Psychometric properties of the Malay version Women’s Views of Birth Postnatal Satisfaction Questionnaire using the Rasch measurement model: a cross-sectional study

Abstract

Abstract Background Maternal satisfaction must be assessed in order to reflect the quality of care, which is considered an outcome of healthcare services. It can also be used to contrast and compare satisfaction with various care models or service configuration or to assess overtime changes. The purpose of this study was to investigate the psychometric properties of the Malay version Women’s Views of Birth Postnatal Satisfaction Questionnaire (WOMBPNSQ) based on the Rasch scale model of labour satisfaction. Methods This is a cross-sectional study. Postpartum women were identified from a tertiary hospital and evaluated at 1-month postnatal period using WOMBLSQ. The Rasch model was used to investigate the reliability, unidimensionality, item and person misfits and distribution map. Results A total of 195 women were involved. The Rasch analysis revealed that the 30 items had a high level of reliability at 0.99 and item separation at 9.02. It has a low level of reliability at 0.45 and persons separation at 0.90. All the items are considered fit. Five people have most misfitting response strings based on item IPS_Q15, ‘I was given little advice on contraception following the birth of my baby’, but extremely trivial differences were found in the parameter estimates after refitting the model. The more difficult item to endorse satisfaction is item CA_Q17 ‘I was given little advice on contraception following the birth of my baby’. Conclusions The WOMBLSQ tested in postpartum women proved to have high item reliability index but with an adequate sample. The analysis shows that the 30 items target the right form of respondents, have similar latent characteristics of postpartum women and a shared sense of satisfaction. For future improvement, more difficult items endorsing satisfaction should be created, and the common items in which satisfaction is expected should be reduced.