Thank you for your reply to make clarity on use of NNT in Meta-analysis.
Amit Raval
M.Pharm, Research Scholar,
Dept. of Pharmacy Practice,
NIPER SAS Nagar,
Panjab
India-160062
On 5/2/09, Rod Jackson <[log in to unmask]> wrote:
> Hi Amit an NNT is a patient-specific estimate of benefit (or harm) which
> is derived from combining the relative effect of an intervention with the
> patientıs pre-treatment risk. An NNT derived from a meta-analysis combines
> the average relative effect of an intervention (across the studies included)
> with the average pre-treatment risk of everyone in the trials. Therefore
> unless the patient or a group of patients, for whom you want to estimate an
> NNT, have the same pre-treatment risk as the average pretreatment risk of
> everyone in the meta-analysed studies, the NNT will be incorrect.
>
> regards
> --
> Rod Jackson
> Professor of Epidemiology
> Head of Epidemiology and Biostatistics
> School of Population Health, Tamaki Campus
> Faculty of Medical & Health Sciences, University of Auckland
> Private Bag 92019
> Auckland, New Zealand
>
> www.epiq.co.nz
>
> [log in to unmask]
> email usually checked within 1 week of receipt
>
>
>
>
> On 1/5/09 6:33 AM, "Amit Raval" <[log in to unmask]> wrote:
>
>> During study of EBM, I found the statement that use of NNT derived from
>> meta-analysis directly can be misleading.
>>
>> NNT is the one of the best way to express the actual outcome in more cost-
>> effective manner. And that's why it is used as measure of effectiveness of
>> treatment (Risk Reduction).
>>
>> So how can we rely on results based as NNT in meta-analysis ?
>>
>
>
>
|