Starting with water and sanitation interventions and then trying to gauge the health impact can actually take us away from our desired goal of securing health improvements. Reversing this approach to start with health impact (first) and then determine causality (second) may create a more effective framework to optimize the trade-offs between water, sanitation and a range of other interventions!
The water and sanitation sector has been subject to numerous health impact studies. These are complex undertakings that require careful intervention and control conditions, extensive and carefully managed data sets, considerable time and money. Even in the best cases, quantifying the health impact of water and sanitation interventions is plagued by the high levels of uncertainty that surround the confounding variables. Furthermore, such studies do not quantify the relative health impact of choosing to invest in water and sanitation rather than breast feeding, or female literacy, or any other intervention. Even worse, such studies can draw a positive correlation between an intervention and the health impact … while the overall health for the particular target population has decreased. In such a scenario, it could be legitimately argued that investing in water and sanitation (and not female literacy) was the wrong choice - if the goal was a positive impact on health.
While quantifying the health impacts of a particular water supply and sanitation intervention is problematic, it is not because measuring health impact is difficult. The primary difficulty of health impact studies lies in starting with a cause, correcting for confounding factors, to then attribute the effect to that predetermined cause. Is it therefore worth considering if we might be better served if we reverse our approach to evaluate health impacts first and then determine causes second?
While deaths are undesirable they are relatively easy to measure. Similarly, new births are relatively simple to measure (all-be-it a given that registration at birth needs improvement). Subtracting deaths from dates of birth enables an average ‘life expectancy’ to be calculated for a particular jurisdiction. This is NOT just a measure of the performance of health professionals, but it is actually a measure of the development effectiveness in that particular jurisdiction. Ultimately law & order, nutrition & employment, water & sanitation, education & transport, breast feeding and the use of ORS will all impact the life expectancy of a particular jurisdiction.
If this data is collected by the lowest tier of local government (responsible for the registration of births and deaths) it should be possible calculate development effectiveness against the average 'life expectancy' trends of a block (upazila) or district. This will create some incentives for local administrators to target those development areas which have the biggest impact on life expectancy. Whether the major causes of fatality are road safety or diabetes, ARI or nutrition, female illiteracy or open defecation, deaths from insurgencies or suicides from crop failure … effective administration of development should reduce premature mortality thus increasing average life expectancy. Developing a simple system for tracking mortality in order to determine ‘life expectancy’ could be a pre-cursor for a more comprehensive system for tracking morbidity to determine ‘quality of life’.
Recognizing those jurisdictions that demonstrate significant improvements in life expectancy can create some incentives for local administrators to mobilize their departments in the most effective manner. Determining ‘what intervention works best in what location’ is then addressed at the local level by ‘schools of practice’ rather than by centralized ‘schools of research’. Central research offices can then focus on learning from the practical experiences of ‘what’ is working ‘where’ and ‘why’ and building this into choice alternatives.
The greatest challenge in such a proposal is not the measuring of births and deaths, as these are easy to measure and verify. Neither is it the collation of data, as a simple Management Information System (MIS) could enable life expectancy to be collated from the lowest local government jurisdictions. Neither is it the risks of internal migration, which could be countered by linking registration of births and deaths to national identify card systems. Neither is it ‘white noise’ because all causes of morbidity (even the unpredictable causes) are a relevant measure of district preparedness and response. The greatest challenge is the introduction of social accountability systems that enable civil society to adjudicate on the authenticity of birth and death registration data.
Authentic data on the registration of births and deaths could ultimately enable citizens to hold local administrators accountable for prioritizing those interventions that have the greatest positive impact on their life expectancy.