Using lot quality assurance sampling methodology to evaluate the u Thukela District Child Survival Project of KwaZulu-Natal

Health services programmes and interventions require objective measures on coverage and quality to ensure that managers have the evidence to adopt and expand health interventions effectively and effciently. Over the past 15 years, the uThukela District Child Survival Project of South Africa achieved this through regular two- to three-year project evaluations using cluster sampling. Results of these surveys have improved the ability of Child Survival Projects to identify priorities, defne objectives based on data, and measure progress towards these objectives. However, the inability to disaggregate information to composite Supervision Areas (SAs) was a limitation. Lot quality assurance sampling (LQAS) was adopted as an improved method based on its ability to assess performance in disaggregated units within a study population and to reliably compare coverage and health status indicators for routine monitoring of child survival activities. This study assessed the feasibility and reliability of LQAS to evaluate the Child Survival Project and compare the results of a rapid manual analysis with a secondary electronic analysis of the same data. The accuracy of the manual analysis was comparable and provides evidence that LQAS and rapid manual analysis can be used to evaluate intervention programmes, particularly in settings of limited electronic technology and expertise.