transitions would capture the full trajectory of costs and outcomes
CThe time horizon does not affect the ICER
DLonger time horizons always favor the new treatment
For chronic diseases, many benefits (and costs) accrue over decades. A 5-year evaluation of a diabetes drug captures early effects on blood sugar control but misses the downstream prevention of retinopathy, nephropathy, and cardiovascular events that develop over 10-30 years. A Markov model simulates annual transitions between health states (controlled diabetes → complications → death) over a lifetime horizon, projecting the full cost and health consequences of each treatment strategy. Most HTA bodies require a lifetime horizon for chronic diseases.
Question 2 Short Answer
An economic evaluation discounts future costs and health outcomes at 3% per year. This means a QALY gained 20 years from now is worth less than a QALY gained today. Why is this appropriate?
Think about your answer, then reveal below.
Model answer: Discounting reflects time preference — both individuals and society prefer benefits sooner rather than later and costs later rather than sooner. A dollar invested today grows over time, so a future dollar is worth less in present terms. The same logic is applied to health outcomes: given a choice, most people prefer to be healthy now rather than in 20 years. Without discounting, interventions with upfront costs and distant benefits (like childhood vaccination) would always look favorable, regardless of how far in the future the benefits occur. The 3% rate is a convention that balances time preference with the ethical concern that future people's health should not be excessively devalued.
There is ongoing debate about whether health outcomes should be discounted at the same rate as costs. Some argue that health does not depreciate like money and should be discounted at a lower rate (or not at all). This is more than an academic question: differential discounting dramatically favors preventive interventions (high upfront cost, benefits decades later). The convention of equal discounting at 3% is a pragmatic compromise used by most major HTA agencies.
Question 3 True / False
Probabilistic sensitivity analysis (PSA) is superior to one-way sensitivity analysis for decision-making because it simultaneously varies all uncertain parameters according to their probability distributions, reflecting the true uncertainty in the model.
TTrue
FFalse
Answer: True
One-way sensitivity analysis varies one parameter at a time while holding others constant — useful for identifying which parameters drive the result but unrealistic because multiple parameters are uncertain simultaneously. PSA draws thousands of random values from each parameter's distribution, runs the model for each draw, and produces a distribution of ICERs. This generates a cost-effectiveness acceptability curve showing the probability that each intervention is cost-effective at different willingness-to-pay thresholds. PSA captures the full uncertainty in the decision, including parameter correlations, and is required by most HTA agencies.