Abstract
In low-income, high HIV prevalence settings, caregiving for a young child who is HIV exposed and uninfected (HEU) can place additional demands on a mother living with HIV while she manages her own health and the well-being of her child. We compared caregiver burden associated with infants’ healthcare needs, perceived by mothers of infants who were HEU and HIV-unexposed, uninfected (HUU). This nested sub-study of a prospective cohort study was conducted within the public primary healthcare setting in Western Cape, South Africa. Eligible mothers were enrolled antenatally with known HIV status and with a child confirmed HIV negative at age 12 months by 20 May 2025. The primary exposure was maternal–child HIV status, and primary outcome was caregiver burden according to the modified 12-item Zarit Caregiver Burden Scale questionnaire (ZCBS; range 0–48), measured by maternal interview at child’s age 12 months. Logistic regression was used to compare caregiver burden (no-to-mild [ZCBS 0–10] vs moderate-to-high [ZCBS 11–48]), unadjusted and adjusted for maternal and household confounders. Among 187 mothers included (n = 91 of infants HEU, n = 96 of infants HUU), the median (interquartile range) ZCBS was 3(1–8); 22/91 (24.2%) mothers with HIV and 11/96 (11.5%) mothers without HIV perceived moderate-to-high caregiver burden (unadjusted odds ratio 2.46, 95% confidence interval (CI) = 1.12–5.43, P = .025). After adjusting for maternal age, income, number of cohabitating adults in the family and maternal depression, the adjusted odds ratio was 2.06, 95% CI = 0.85–4.98, P = .108. Living with HIV while caring for an infant who is HEU increases mothers’ caregiving burden during the critical period of early motherhood.
Introduction
Responsibilities of caregiving during early childhood are numerous—maintaining child’s health and nutrition, seeking care during sickness, promoting play and early learning, managing household, family, and financial duties. The weight of these responsibilities is intensified if the child has additional needs or a chronic condition [1], especially in low-income settings due to absence of structural supports [2–4].
The impact of caregiving on caregivers has been conceptualized as ‘caregiver burden’—the physical, psychological, social, and financial hardship perceived by a caregiver associated with demands of a care-dependent family member (sick child, elderly, chronically ill person, or person with a disability). While caregiver burden has been used as an abstract construct in qualitative studies, it has also been quantified through use of standardized tools like the Zarit Caregiver Burden Scale (ZCBS) and the Caregiver Burden Strain questionnaire [5, 6]. While this construct is not intended to imply that the person requiring care is burdensome, it can be a helpful measure to identify caregivers struggling to cope with caregiving demands and in need of support.
Among the many reasons why caregiver burden may become high are when caregivers’ well-being is compromised or care recipients’ needs are high. For instance, caregiving for a young child can place additional responsibilities on a mother living with HIV (MLHIV) while she simultaneously manages their own illness, fear of vertical HIV transmission (VHT), breastfeeding dilemmas, stigma, depression, and poor family or social support [7–9]. Advances in antiretroviral therapy (ART) have reduced morbidity in MLHIV as well as VHT rates. Despite this, in high HIV prevalence settings, in the absence of reduction in antenatal HIV prevalence, a substantial number of women still live with HIV, conceive and give birth. Providing care to a young child exposed to HIV, who requires post-natal prophylactic medication, repeated HIV testing, and may have additional health, growth and developmental challenges, could place unique caregiving demands on mothers through the need for additional care seeking and caregiving [10–16]. Managing these needs could negatively impact a mother’s own health and retention in HIV care [17], and alter the responsiveness of caregiving, thus affecting the child’s well-being.
Therefore, it is important to quantify the caregiver burden perceived by MLHIV and explore contributing factors. In this regard, children living with HIV (CLHIV) and HIV orphans have been well studied with caregivers reporting a high burden [18–20]. However, there is a lack of evidence on the burden experienced by mothers of children who are HIV exposed and uninfected (HEU).
Given the package of early life exposures, risk factors and additional vulnerability in children who are HEU, we hypothesize that caregiver burden associated with their healthcare needs may differ from that of children who are HIV-unexposed, uninfected (HUU). This study measured caregiver burden related to the healthcare needs of children during infancy, as perceived by mothers of children who are HEU compared to those of children who are HUU. Secondly, this study aimed to determine maternal, child, and socio-economic factors associated with this burden.
Materials and methods
Study design and participants
This was a sub-study of a larger prospective cohort study CHERISH (Children HIV Exposed Uninfected—Research to Inform Survival and Health) conducted in the Western Cape province of South Africa from 2020 to 2025 [21]. The Western Cape has an antenatal HIV prevalence of 16% (73% of all pregnant women living with HIV on ART prior to pregnancy) and VHT rate was 1.5% at 14 weeks in 2022 [22]. This sub-study was conducted at the rural CHERISH site in Worcester, Breede Valley [22].
