Water Supply and Sanitation (WSS) Coverage

  • 13
These maps show the predicted proportion of households using an improved drinking-water source, an improved sanitation facility, and habitually defecating in the open sub-nationally in 2012. They can be used to identify areas where WSS coverage remains low, and so where drug-based interventions alone are unlikely to eliminate or control STH and schistosomiasis.

To generate these maps, information on household use of WSS was sourced from population-based national household surveys. Statistical models were fit to these data and used to estimate coverage sub-nationally (district-level or equivalent) across sub-Saharan Africa for 2012. For each map, the large figure shows predicted coverage for the overall population, whilst inset maps show predicted coverage separately for urban and rural populations.

 Defining an ‘improved’ source

Access to improved drinking water and sanitation were defined using the criteria outlined by the Joint Monitoring Programme for Water and Sanitation, based upon criteria set down for monitoring the Millennium Development Goal 7C, and are measured by reported use.

  • Improved drinking-water source (i.e. protected from outside faecal contamination): piped water, standpipes, tubewells, borewells, protected dug wells, protected springs and rainwater.
  • Improved sanitation facility (i.e. hygienically separates excreta from human contact): flush toilets, piped sewer systems, septic tanks, ventilated improved pit latrines, pit latrines with a slab and composting toilets. This metric includes households that share access to an improved facility
  • Open defecation (i.e.habitual defecation in the open): no access to an improved or unimproved sanitation facility

 Estimating access at sub-national levels

Spatially-explicit statistical models that took into account the spatial distribution of surveys, urban/rural designation, and also when the survey were conducted were developed and used to predict household access at high spatial resolution (district-level or equivalent) for 2012. These models consider the hierarchical structure of the whole dataset; this allows ‘borrowing’ of information from neighbouring districts to supplement available date, thus creating sufficient statistical power to generate reliable estimates for comparison of district estimates for evaluation purposes. The model also allows different time trends for urban and rural populations by country, but assumes an overall linear change over time. In practice, this means that countries (and districts within countries) are assumed to follow the regional (and national) average in cases where trend (or coverage) information is sparse or absent. Further details are provided in Pullan et al (2014).