Risk Factors for Postpartum Depressive Symptoms among Oregon Women

Introduction:

Postpartum depression is a major depressive episode that is temporally associated with childbirth. Generally, it has onset within four weeks of delivery and can last up to one year. Postpartum depression has been recognized as a very important maternal and child health concern over the last few years because it affects about 15% of new mothers and can have adverse outcomes for the entire family.

The causes of postpartum depression are not clear. Some studies suggest the onset of depression may be related to the rapid decline in hormonal levels after delivery. Other studies have investigated the link between social and environmental risk factors associated with postpartum depression. Risk factors identified in prior studies includes prenatal depression, childcare stress, prenatal anxiety, life stress, social support, marital relationship, having a history of previous depression, infant temperament, and maternity blues, low-self-esteem, marital status, socioeconomic status, and unplanned/unwanted pregnancy.

The objectives of this study were three-fold. The first was to determine the prevalence of postpartum depressive symptoms among Oregon women. The second objective was to determine which of the risk factors selected based on prior literature were significant in this sample of Oregon women. The last objective was to evaluate the relationship of individual race/ethnic groups and postpartum depressive symptoms.

Methods:

This was a an exploratory cross-sectional study utilizing data from the Oregon Pregnancy Risk Assessment Monitoring System (PRAMS) survey for live births from the year 2004. Oregon PRAMS is surveillance project run by the Centers for Disease Control (CDC) and collects state-specific, population-based data on maternal experiences, attitudes and feelings before, during and shortly after pregnancy. Respondents are selected as a stratified random sample with oversampling of women from racial/ethnic minorities. In 2004, Oregon added four new questions to the PRAMS survey, which asked about depressive symptoms. This new data gives us an opportunity to learn more about the prevalence of postpartum depressive symptoms and associated risk factors specific to Oregon. Determining possible variation or similarities between state risk factors for postpartum depressive symptoms is important for developing public health interventions or programs specific to each state. Early recognition and screening for postpartum depressive symptoms is an important public health issue because it can improve health outcomes for mothers, children and families in Oregon. For 2004, a total of 2814 surveys were sent and 1968 completed surveys were returned; the unweighted response rate was 69.9%. The weighted response weight was 74.4%.

The independent variables initially selected were previously identified significant risk factors with some additional variables of interest and included the following: race/ethnicity, age and education of mother, federal poverty level, and marital status, pregnancy intention, WIC (Women, Infant and Children nutrition program) enrollment status, physical activity in past month, Medicaid recipient, prenatal dental care, smoking before and during pregnancy, alcohol use before and during pregnancy, previous live births, previous terminations, breastfeeding status at 8 weeks, exclusive breast feeding at 8 weeks postpartum, living urban or rural, physical abuse by husband or partner before and during pregnancy, prenatal care initiation and duration, mom's body mass index (BMI), and prenatal stress.

The dependent variable combined two questions from the survey to represent whether a woman has postpartum depressive symptoms or not. As a point of clarification, postpartum depression is a clinical diagnosis that must be made by a physician using a validated screening and diagnostic tool. Oregon PRAMS includes two questions asking women about depressive symptoms. The first question was “Since your new baby was born, how often have you felt down, depressed, or hopeless?” and the second question asked, “Since your new baby was born, how often have you had little interest or little pleasure in doing things?” Despite the fact that these two questions have not yet been validated, one study did find that the following two questions “during the past month, have you often been bothered by feeling down, depressed, or hopeless?” and “during the past month, have you often been bother by little interest or pleasure in doing things?” did have similar sensitivity and specificity to using validated tools for diagnosing major depression (1). Postpartum depression is considered a major depression (as defined below), therefore this finding of similarity between validated tools and these two questions suggests that women experiencing postpartum depressive symptoms are likely also experiencing postpartum depression. However, for accuracy, the two will be distinguished appropriately.