For the main CHERISH cohort study, pregnant women were recruited antenatally between 24 and 36 weeks gestation in a 1:1 ratio with and without HIV from the three largest antenatal clinics in Breede Valley. This sub-study included CHERISH mother–child dyads if child participants had reached 12 months of age by May 2025. For infants born to MLHIV, HIV transmission were excluded through HIV-PCR tests at birth and 10 weeks and HIV antibody rapid test at 12-month study visit (for all infants regardless of maternal HIV status). All mothers enrolled as living without HIV had repeat HIV rapid antibody testing at their child’s 12 month visit to exclude post-natal maternal HIV acquisition. This fulfils DECIPHER criteria for classification of children HEU and HUU with high certainty at age 12 months [23]. Only those mothers whose infants were confirmed as HIV negative at 12 months retest were included in the analysis. For mothers who delivered twins, only one child was enrolled to avoid maternal and household clustering in the analysis.
Data collection
The primary exposure variable was mother’s HIV status (with or without HIV) collected at enrolment in CHERISH along with her child’s HIV status, established around 12 months of age (HEU or HUU). For clarity in this analysis, mothers were classified as mothers of children HEU and HUU. Additional demographic, socio-economic, medical, and HIV-related characteristics were collected at scheduled time points from antenatal enrolment through to delivery, 3, 6, 9, and 12 months (Supplementary Table S1). For the aims of this sub-study at 12 months, additional questions were asked of mothers related to caregiving support, perceptions about infant’s overall health and satisfaction with health services.
The primary outcome variable was caregiver burden perceived by mothers related to their child’s healthcare needs during infancy measured with the ZCBS at 12 months. The original tool includes a 29-item questionnaire on how often caregiving created a physical, social, financial, or psychological burden and answers are recorded on a 5-point Likert scale ranging from never (0) to nearly always (4) [5]. We used the modified 12-item version which has been shown to be reliable and valid in a South African setting [19]. The questionnaire was translated to isiXhosa and Afrikaans, piloted on a small group of mothers and subsequently modified for contextual relevance (Supplementary Methods). Cronbach’s alpha was 0.82 reflecting good internal consistency of the 12-item ZCBS for this study.
Data management and statistical analysis
Previous research on caregiver burden has shown that 40% of caregivers reported a high burden associated with needs of children who are HIV affected [19]. Using this baseline proportion as a reference and an estimated 20% difference in perceived burden between HEU and HUU groups, (α = 0.05, power = 80%), a sample size of approximately 197 was computed using the OpenEpi calculator [24]. This sample size was achieved by 20 May 2025 (cut-off point for data collection and extraction).
Questionnaires were administered on a tablet and entered in the study database within the Stellenbosch University instance of REDCap [25]. We constructed the final exposure variables by aggregating maternal responses across time points which were utilized in subsequent analysis (Supplementary Table S1).
The unit of analysis was the mother–child dyad. Maternal and infant exposure variables were summarized as frequency and percentages for categorical variables, whereas mean with standard deviation (for normally distributed data) and median with interquartile range (IQR) (for skewed data) were calculated for numeric variables. Differences in summary measures of exposure variables for both groups were calculated by two sample t-test or Wilcoxon rank sum test (numeric variables) and Chi-square or Fisher exact test (categorical variables) with P-values. Summary measures for HIV-related exposures were separately computed for mothers of infants who are HEU. Observations with missing or unknown values were excluded and summary measures were calculated using the revised denominators.
Caregiver burden was analysed both as a continuous and categorical variable. ZCBS median (IQR) score was calculated and compared between groups. A cut-off of 17 has previously been used on the 12-item version to denote higher caregiver burden [19]. Due to the distribution of total scores in our sample (Supplementary Fig. S2), we classified caregiver burden into two categories for analysis (0–10: no-to-mild burden, 11–48: moderate-to-high burden). Additionally, frequency of high burden responses for each ZCBS item was calculated as the proportion of mothers who reported ‘quite frequently’ (Likert scale response 3) or ‘nearly always’ (Likert scale response 4).
Unadjusted odds ratio with 95% confidence intervals (CIs) was estimated for HIV status and perceived caregiver burden, comparing mothers of children HEU and HUU. Unadjusted odds ratios were also calculated to assess the associations between individual exposure variables and caregiver burden. Multivariable logistic regression analysis was subsequently conducted to evaluate factors associated with caregiver burden. Initially, we examined the odds ratio for HIV status and perceived caregiver burden after adjusting for each additional exposure variable. A final adjusted model was then constructed using a backward stepwise regression approach to control for eligible covariates (potential confounders). These covariates were selected a priori (based on a directed acyclic graph and known predictors from the literature; Supplementary Fig. S1) as well as those which had the highest effect on unadjusted odds ratio of association between exposure and outcome in univariable analysis. We intentionally excluded covariates related to mothers’ health during pregnancy and infancy, support received with caregiving and various infant related factors since they appear to be on the causal pathway between exposure and outcome and may act as potential mediators. Birth weight and preterm birth were explored as potential effect modifiers through stratified analysis.
Ethical approval was obtained from Health Research Ethics Committee (HREC) at Stellenbosch University (approval number S22/09/182). Mothers provided written informed consent in Afrikaans, isiXhosa, or English at antenatal enrolment, again after delivery for enrolment of their child and consent was affirmed at the 12-month visit including for participation in this sub-study.
