Psychometric Properties of Breastfeeding Attrition Prediction Tool (BAPT): A Systematic Review

Background It is necessary to identify women who discontinue the breastfeeding to achieve the purpose of increasing the length of breastfeeding. To this end, Breastfeeding Attrition Prediction Tool (BAPT) has been developed. The current systematic review aimed to comprehensively review the validity and reliability of BAPT scale with different versions to give comprehensive information for authorities in this field. Materials and Methods Three English databases including Medline (via PubMed), Scopus, and Cochran library until May 2018 with no date restriction were searched. The search strategy was developed based on main terms of (Breastfeeding Attrition Prediction Tool OR BAPT) AND (Reliability OR validity OR Psychometrics OR Factor Analysis). Two reviewers separately extracted the required data available in full-text of all the quality of related studies was investigated using COSMIN checklist. Results All Cronbach alpha coefficient (both overall and subscale) were in excess of 0.7 except for subscale "NBS" in American version and subscale PBS in Persian version. In term of discriminant and predictive validity, BAPT revealed a good ability to classify women with or without breastfeeding. BAPT predicted breastfeeding status at postpartum period but it was not able to predict breastfeeding status in the third trimester of pregnancy. In term construct validity, four-factor solution of original English version was confirmed in Turkish, Persian and English other version. Conclusion Overall, the findings of systematic review supported that BAPT may be a valid (content, predictive and construct validity), and reliable (internal consistency and re-test reliability) instrument to use in both researches and clinics to identify mothers who are at risk of breastfeeding stop.