The European Journal of Public Health 2008 18(5):438-439; doi:10.1093/eurpub/ckn087
© The Author 2008. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.
Comments on HIA forecast: cloudy with sunny spells
John Kemm
West Midlands Public Health Observatory, Birmingham, England, UK
Correspondence: e-mail: John.kemm{at}wmpho.org.uk
In her forecast, Thomson makes three criticisms of current Health Impact Assessment (HIA) practice. That it is unreasonable to expect decision makers to be influenced by HIA; that HIA fails to acknowledge the uncertainty attached to its predictions; and that the predictions in HIA are inadequately based on evidence. Each of these criticisms deserves consideration.
 |
Influence
|
|---|
The purpose of HIA is to assist decision makers and Thomson
is right to focus attention on the relation between decision
maker and health impact assessor. It is unreasonable to expect
that health considerations will always outweigh non-health considerations,
since the aim of HIA is to ensure that health considerations
are not overlooked rather than to ensure that they always take
precedence over all others. Where HIA is part of certain statutory
processes (which in UK include planning applications and operating
license applications for certain industrial processes) one can
be confident that the HIA will be taken into account. In other
situations increasing the chances that it will be considered
by the decision makers is an essential element in design and
planning of the HIA. The proponents of the proposal being considered
in the HIA must be aware of and given opportunity to contribute
to the HIA. Preferably the HIA process should start early before
the proposals are finalized so that they can readily be modified.
It is still easy to find examples of HIA, which have been done
in such a way that they had no prospect of influencing decision
makers, but the problem has received much more thought than
Thomson suggests.
|
Admitting uncertainty
|
|---|
Thomson is right to highlight the uncertainty, which attaches
to predictions in HIA or any other field, and it is indeed a
‘misguided conceit’ to pretend otherwise. However,
any competent HIA report will make clear that it offers a judgement
(with luck a best judgement) of likely consequences rather than
a statement of certainty. However, it is not reasonable to suggest
that uncertainty should prevent HIA from making any predictions.
Certainty or even very high probability are luxuries, which
HIA practitioners along with others concerned with public health
do not enjoy. It is little help to a decision maker to suggest
that they should wait 5 years while academics research the question
(and all too probably come back with the conclusion that more
research is needed). Usually the decision has to be made in
the near future and unless an HIA is prepared to offer its best
judgement, which may well be wrong, it is no use. Statistics
typically aim to reduce the chance of error to 1 in 20 (
P <
0.05). Most decision makers have to be content with a much higher
error rate.
|
Evidence for predictions
|
|---|
As Thomson suggests the starting point for prediction should
be a causal pathway diagram so that the assessor can attempt
to investigate the importance of each causal link using such
evidence as is available. Thomson criticises predictions based
on the assumption that community severance produces negative
impacts. While conceding that the effect of social networks
on health are complex and poorly understood it is surely unreasonable
to go on to suggest that no prediction can be made of how community
severance will effect health. Similar arguments could be applied
to transport policies, which rely on use of private vehicles
or policies which tolerate bad housing.
One can readily agree that evidence derived from research studies and synthesized in systematic reviews are needed for HIA and that HIA will get better as more of these become available. It is unrealistic, however, to think that research could ever provide a library of health consequences of changes or a ‘list of interventions that work’, since there will always be questions as to whether they are relevant to a particular context.
|
Conclusion
|
|---|
Thomson's conclusion that ‘it is difficult to see how
HIA can or should have any influence on decision making’
is far too sweeping. Unfortunately, her article was based on
a totally misdirected search strategy, which looked for HIA
reports in academic journals. A few minutes on Google would
have found far more than the 50 or so HIA reports found by Thomson.
A visit to the HIA gateway (
www.HIAgateway.org.uk) a year ago
would have produced more than 100 and today would produce nearly
200 HIA reports. A good HIA report is unlikely to be accepted
by a peer reviewed journal and an accepted paper would almost
certainly have to omit half the material that a good HIA report
should contain. None the less, the HIA community should pay
a great deal of attention to Thomson's critique. A more balanced
summary of the state of HIA would have been ‘has made
worthwhile progress – must try harder’.
CiteULike 
Connotea 
Del.icio.us What's this?
Top
Influence
Admitting uncertainty