Evaluation of the psychometric properties of the London Measure of Unplanned Pregnancy in Brazilian Portuguese

Ana Luiza Vilela Borges, Geraldine Barrett, Osmara Alves dos Santos, Natalia de Castro Nascimento, Fernanda Bigio Cavalhieri, Elizabeth Fujimori

Estimates of unplanned pregnancy worldwide are of concern, especially in low and middle-income countries, including Brazil. Although the contraceptive prevalence rate is high in Brazil, almost half of all pregnancies are reported as unintended. The only source of nationally representative data about pregnancy intention is the Demographic and Health Survey, as with many other countries. In more recent years, however, it has been realized that concept of unintended pregnancy is potentially more complex and requires more sophisticated measurement strategies, such as the London Measure of Unplanned Pregnancy (LMUP). The LMUP has been translated and validated in other languages, but not Portuguese yet. In this study, we evaluate the psychometric properties of the LMUP in the Portuguese language, Brazilian version. A Brazilian Portuguese version of the LMUP was produced via translation and back-translation. After piloting, the mode of administration was changed from self-completion to interviewer-administration. The measure was field tested with pregnant, postpartum, and postabortion women recruited at maternity and primary health care services in Sao Paulo city. Reliability (internal consistency) was assessed using Cronbach's alpha and item-total correlations. Construct validity was assessed using principal components analysis and hypothesis testing. Scaling was assessed with Mokken analysis. 759 women aged 15-44 completed the Brazilian Portuguese LMUP. There were no missing data. The measure was acceptable and well targeted. Reliability testing demonstrated good internal consistency (alpha = 0.81, all item-rest correlations >0.2). Validity testing confirmed that the measure was unidimensional and that all hypotheses were met: there were lower LMUP median scores among women in the extreme age groups (p < 0.001), among non-married women (p < 0.001) and those with lower educational attainment (p < 0.001). The Loevinger H coefficient was 0.60, indicating a strong scale. The Brazilian Portuguese LMUP is a valid and reliable measure of pregnancy planning/intention that is now available for use in Brazil. It represents a useful addition to the public health research and surveillance toolkit in Brazil.