Risk factors for postpartum depression in women living with HIV attending Prevention of Mother–to-Child Transmission (PMTCT) Clinic at Kenyatta National Hospital, Nairobi

Abstract

Mothers with HIV are at high risk of a range of psychosocial issues that may impact HIV disease progression for themselves and their children. Stigma has also become a substantial barrier to accessing HIV/AIDS care and prevention services. The study objective was to determine the prevalence and severity of postpartum depression (PPD) amongst women living with HIV and to further understand the impact of stigma and other psychosocial factors in 123 women living with HIV attending Prevention of Mother to Child transmission (PMTCT) clinic at Kenyatta National Hospital (KNH) located in Nairobi, Kenya. We used the EPDS scale and HIV/AIDS Stigma Instrument – PLWHA (HASI – P). Forty-eight percent (N=59) of women screened positive for elevated depressive symptoms. Eleven (9%) of the participants reported high levels of stigma. Multivariate analyses showed that lower education (OR=0.14, 95% CI [0.04 – 0.46], p=0.001) and lack of family support (OR=2.49, 95% CI [1.14 – 5.42], p=0.02) were associated with presence of elevated depressive symptoms. The presence of stigma implied more than 9 fold risk of development of PPD (OR=9.44, 95% CI [1.132–78.79], p=0.04). Stigma was positively correlated with an increase in PPD. PMTCT is an ideal context to reach out to women to address mental health problems especially depression screening and offering psychosocial treatments bolstering quality of life of the mother-baby dyad.

Background

Mothers with HIV face a range of psychosocial problems, including postpartum depression (PPD) (Vesga-Lopez, Blanco, Keyes, Olfson, Grant, & Hasin, 2008) which impacts HIV disease progression in the mother and has lasting impacts for child health (Hartley, Pretorius, Mohamed, Laughton, Madhi, Cotton & Seedat, 2010). Depression is a highly prevalent co-morbidity among HIV+ individuals (Owe-Larsson, Sall, Salamon, Allgulander, 2009). It is inversely correlated with self-esteem, infant health status, and years of formal education (Ross, Sawatphanit, Mizuno & Takeo, 2011). High prevalence of depressive symptoms amongst pregnant HIV+ women areassociated with increased risk of adverse pregnancy outcomes and poor quality of life (Kapetanovic, Dass-Brailsford, Nora & Talisman, 2014). Additionally, women with HIV experience lower levels of emotional support available to them (Bonacquisti, Geller, Aaron, 2014).

Perinatal depression is reported to be as high as 30–50 % in South Africa (Chibanda et al., 2010; Hartley et al., 2011; Rochat, Tomlinson, Barnighausen, Newell, and Stein, 2011; Stewart et al., 2010). InNyanza province of Kenya HIV prevalence is as high as 20.7% in antenatal care settings (Dillabaugh et al., 2012). Stigma is known as a substantial barrier in adhering to and accessing HIV/AIDS care. Furthermore stigma contributes to depressive symptomatology (Rao et al., 2012) compounding the negative impact on women living with HIV.

Methods

Setting and Participants

We conducted a cross-sectional study with women attending PMTCT clinic at Kenyatta National Hospital (KNH). The clinic serves an average of 160–240 postnatal women every month who are primarily from urban and periurban settlements within Nairobi. Our participants were 18–50 years oldpostnatal women living with HIV recruited at 8-weeks post-delivery. This allowed time for the PCR testing from 6 weeks onwards to ascertain HIV status of their baby. This study was approved by University of Nairobi/Kenyatta National Hospital Ethics and Review Committee (ERC no. P171/03/2014). All postnatal women with severe depressive symptoms, suicidal ideation and alcohol abuse disorder were offered psychosocial support by the researcher and thereafter referred to the Department of Mental health at KNH.

Measurements

We gathered information on participants’ socio-demographics (age, marital status, educational level, occupation, and socio-economic status), clinical information, and psychosocial information. Probes were made on quality of support received from family, significant others on alcohol use and experience with domestic violence. We also assessed presence of STIs, HIV status of the child, and breast/formula feeding practice to understand associated challenges better. On hindsight the response options assessing these associated risks were not as elaborate generating limited information and we do acknowledge this as a limitation of our study. We used the 10-item Edinburgh Postnatal Depression Scale (EPDS) to screen for and examine severity of depressive symptoms (Cox, Holden, & Sagovsky, 1987). EPDS is an internationally validated tool for identifying patients at risk for perinatal depression and frequently used in Sub-Saharan Africa (Tsai, Bangsberg, Frongillo, Hunt, Muzoora, Martin, & Weiser, 2013)., Women who scored above 12 were identified as having elevated untreated depressive symptoms. HIV/AIDS Stigma Instrument – PLWHA (HASI–P) is a 33-item instrument covering six dimensions of HIV-related stigma: verbal abuse, negative self-perception, and health care neglect, as well as dimensions such as social isolation, fear of contagion, workplace stigma and total perceived stigma (Holzemer et al., 2007). It is built on a four point likert scale with responses ranging from 0 (never) to 3 (mostly). We calculated the median overall stigma and used the median value to categorize the participant to either: 0-never, 1-once or twice, 2-severally, 3-mostly. We found that those with median of 1, 2, and 3 were few and combined them to have a new variable: no stigma and presence of stigma. We found very few responses ranging from once to most options thus we decided to turn stigma from a continuous score into a binary one..

