Statistical analysis
Based on the study by Thompson who found depression in 10% of the unexposed and 16% of the exposed women at 4 months postpartum,16 and including 2 women in the group without PPH for every woman with it, we required 1542 women, including 514 exposed and 1028 unexposed, to have a power of 90%, with a bilateral alpha equal to 5%.
We conducted a descriptive analysis of both groups — exposed and unexposed — for the social, demographic, medical, and perinatal data. Categorical variables were compared by chi-square (χ2) tests (or Fisher’s exact test, as appropriate) and continuous variables by Student’s t-test (or a Mann-Whitney test).
Prevalence rates and their 95%CIs for depression, anxiety, and PTSD were calculated at 2, 6, and 12 months by χ2 tests (or Fisher’s exact test, as appropriate).
The scores of each questionnaire were transformed in log form to obtain a normal distribution. The results are presented as geometric means. The geometric means were compared between the groups at each study time point with Student’s t-test (or a Mann-Whitney test). A multivariate analysis by a manual backward stepwise generalized linear model was performed to study the association between exposure and the scores obtained on each questionnaire. The adjusted geometric means were compared between the groups. The results are also presented as crude and adjusted odds ratios (ORs) with their 95%CIs.
Statistical significance was set at p<0.05. All statistical analyses were performed with SAS software version 9, 2002-2010 (SAS Institute Inc.).