Area selection for post-hoc impact evaluation of the delegated management model of urban water service delivery in Kisumu, Kenya

Okotto, L.G., Okotto-Okotto, Joseph, Damkjaer, Simon and Wright, Jim (2021) Area selection for post-hoc impact evaluation of the delegated management model of urban water service delivery in Kisumu, Kenya. In Amsterdam International Water Week: Urban Water Cycles, Water Service Delivery Systems and Removal of Pollutants.

Abstract

Background: delegated management, also known as a water operator partnership, is a model under which a utility delegates operational responsibility for urban water services to a small-scale provider. The model is currently operational in parts of Maputo(Mozambique), Arusha ( Tanzania), and Kisumu in Kenya. To date, studies of the impact of delegated management model have focused on expansion of services, pipeline breakages and supply interruptions, and revenue recovery, but have not considered its impacts on water quality and thereby ultimately health.

Aim: this study therefore aims to evaluate the impact of delegated management on microbiological water quality at kiosks through to point-of-consumption in the home in off-grid neighbourhoods of Kisumu.

Methods: this post hoc impact evaluation study will compare E. coli contamination of household stored water in a matched sample of Enumeration Areas (EA) under delegated management versus a control group. Characteristics of urban EAs under delegated management (including metered water connections per household; kiosks per capita; population density; proximity to sewerage lines; and probability of built-up land cover measured via Sentinel-2 satellite imagery) were collated within a GIS and compared with those not under delegated management using logistic regression. To minimise differences between EAs under delegated management and control EAs, a balanced sample of EAs will be selected using coarsened exact matching based on these characteristics. Within selected EAs, E. coli will be measured in water at kiosks, in vendors’ hand-carts, and in water stored by consumers, targeting 90 households in intervention and control groups. Final analysis will examine the effect of delegated management on E. coli in household stored water using logistic or multinomial regression, controlling for risk factors related to the vendor (e.g. vessel cleaning and handwashing), the household (e.g. water storage and handling practices), and the piped supply (e.g. adequate residual chlorine).

Preliminary results: of 342 urban EAs in Kisumu County, the 16 under delegated management had significantly less built-up land cover (mean probability of built-up land cover 19.7% compared to 27.6% in comparison EAs, p=0.008). Examination of other EA characteristics indicated non-significant differences in mean population density and imbalance in their joint distribution, indicating the need to select comparable EAs for the control group. Final EA selection and fieldwork is ongoing.

Conclusions: these preliminary findings suggest there are systematic differences between EAs under delegated management and those directly managed by the utility. Such differences need to be controlled for in developing evidence concerning the impact of delegated management on the quality of water services in Kisumu.

This record has no associated files available for download.

Contributors

Author: Jim Wright

Download statistics