I've been reading this discussion with interest as our team at the
Social Science Research Unit is currently conducting a systematic review
on health inequalities, and we're particularly interested in how to
measure the effectiveness of interventions. We've found relatively few
suggestions in the literature as to how to evaluate the impact of
interventions on health inequalities, as opposed to measuring
inequalities in a population in an observational sense (although of
course you have to do the latter to be able to do the former).
There seems to be a difficulty with carrying observational measures over
to outcome evaluation research. Given a delimited sample - say a school
or a workplace - standard measures of inequality could be applied to it
before and after the intervention to obtain a measure of the
intervention effect. However, this wouldn't tell you about the effect of
the intervention beyond that sample, on the inequalities which are
present in the wider society. This problem could be solved by using a
sample which is demographically representative of the population as a
whole - but then the intervention would have to have a very broad focus,
and the existing research seems to show that this kind of broad-based
intervention usually benefits high-status groups more than low-status
ones. That is, interventions which work directly with groups at risk of
disadvantage to improve their health outcomes seem most promising in
reducing health inequalities, but their effectiveness in achieving this
is hard to measure; on the other hand, interventions across a whole
population can be readily evaluated, but are generally less effective
(or even counter-productive) in terms of reducing inequalities.
We'd be grateful for any input on this problem - could it be that no
single measure of effectiveness is appropriate for both small-scale
interventions with disadvantaged groups and large-scale interventions
such as national policy initiatives? Is it sufficient to identify groups
at risk of disadvantage to target interventions, and then use standard
measures of effectiveness, on the assumption that, if those groups'
outcomes are improved, then inequalities have been reduced?
Many thanks,
Theo Lorenc
Social Science Research Unit
Institute of Education
18 Woburn Square
London WC1H 0NR
UK
[log in to unmask]
|