This is an idealised representation of an HIA's influence. Other assessment processes, organisational considerations and even broader social conditions will play much larger roles. In fact the process depicted is never linear either. Health outcomes and determinants are constantly changing, and decisions are constantly revisited. The delay between an activity and eventual health outcomes can sometimes stretch to decades. At each step there are a multitude of other factors that exert influence, apart from the HIA.
Though HIA's goal is to protect health, like EIA's is to protect the environment, the practical purpose of an HIA is to change decisions and implementation - the first step in the process depicted above. We should think about the purpose of EIA the same way. It's not an environmental intervention, it's a decision-making intervention. Burdening it with expectations of environmental protection isn't realistic.
If we do acknowledge that it will be difficult if not impossible for an HIA to demonstrate its role in changing health outcomes, we should redouble our efforts to prove its effectiveness in influencing decisions and implementation. I made a related argument in a recent article:
In some ways the issue of effectiveness may have less currency in relation to other forms of IA [than HIA]. Impact assessment, in particular environmental impact assessment, is used in some form in almost every country. Its use is common, accepted, well understood and not usually actively compared to other interventions or activities. This is not necessarily true for health impact assessment though because of the resource constraints and associated health disciplinary and epistemological concerns... if HIA's use is to continue to be supported in increasingly resource-constrained health systems that demand evidence of the comparative effectiveness of interventions.
The right question is not whether HIA changes health outcomes. Instead it's does HIA change decisions, implementation and ways of working?