Using Bayesian networks model predicting pregnancy after psychiatric interventions in infertile couple

Background Aim: Considering the psychosocial model of diseases, the aim of this study was to evaluate the effect of psychiatric intervention with regard to demographic and marriage characteristics on the pregnancy rate using Bayesian network model in infertile women. Methods Materials: In a randomized clinical trial, 638 infertile patients referred to an infertility clinic were evaluated. Among them, 140 couples with different levels of depression in at least one of the spouses were included in this substudy. These couples were divided randomly into two groups. After psychiatric intervention the clinical pregnancy rates of the two groups. The data were divided into two groups: demographic characteristics and marriage specifications, and by drawing Bayesian networks using Grow-Shrink (GS) algorithm, the conditional probability of pregnancy was estimated. Results: According to the results, Bayesian network model of the GS algorithm was significant (P = 0.548) and given that the fertility in the intervention group who were concurrently treated with antiretroviral treatment, the conditional probability was 38.5%, and this amount in the control group is 3.5% and group who were concurrently treated with induction of ovulation or did not receive any treatment the conditional probability was 72.2% and this amount in the control group is 23.1% comparing the values shows the importance of psychiatric intervention in increasing pregnancy rate. Conclusion: Results obtained from Bayesian network model are in line with results obtained from logistic model in terms of the significance of the variables with the difference that apart from the graphic structure, Bayesian network model also estimates conditional probabilities. This study shows that psychiatric and psychological treatments play an important role in curing infertility that will increase the chances of pregnancy.