Patients with chronic diseases that include HIV infection are at increased risk of experiencing postpartum depression. In addition, social isolation has been associated with depression among women with HIV. Yet, it is unclear whether disclosure of HIV serostatus before the birth is associated with the risk of postpartum depression.
The purpose of this study was to determine whether maternal disclosure of her positive HIV serostatus before the delivery is associated with the risk of early postpartum depression.
In this retrospective cohort study, women who received obstetric care in a specialty perinatal HIV clinic (2007-2014) were stratified by whether, before the delivery, they had disclosed their HIV serostatus to (1) their sexual partner(s) or (2) at least 1 family member aside from sexual partner(s). Postpartum depression was identified initially by a positive result on a validated depression screening tool (Patient Health Questionnaire-9 or Edinburgh Postnatal Depression Scale) at the 6-week postpartum visit and then confirmed by evaluation with a mental health professional. Postpartum depression rates were compared by disclosure status. Multivariable logistic regression was performed to identify whether disclosure to either sexual partner(s) or family members remained associated independently with postpartum depression after we controlled for potential confounders that included antenatal mental health disorders.
Of the 215 women who received perinatal HIV care in this center and who had a documented disclosure status, 149 women (71.3%) had disclosed to their sexual partner(s), and 78 women (42.9%) had disclosed to at least 1 family member who was not a sexual partner. Although disclosure to sexual partner(s) was associated with a reduction in the proportion of women with postpartum depression (15.6% vs 25.5%), this difference did not reach statistical significance ( P = .126) and remained statistically insignificant after we controlled for potential confounders (adjusted odds ratio, 0.47; 95% confidence interval, 0.15–1.41). In contrast, disclosure to family member(s) was associated with a decreased prevalence of postpartum depression (11.4% vs 24.7%; P = .03), and this difference persisted in multivariable regression (adjusted odds ratio, 0.35; 95% confidence interval, 0.13–0.95).
In this cohort, maternal disclosure of HIV serostatus to family members (other than sexual partner[s]) was associated independently with a reduction in postpartum depression by more than one-half. Disclosure of HIV serostatus to a family member may be a marker for psychosocial well-being and enhanced support that affords protection against postpartum depression.
There are 15,000 women newly diagnosed with HIV annually in the United States, of whom 80% are of childbearing age. This leads to an estimated 5000–6000 HIV-seropositive women giving birth in the United States annually. Substantial progress has been made over the past 30 years in the prevention of mother-to-child transmission of HIV. However other complications of pregnancy that can affect women with HIV, such as postpartum depression, have received less attention.
Previous research has shown that rates of depression are higher in HIV-seropositive individuals (4-23%) compared with those without HIV infection (2-9%), with an associated odds ratio for depression associated with HIV seropositivity of 1.99 (95% confidence interval, 1.32–3.00). Similarly, perinatal depression is prevalent among HIV-seropositive pregnant women, with published rates that range from 31-44%. Depression in HIV-seropositive individuals has been associated with poor viral suppression and nonadherence to antiretroviral regimens, both with detrimental ramifications for maternal health and prevention of mother-to-child transmission. In addition, perinatal depression among all women remains a major cause of maternal morbidity and has significant implications on long-term child neurodevelopment. Accordingly, sound mental health during and after a pregnancy has been identified as a high-priority goal for clinical care and research.
Social isolation has been associated with an increased risk of depression in HIV-seropositive individuals. Nondisclosure of HIV serostatus can lead women to feel socially isolated, particularly around the time of childbirth. In contrast, continued stigma surrounding the diagnosis of HIV can lead to a woman feeling ostracized from her community after disclosure and thereby may increase social isolation. To our knowledge, the association between HIV serostatus disclosure and the risk of postpartum depression has not been examined previously. Therefore, the objective of this study was to determine whether disclosure of HIV serostatus to sexual partner(s) or to family members is associated with the risk of postpartum depression.
