Table 1.

Steps in designing a delivery system evaluation

	Examples	Comments
Determine the purpose of the evaluation	- To evaluate a new delivery system for ITNs - To evaluate a new delivery system for IPTc - To assess the delivery of IPTp through ANC	Evaluation of a new delivery system for an existing intervention requires a pre-post survey with attribution of nets by source. A process analysis is required to assess the outputs at each intermediate step in the causal chain of delivery
Select the evaluation method	- cross sectional pre-post survey with attribution of outcomes by source of intervention - cross sectional post intervention survey with attribution of outcomes by source of intervention - cross sectional post only survey with no control	For evaluation of a new delivery system for an existing intervention a pre-post survey with attribution of outcomes by source would provide causality for proximal indicators and plausibility inference for distal indicators.
Define the outcome indicators	- the proportion of children under 5years using an ITN - the proportion of children under 5 years taking a full course of IPT - the proportion of pregnant women who attend ANC receiving at least 2 doses of IPT	The primary outcome indicator may be a distal or proximal indicator
Define the pathway of delivery	- include several proximal and more than one distal steps - includes several proximal and one distal steps - the evaluation terminates at proximal steps	The number of steps varies with interventions and with delivery systems.Many pathways are linearNot all pathways are linear
Characterise the contextual factors	- malaria transmission levels - structure of the health system - socio-demographics of the population	Disaggregate outcomes by contextual factorsDescribe contextual factors