In a clinical trial I have 1000 patients with disease A. Their mean treatment cost is 1000$ per patient with a 95% confidence interval of [900;1100] for example. The whole population of country X is 10 Mio. Prevalence of disease A in country X is 5%, with a 95% confidence interval of [4.5%,5.5%].

How do I calculate the total costs including the 95% interval for country X?

I'd also like to see a Bayesian approach if appropriate.


I don't think you can do this from the data you have. The cost per treatment in a clinical trial will be very different from the cost in a general population because they involve very different costs (e.g. different amounts of doctor time, different travel, different supervision, different evaluation and often different demographics on the people).

I'd look for other people's experience in scaling things up in this way; but such data may be hard to get as it may be proprietary.

| cite | improve this answer | |
  • $\begingroup$ thanks Peter for your remark, but let us assume the cost value is a representative value and is valid for scaling up to the total population. My main concern is how I handle these two confidence intervals. $\endgroup$ – spore234 May 11 '15 at 11:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.