Would a realist evaluation (Pawson and Tilley 1997) be beneficial? This would attempt to establish what works (nor not),why, for whom and in what context/circumstances.
-----Original Message-----
From: Evidence based health (EBH) [mailto:[log in to unmask]] On Behalf Of Anne Gray
Sent: 24 August 2012 12:01
To: [log in to unmask]
Subject: Evidence based healthcase delivery
How useful are randomised controlled trials in evaluating new ways of delivering care?
This question is asked in an editorial in HSJ today (http://www.hsj.co.uk/5048586.article) relating to the Whole system demonstrator project which the government is using to justify spending huge amounts of money on telehealth and telecare.
The Author raises really useful comments about variables in service delivery - such as how long it takes to recruit people to new services. The length of time to recruit does not affect the outcome of the RCT but definitely does if you are trying to work out in which year the service will give you your QIPP savings.
It would be good for those of us supporting comisisoners to get a better grip on the evidence needed about health services so that we can support the spread of good services. RCTs work when the variables can be controlled ie when we can control WHAT we are doing. Services are about HOW that is delivered. Having worked with comissioners I am well aware that we do not understand (nor does anyone else I suspect) the variables needed to predict outcomes of transferring services from one location to another ie the spread of good practice.
Could we develop a critical appraisal tool for service delivery? Some of the case studies and QIPP study templates have made a start.
|