Statistical Analysis

We analyzed data using the SPSS version 20. We employed descriptive univariate analysis to describe the socio-demographics, psychosocial risk factors and depression prevalence. To test relationships amongst these variables, we performed bivariate analyses using chi-square/Fisher’s exact and Kendall’s tau-b tests. Finally multivariate regression model was fitted with depression as an outcome and predictors such as education, family support and levels of HIV/AIDS stigma that were associated in the bivariate analysis with a p value of 0.05. All relationships were described with their odds ratio (OR) with their 95% confidence intervals..

Results

Prevalence of Postpartum depression

The mean age of women in our study was 31years (N=123, SD= 5.2). The EPDS mean score was 11.53 (SD= 5.7) and fifty nine (48%) of our participants met screening criteria for elevated depressive symptoms. We did find it a matter of concern that 29% (n=36) of our participants had suicidal ideation (see Table 2).

Table 2.

Prevalence of PPD and Associated features

PPD and associated features	Category	N	%
Total EPDS score	Mean 11.53, SD 5.7
EPDS score (Ranges 0–30 for Non-Elevated depressive symptoms and Elevated depressive symptoms)	Non-Elevated depressive symptoms	64	52
	Elevated depressive symptoms	59	48
EPDS suicidal ideation intensity	None	87	70.7
	Mild	7	5.7
	Moderate	23	18.7
	Severe	6	4.9
Suicidal ideation	Absent	87	70.7
	Present	36	29.3
Months since birth of child	0–3 months	22	17.9
	4–6 months	32	26.0
	7–9 months	27	22.0
	10–12 months	21	17.1
	13–15 months	8	6.5
	16–18 months	8	6.5
	19–21 months	4	3.3
	22–24 months	1	0.8

Predictors of PPD among women living with HIV

On multivariate analyses (see Figure 1), EPDS score >12 was strongly associated with level of education (χ²(1) =13.60, p<0.0001). Elevated depressive symptoms were also associated with lack of family support (χ² (1) =6.30, p=0.012). Elevated depressive symptoms were also associated with overall stigma (χ² (3) =9.23, p=0.03) and particularly with negative self-perceived stigma type (χ² (3) =23.17, p<0.0001) where poor self-efficacy is co-terminus with the experience of stigma.

Figure 1.

Predictors of PPD among women living with HIV

*Association is significant at the 0.05 level (2-tailed) **Association is significant at the 0.01 level (2-tailed)

Eleven (9%) of the participants reported high levels of stigma. Multivariate analyses showed that lower education (OR=0.14, 95% CI [0.04 – 0.46], p=0.001) and lack of family support (OR=2.49, 95% CI [1.14 – 5.42], p=0.02) were associated with presence of elevated depressive symptoms. Furthermore, the presence of stigma implied more than 9 fold risk of development of PPD (OR=9.44, 95% CI [1.132–78.79], p=0.04).

Discussion

We found a large proportion of postpartum women living with HIV experience elevated depressive symptoms. Family social support, educational level, and stigma are common set of risk factors aggravating distress in PPD (Fisher, Mello, Patel, Rahman, Tran, Holton & Holme, 2012; Rao et al., 2012).

PPD was shown to be as high as 30–50 % in multiple studies in Africa (Hartley et al., 2011; Chibanda et al., 2010; Rochat et al., 2011; Stewart et al., 2010). High prevalence of clinically significant depressive symptoms and suicidal ideation reported by women in our study is consistent with studies conducted in the region (Gavin, Tabb, Melville, Guo, & Katon, 2011). Our findings can be situated along the following themes.

1. Depression as a significant challenge in HIV positive women in perinatal spectrum: Fewer studies have been conducted on PPD among women living with HIV. A Ugandan study found 39% of their 447 HIV positive participants from ages 18–49 years screened positive for probable depression (Kaida et al., 2012). Higher levels of HIV-related stigma were significantly associated with elevated depressive symptoms (Endeshaw, Walson, Rawlins, Dessie, Alemu, Andrews, & Rao, 2014). Thus intervening at the PMTCT level itself might reduce psychiatric morbidity and improve adherence and engagement in the program (Abrams, Myer, Rosenfield, & El-Sadr, 2007).