Materials and Methods
This is a retrospective cohort study of all HIV-positive pregnant women who received obstetric care in a specialty perinatal HIV clinic at Northwestern University from 2007-2014. Each patient underwent an evaluation by a licensed clinical social worker (LCSW) during prenatal care, wherein disclosure status to sexual partner(s) and/or family members was assessed. Additionally, the obstetric and infectious disease clinical teams routinely inquired about disclosure status as a part of clinical HIV care. If disclosure had not occurred previously, efforts to assist with disclosure were discussed. Social work and medical records were reviewed in detail to determine whether a patient had disclosed her HIV serostatus to either her partner(s) or family. Because disclosure to each represents a different aspect of social support, women were dichotomized by disclosure status to (1) their sexual partner(s) and (2) at least 1 family member (other than her partner) before the delivery. Women without a known disclosure status to their partner(s) and family were excluded.
Postpartum depression was identified initially by a positive result on a depression screening tool at the 6-week postpartum visit. The screens that were used included either the Edinburgh Postnatal Depression Scale or the Patient Health Questionnaire-9, which both are validated screens endorsed by American College of Obstetricians and Gynecologists for perinatal depression screening. For women who screened positive for perinatal depression, the diagnosis was confirmed by interview with an LCSW (2007-2009) or a clinical psychologist (2009-2014) who were members of the perinatal HIV clinic staff. All women with a positive depression screen underwent evaluation by either the LCSW or psychologist. In addition, as a part of routine clinical care during pregnancy, the LCSW or psychologist interviewed all patients to screen for previous mental health disorders or substance abuse history. Their notes were reviewed to identify whether either of these were present.
The medical records were abstracted for demographic and clinical characteristics that are associated with postpartum depression. These data included maternal age, self-reported race/ethnicity, immigrant status, marital status, current employment, insurance status, and duration of HIV diagnosis. These variables and the presence of any antenatal mental health or substance abuse history (excluding tobacco use only) were compared, with the use of bivariable analysis, between women who had disclosed to their partner and those who had not. In a similar manner, these potential confounders were compared between women who had disclosed to at least 1 member of their family (other than their partner) and those who had not. Chi squared and Fisher exact tests were used for categoric variables, as appropriate. The Student t test and Mann-Whitney U were used for continuous variables, as appropriate. Variables with a significant ( P < .05) association with disclosure in bivariable analysis were included in a multivariable logistic regression model for the outcome of postpartum depression. Odds ratios and 95% confidence intervals were estimated. An interaction term between disclosure to partner(s) and disclosure to family member(s) was entered into the regression to determine whether effect modification was present if disclosure to both occurred.
All tests were 2-tailed, and a probability value of <.05 was used to define statistical significance. Analyses were performed with Stata software (version 13; Stata Corporation, College Station, TX). The Institutional Review Board of Northwestern University approved this study with a waiver of informed consent.
A total of 215 women received prenatal care in the specialty perinatal HIV clinic during the study period. Of these, 209 women (97.2%) had a continued sexual partner with a documented partner disclosure status retrievable in their prenatal record and constituted the analyzable sample for the partner disclosure objective. Disclosure to the partner(s) before the delivery occurred in 149 women (71.3%). Characteristics of the cohort, stratified by partner disclosure status, are shown in Table 1 . Women who disclosed to their partner(s) were more likely to be married, employed, privately insured, and have obtained at least a college degree. Women who disclosed were less likely to have a history of illicit substance abuse. There were no differences in age, race/ethnicity, immigration status, parity, duration of HIV diagnosis, and a history of mental illness or tobacco use between women who disclosed to their partner(s) and those who did not disclose.
