Absence from Work in Pregnancy Related to Racial Factors: A Bayesian Analysis in the State of Bahia—Brazil

Abstract

Concerning the growing participation of women in the labor market, there are few studies on the influence of work on pregnancy and the absence from work during pregnancy, which makes difficult to develop public health policies for pregnant workers. This work evaluated the self-reported color/race and the absence from work in 502 puerperal women aged 19 years or older and non-indigenous, at Manoel Novaes Hospital, in Itabuna, Bahia—Brazil, through the application of a form. A Bayesian network was created using Bayesian Search (BS) learning algorithm. In the sample, 6 puerperal women declared themselves yellow (1.20%) and of these 2 (33.33%) absented from work during pregnancy, 49 declared themselves white (9.76%) and of these 24 (48.98%) absented from work, 322 declared themselves brown (64.14%) and of these 130 (40.38%) absented from work and 125 declared themselves black (24.90%) and of these 54 (43.20%) absented from work. In the Bayesian analysis, the largest inferences related to absenteeism was the black-race with 46%, contrasting with data from the frequentist model, where the white-race has the highest rate of absenteeism, similarly the lowest rate was in the mixed-race (mulatto) with 39% and the frequentist method was the yellow one with the lowest rate. The collected data increase the knowledge of the main causes of absence in pregnancy and the importance of the topic and the lack of studies justifies that more studies be carried out. It was possible to create a Bayesian network from data collected from puerperal women using the BS learning algorithm and infer about absenteeism considering the self-declared color/race as the input nodes. The Bayesian analysis of the absence from work during pregnancy is an important tool for the study of the topic and makes it possible to develop software for decision making by health professionals during pregnancy.