The majority of environmental processes and agents that affect human health are not singular in their effects, but cause or contribute to a range of different health outcomes.  Atmospheric particulates, for example, are implicated in a range of illnesses, including upper and lower respiratory tract infections, cardivascular diseases and cancers.  Exposure to noise may impair learning abilities, cause sleep disturbance and annoyance, and contribute to raised blood pressure and cardiovascular illness.  These various effects may also manifest themselves differently between different individuals (and in one individual over time), depending on the level of exposure, the susceptibility of the people concerned, and the course of the illness.  Health effects are therefore both varied and multivariate phenomena.  Combining them into a single, interpretable measure of impact is not always easy. 

Five sets of factors, in particular, have to be considered and allowed for in devising and applying measures of impact: 

1. Type of effect. The majority of environmental processes and agents that affect human health are not singular in their effects, but cause or contribute to a range of different health outcomes.  Air pollution, for example, is implicated in both upper and lower respiratory tract infections, cardivascular diseases and cancers.  Nor are the effects necessarily adverse; many policies and other interventions have important positive effects for health - e.g. by providing protection from disease or by enhancing people's physical condition, sense of well-being and quality of life. 
2. Scale of effect.  Any outcome, whether adverse or beneficial, may vary greatly in its degree.  Diseases, for example, may vary in their severity from mild discomfort (or concern) through to total incapacity and, ultimately, death.  Cases of any disease are therefore not necessarily equal, but need to be weighted in some way to reflect their differing severity.   Likewise, beneficial effects may vary in terms of their degree of benefit, from mild alleviation of symptoms through full protection against adverse effects to full enhancement of quality of life.  
3. Duration of effect.  Likewise, diseases may persist for different lengths of time (either as a single period or repeated events).  The overall impact of the disease thus has to be considered as the sum of its effects over this entire duration. 
4. Timing of effect.  Health effects may arise at different life-stages, from pre-birth (in utero) to old age, and in some cases, may even be delayed to future generations.  Independently of their duration, this may affect people's attitudes to the effects: there is a tendency, for example, to attach higher values to immediate events compared to those that are deferred until later (discounting).
5. Target of effect.  In addition to the above, the way in which any effect is valued depends on the underlying social priorities of those concerned.  These may not be based on simple utilitarian principles, but may give preference to some groups over others (e.g. in the endeavour to maintain or reduce social inequalities).  Thus perceived differences in impact may exist depending on who is affected.

Impact analysis thus involves more than simply summing the incidence of disease across different outcomes.  It requires the incorporation of, and trades-off with, beneficial effects; and it requires all effects to be weighted according to their scale, duration and timing and the social values that may be attached to different groups of individuals.   All impact analyses are thus contextual in that they are rooted within a specific (though often implicit) value system.  If the analysis is to be acceptable and meaningful to the users, the value system needs to be made explicit, and should reflect the needs of the users concerned.