|Variable||Did not disclose to partner (n = 60)||Disclosed to partner (n = 149)||P value|
|Age, y a||28.7 ± 6.6||29.9 ± 6.1||.209|
|Race/ethnicity, n (%)||.180|
|Non-Hispanic white||3 (5.0)||19 (12.8)|
|Non-Hispanic black||46 (76.7)||103 (69.1)|
|Hispanic||11 (18.3)||22 (14.8)|
|Nulliparous, n (%)||20 (33.3)||58 (38.9)||.449|
|Married (n = 207), n (%)||2 (3.3)||68 (46.3)||<.001|
|Foreign-born (n = 206), n (%)||21 (35.0)||62 (42.5)||.351|
|Employed (n = 205), n (%)||19 (33.3)||74 (50.0)||.032|
|Educational level (n = 167), n (%)||.031|
|Less than high school||10 (21.7)||21 (17.4)|
|High school graduate/general education development||24 (41.6)||48 (39.7)|
|Some college||10 (21.7)||25 (20.7)|
|College graduate or beyond||2 (4.4)||27 (22.3)|
|Insurance status (at first prenatal visit), n (%)||<.001|
|Uninsured||9 (15.0)||18 (12.1)|
|Public||49 (81.7)||94 (63.1)|
|Private||2 (3.3)||37 (24.8)|
|Disclosed to family (n = 179), n (%)||19 (33.9)||57 (46.3)||.119|
|New HIV diagnosis during pregnancy, n (%)||21 (35.0)||42 (28.2)||.332|
|Years since HIV diagnosis b||2.5 (0–6)||4 (1–8)||.088|
|History of mental health diagnosis, n (%)||26 (43.3)||52 (34.9)||.254|
|Substance abuse, n (%)||19 (31.7)||21 (14.1)||.003|
|Tobacco use (n = 204), n (%)||18 (31.6)||30 (20.4)||.091|
Of the total sample, 182 women (84.7%) had a documented family disclosure status. These women constituted the analyzable sample for the family disclosure objective. Disclosure of HIV serostatus to at least 1 family member (aside from the sexual partner) before the delivery occurred in 78 of the cohort (42.9%). Characteristics of the cohort, stratified by family disclosure status, are shown in Table 2 . Compared with women who did not disclose, women who disclosed to their family were younger and had a longer duration HIV diagnosis. They were also less likely to be foreign-born or to have been diagnosed with HIV in the index pregnancy. There were no differences in race/ethnicity, marital status, employment status, insurance status, antenatal mental health disorders, or substance abuse history between women who did and did not disclose. Fifty-seven women (31.8%) disclosed to both their partner(s) and at least 1 additional family member. However, there was no significant association between partner disclosure and family disclosure ( P = .12). Only 37 women (17%) in the sample had not disclosed to either their partner(s) or family. Given the small sample size and therefore a lack of power to identify potentially clinically meaningful differences, this category was not analyzed separately.
|Variable||Did not disclose to family (n = 104)||Disclosed to family (n = 78)||P value|
|Age, y a||30.4 ± 5.9||27.8 ± 6.3||.004|
|Race/ethnicity, n (%)||.586|
|Non-Hispanic white||11 (10.6)||8 (10.3)|
|Non-Hispanic black||72 (69.2)||60 (76.9)|
|Hispanic||18 (17.3)||8 (10.3)|
|Asian||3 (2.9)||2 (2.6)|
|Nulliparous, n (%)||37 (35.6)||33 (42.3)||.356|
|Married (n = 181), n (%)||33 (31.7)||26 (33.7)||.773|
|Foreign-born (n = 180), n (%)||51 (49.5)||22 (28.6)||.005|
|Employed (n = 177), n (%)||47 (46.5)||30 (40.8)||.446|
|Educational level (n = 148), n (%)||.343|
|Less than high school||19 (21.4)||7 (11.9)|
|High school graduate/general educational development||37 (41.6)||28 (47.5)|
|Some college||16 (18.0)||15 (25.4)|
|College graduate or beyond||17 (19.1)||9 (15.3)|
|Insurance status (at first prenatal visit), n (%)||.935|
|Uninsured||14 (13.5)||12 (15.4)|
|Public||71 (68.3)||52 (66.7)|
|Private||19 (18.3)||14 (18.0)|
|Disclosed to sexual partner (n = 179), n (%)||66 (64.1)||57 (75.0)||.119|
|New HIV diagnosis during pregnancy, n (%)||38 (36.5)||15 (19.2)||.011|
|Years since HIV diagnosis b||2 (0–6)||5 (1–11)||<.001|
|History of mental health diagnosis, n (%)||40 (38.5)||33 (42.3)||.600|
|Substance abuse, n (%)||21 (20.2)||15 (19.2)||.872|
|Tobacco use (n = 179), n (%)||23 (22.3)||18 (23.7)||.831|