Developing and validating a knowledge and skills test for Tanzanian birth attendants trained in PartoMa safe childbirth care.

IntroductionImproving quality of maternity care in low- and middle-income countries is essential for reducing maternal and perinatal mortality and morbidity. Low dose-high frequency in-house training in routine and emergency maternity care is suggested to be central for this. To ensure the effectiveness and resource-efficiency of such training, a knowledge and skills test requiring minimal time and resources is needed. Therefore, we set out to develop and validate a test for routine use in busy, low-resource maternity units; to efficiently and effectively assess potential knowledge and skills gains over time when attending low-dose, high-frequency training in integrated care during childbirth. Materials and MethodsUsing Messicks standards, we developed a comprehensive yet time-efficient test covering childbirth surveillance, respectful care, and management of maternity and neonatal complications. Both expert and participant feedback informed the test design. The test was applied during a Tanzanian in-house training intervention (PartoMa) to assess its performance. Reliability was determined by Cronbachs Alpha, while validity was evaluated through face and content validity, factor analysis, and Rasch analysis. ResultsAfter multiple revisions, Cronbachs Alpha was below 0.7, indicating limited reliability. Experts agreed the test was well-designed for the intended content, with no concerns about clarity or relevance. Content validity was confirmed through expert judgment, reflecting the tests goals. In pilot testing, 160 (84.6%) participants rated it excellent in achieving the study objectives. Exploratory factor analysis showed the test did not measure a single latent trait, and Rasch analysis revealed discrepancies between observed and expected responses on item scores. ConclusionThe knowledge and skills test has shown promise for assessing healthcare providers in high-pressure, resource-limited settings, but its low Cronbachs Alpha and limitations in test results highlight the need for refinement.