Dr.Lubna A.Al-Ansary writes:
> You might consider this a naive question but I could
> not find an answer in any of the EBM resource that I
> have & therefore I would greatly appreciate your help.
> In prospective trials that run for many years, would
> it be more accurate/useful to calculte the number
> needed to treat (NNT) in reference to the absolute
> risk reduction (ARR) in general or in reference to the
> ARR in patient-years?
It's not at all naive. The answer is quite subtle, and it depends a lot on
the context of the problem.
If the event can occur more than once per patient, and if you are
comfortable with certain assumptions like independence in non-overlapping
time intervals and no trends or seasonal changes in the probability of an
event, then calculate NNT in terms of patient years. An intervention might
prevent one extra nosocomial infection for every seventeen patient days.
If there is a single event, then it may make sense to look at the NNT at a
specific time point. For example, we need to treat 12 patients to prevent
one relapse within two years of surgery.
A key question is whether you think that following a hundred patients for
two years each will give you numbers comparable to following twenty patients
for ten years each. In other words, does 100 pts * 2 yrs = 20 pts * 10 yrs?
Steve Simon, [log in to unmask], Standard Disclaimer.
The STATS web page has moved to
http://www.childrens-mercy.org/stats.
|