The statistical software STATA version 8.2 was used for data analysis. Data analysis included descriptive statistics, univariate and multivariate logistic regression. Simple logistic regression was done on all the independent variables selected and those that were significant (based on confidence intervals) were included in the full multivariate logistic model. A backward progression approach was used for multivariate regression analysis. Specifically, insignificant variables were sequentially removed from the model one at a time based on the highest p-value above 0.5 and confidence intervals. Potential confounding was evaluated when each variable was removed by identifying any change of more than 10% in the remaining variables. If a removed variable was a confounder, it was left in the final multivariate model. The significant variables in the final model had a confidence interval that did not include 1 and a p-value <0.5. Individual race and ethnic subgroups were further evaluated in the multivariate model by creating a set of dummy variables for each race and ethnic category. Each race and ethnic category was labeled "1" and all others were labeled "0" as the reference. Each dummy variable was evaluated in the model independently, adjusted for other significant risk factors. Results:

The prevalence of postpartum depressive symptoms in this sample of women was 13%.

In multivariate logistic regression, the following risk factors remained statistically significant:

Non-White race and ethnicity (OR 1.17: 95% CI 1.03, 1.33)

Mother's BMI>=25 (OR 1.76: 95% CI 1.12, 2.75)

Partner-related stress (OR 3.45: 95% CI 2.18, 5.47)

Income: Federal poverty level <300% (OR 1.28: 95% CI 1.11, 1.48)

Insufficient dental care (OR 1.94: 95% CI 1.07, 3.49).

Subsequently, each race and ethnic group was added into the multivariate model one at a time. Asian and Pacific Islander women were the only group to have a statistically significant elevated odds ratio, when adjusting for federal poverty level, BMI, partner-related stress and insufficient dental care.

Hispanic (OR 1.25: 95% CI 0.75, 2.09)

Black (OR 1.38: 95% CI 0.79, 2.41)

American Indian (OR 1.32: 95% CI 0.77, 2.25)

Asian/Pacific Islander (OR 1.98: 95% CI 1.17, 3.35)

Conclusion:

This study analyzed the prevalence and predictor variables for postpartum depressive symptoms among a sample of women in Oregon. Based on the results, the prevalence of postpartum depressive symptoms in Oregon reflects prior research at 13%. Similarly, the risk factors found to be significant were similar to what has been found in prior literature with the exception of insufficient dental care, which has not been analyzed previously.

Two findings were unexpected in multivariate analysis. The first was that of all non-White race and ethnic groups, Asian and Pacific Islander women had the highest odds ratio for postpartum depressive symptoms, when adjusting for federal poverty level, mother's BMI, insufficient dental care and partner-related stress. The second finding was that Asian and Pacific Islander women were the only non-White group to have a statistically significant association with postpartum depressive symptoms. These results suggest that other factors pertaining to Asian and Pacific Islander women increase the risk for postpartum depressive symptoms. These factors may likely be cultural, but need to be further elucidated.

There is limited research estimating risk ratios for individual race and ethnic groups and postpartum depressive symptoms. Much of the literature focuses on non-Hispanic Black and Hispanic women. Despite important literature contributions on subgroups of Asian and Pacific Islander women, no studies were found evaluating risk ratios. As a result of limited epidemiological research, the status of Asian and Pacific Islander women is largely unknown and information about postpartum depressive symptoms is even less available. This study is important because it has expanded upon limited data available regarding the health of Asian and Pacific Islander women. To better validate the results of this study, the same analysis will be done for 2005 data. Further epidemiological research that includes data on culturally relevant risk factors on Asian and Pacific Islander women as a whole and the many subgroups in this broad categorization would be helpful in providing additional insight into targeted screening and intervention for postpartum depressive symptoms.

1. Wooley, MA, Avins, AL, Miranda, J et al. Case-Finding Instruments for Depression, Two Questions Are as Good as Many. J Gen Intern Med 1997;12: 439-445