Thinking of community based interventions always raises the question of
statistical power for the findings (or no findings) of the study. These
sorts of interventions are naturally situated within community
'clusters' (hospital, neighbourhood, town etc). The best available
evidence on the analysis of clustered RCTs - that I am aware of -
recommends that randomisation and analysis both should be conducted at
the cluster level. However public health studies are costly and
difficult to run and are usually done in a limited number of these
clusters.
I practically faced with this question when I was preparing a
presentation on 'Sure Start' programme which is a recent public health
policy in the UK to tackle health inequality. While the interventions
seemed (in essence) to be evidence based, the researchers couldn't be
sure that they can provide statistical power which is necessary to
defend the policy in EBHC terms. For example how to compare the outcomes
such as the incidence of Low Birth Weight babies in the intervention and
control communities.
My question is there any RCT planning technique (perhaps a complex one!)
to address this issue?
Regards
Arash
--
Arash Rashidian, MD
Health Services Research Scholar
Department of Health Sciences
Alcuin A Block, University of York
York, YO10 5DD, UK
Tel: +44 (0)1904 434498
Mobile: +44 (0)7786323559
Fax: +44 (0)1904 434517
http://www-users.york.ac.uk/~ar130/
Studying Adherence to Guidelines and Evidence (SAGE)
"Sontheimer, Daniel MD" wrote:
>
> The big questions I see are:
>
> Can RCTs handle complex intervention and complex evaluations?
> Are they the best-suited for this?
>
> Remember the Ca++ blocker trial (NORDIC??? or something, i pitched the
> article) of nitrendipine or diltiazem, I can't remember, but they used a
> method
> that was supposed to simulate 'real life' practice. My first read on the
> methods was that I thought it was a sham.
> Second read is that it still is for a single-agent trial. As complexity
> grows there may be more to it.
> Dan
|