2. Addressing disempowerment and the absence of support at familial and social levels: Our study participants experienced elevated depressive symptoms and a veritable lack of social support. Social support is known as a strong protective factor against PPD (Robertson, Grace, Wallington. & Stewart, 2004) and support from family members acts as a buffer against depression in women in LMIC settings (Broadhead, Abas, Khumalo, Sakutukwa, Chigwanda, & Garura, 2001). Education was positively associated with PPD in that women who report low rates of depressive symptoms comparatively have higher education (Rao et al., 2012; Prachakul, Grant & Keltner, 2007; Bennetts et al., 1999).

3. Stigma as a strong determinant of depression in women with HIV: Our results reconfirm that stigma has a strong association with PPD especially in Kenyan cultural context (Dlamini et al., 2007). Both experienced stigma and internalized stigma were strong predictors of PPD among HIV positive South African women (Peltzer & Shikwane, 2011). Other studies in the region too have found that women who had primary education or less have greater adjusted odds of substantial stigma (Cuca et al., 2012).

Study Limitations

EPDS is a screening instrument and not a clinical tool as such limiting our scope. As a pilot project we did not use elaborate psychometric tools to assess risk factors such as social support, alcohol consumption and intimate partner violence in greatrigor.

Caveats and Conclusion

PPD, if left untreated, has adverse effects on mothers and their infants. For the mother, the episode can be the precursor of chronic recurrent depression and for her children PPD can impede their social, emotional, cognitive and physical development (Logsdon, Wisner & Pinto-Foltz, 2006). A concern we noted in our study was select participants’ alcohol abuse despite being on ARVs. Substance use is one of the known barriers to HIV treatment adherence apart from medication side effects and depression (Berg, Michelson, & Safren, 2007) and is also a correlate of PPD (Rubin et al., 2011). Our findings point to the high prevalence of PPD in HIV context. We think that PMTCT might be an ideal context to reach out to women to address mental health problems especially depression screening and offering psychosocial treatments benefitting the mother-baby dyad.

Table 1.

Socio-demographics, clinical, and psychosocial characteristics of the sample (N=123)

Characteristic	Category	N (N %)	P value
Age	M 31.2, SD 5.2, Median 32 Range 19–48 years		NS
Number of children	M 2, SD 1, Median 2
Religion	Christian	121(98.4)	NS
	Muslim	2(1.6)
	Others	0(0.0)
Marital status	Single	29(23.6)	NS
	Married	84(68.3)
	Divorced	2(1.6)
	Separated	7(5.7)
	Co-habiting	1(0.8)
Educational level	No formal education	1(0.8)	<0.0001**
	Primary	22(17.9)
	Secondary	38(30.9)
	Diploma courses/Midlevel colleges	56(45.5)
	UG/PG University	6(4.9)
Occupation	Unemployed	41(33.3)	NS
	Employed	26(21.1)
	Self employed	56(45.5)
Income in KES	Up to 10,000	68(55.3)	NS
	10,000–20,000	25(20.3)
	20,000–30,000	8(6.5)
	30,000–40,000	10(8.1)
	40,000–50,000	3(2.4)
	above 50,000	9(7.3)
HIV and PMTCT related factors			NS
When were you diagnosed as HIV positive	Before pregnancy	64(52.0)
	Antenatal clinic	50(40.7)
	During delivery	5(4.1)
	After delivery	4(3.3)
Where did you deliver the child	Hospital	123(100.0)	NS
	Home	0(0.0)
Did your child cry immediately after delivery	Yes	113(91.9)	0.034*
	No	10(8.1)
Do you know HIV status of child	Positive	4(3.3)	NS
	Negative	107(87.0)
	not sure	12(9.8)
Does your child experience frequent sickness	Yes	14(11.4)	NS
	No	109(88.6)
How do you feed your child at the moment	Exclusive breastfeeding	64(52.0)	NS
	Formula feeding	10(8.1)
	Mixed feeding	49(39.8)
Have you been treated from STI in the past one month	Yes	12(9.8)	NS
	No	111(90.2)
Rate your social support from family	Good	54(43.9)	0.012*
	Not good	69(56.1)
Rate your social support from friends	Good	37(30.1)	NS
	Not good	86(69.9)
Rate your social support from significant others	Good	19(15.4)	NS
	Not good	104(84.6)
Did you drink alcohol (beer, wine, home-brewed beer or spirits) in the past one month	Yes	20(16.3)	NS
	No	103(83.7)
Has your male partner abused you since the delivery of this child?	Physically	9(8.4)	NS
	Emotionally	21(19.6)
	None (supportive)	77(72.0)
Does your male partner engage in extramarital sexual affairs?	Yes	25(23.6)	NS
	No	78(73.6)
	Not sure	3(2.8)

^*chi-square and kendell’s tau b significant at the 0.05 level

^**significant at the 0.01 level