Results
One hundred and eighty-seven mothers were included in the final analysis (91 with HIV and 96 without HIV), all with infants confirmed to be HIV negative at age 12 months (Fig. 1).
Figure 1.
Flow diagram illustrating participant enrolment, inclusion, exclusion, and follow-up.
Compared to mothers without HIV, MLHIV were older, had higher parity, less likely to complete grade 12 or have any income, more often lived in informal dwellings and less often had electricity and running water in the house (Table 1). Mothers were living in households with median of two cohabitating adults (IQR 2–3), and they were caring for median of one child (IQR 0–3) (including index child).
Table 1.
Maternal characteristics compared by HIV status.
| Characteristic | Tot al (N = 187) | Mothers without HIV (N = 96) | Mothers with HIV (N = 91) | P-values |
|---|---|---|---|---|
| Demographic and socio-economic characteristics (at antenatal baseline) | ||||
| Age (years)—mean (SD) | 30.2 (6.2) | 28.3 (5.8) | 32.1 (6.0) | .000 |
| Parity—median (IQR) | 1 (0–2) | 1 (0–1) | 2 (1–3) | .000 |
| Cohabitating with partner—n (%) | 86 (46.0) | 38 (39.6) | 48 (52.7) | .071 |
| Cohabitating adults in family—median (IQR) | 2 (2–3) | 3 (2–4) | 2 (2–3) | .015 |
| Cohabitating children in family—median (IQR) | 1 (0–3) | 1 (0–2.5) | 1 (0–3) | .753 |
| Education—n (%) | .000 | |||
| Grade 12 | 74 (39.6) | 54 (56.2) | 20 (21.9) | |
| Secondary | 104 (55.6) | 39 (40.6) | 65 (71.4) | |
| Primary or none | 9 (4.8) | 3 (3.1) | 6 (6.6) | |
| Employed (full time or part time)—n (%) | 53 (28.3) | 33 (34.4) | 20 (21.9) | .060 |
| Any income—n (%) | 56 (29.9) | 36 (37.5) | 20 (21.9) | .021 |
| Child support grant received—n (%) | 133 (71.1) | 72 (75.0) | 61 (67.0) | .186 |
| Type of housing—n (%) | .003 | |||
| Informal dwelling | 98 (52.4) | 40 (41.7) | 58 (63.7) | |
| Formal dwelling | 89 (47.6) | 56 (58.3) | 33 (36.3) | |
| Electricity in house—n (%) | 159 (85.0) | 88 (91.7) | 71 (78.0) | .009 |
| Running water inside house—n (%) | 74 (39.6) | 45 (46.9) | 29 (31.9) | .036 |
| Food insecurity ever reported during child’s infancy—n (%) | 42 (22.5) | 18 (18.7) | 24 (26.4) | .212 |
| Maternal health (ever during pregnancy or child’s infancy) | ||||
| Ever treated for any chronic conditionsa (other than HIV)—n (%) | 23 (12.3) | 10 (10.4) | 13 (14.3) | .421 |
| Ever hospitalized—n (%) | 18 (9.6) | 10 (10.4) | 8 (8.8) | .706 |
| Ever screened positive for depression on EPDSb—n/N (%) | 16/181 (8.8) | 7/93 (7.5) | 9/88 (10.2) | .522 |
| Ever reported smoking—n (%) | 31 (16.6) | 21 (21.9) | 10 (10.9) | .045 |
| Ever reported consuming alcohol—n (%) | 47 (25.1) | 29 (30.2) | 18 (19.8) | .100 |
| Support with caregiving (at 12-month visit) | ||||
| Enough support from father of child—n (%) | 133 (71.1) | 70 (72.9) | 63 (69.2) | 0.578 |
| Enough support from family—n (%) | 162 (86.6) | 90 (93.7) | 72 (79.1) | 0.003 |
| Enough support from community—n (%) | 105 (56.1) | 52 (54.2) | 53 (58.2) | 0.575 |
| Completely satisfied with healthcare services support—n (%) | 161 (86.1) | 83 (86.5) | 78 (85.7) | 0.883 |
| Caregiver burden perceived (Zarit Caregiver Burden Scale [ZCBS] score—out of 48) | ||||
| ZCBS score—median (IQR) | 3 (1–8) | 3 (1–8) | 4 (1–9) | .320c |
| ZCBS score—mean (SD) | 5.5 (6.4) | 5 (5.9) | 6.0 (6.9) | |
| ZCBS score categories—n (%) | .023d | |||
| Moderate-to-high burden (11–48) | 33 (17.6) | 11 (11.5) | 22 (24.2) | |
| Low-or-no burden (0–10) | 154 (82.3) | 85 (88.5) | 69 (75.8) | |
Any medical or obstetric conditions that received treatment during pregnancy, delivery, or infancy—tuberculosis, hypertension, diabetes, depression, other mental health condition, hypertensive disorders of pregnancy, gestational diabetes, asthma, anaemia.
Edinburgh Postnatal Depression Scale (EPDS) score ≥10.
Based on Wilcoxon rank sum test (comparing median for non-parametric outcome).
Based on Chi-square test (compare proportions between two groups).
Of 187 mothers, 54 (28.9%) reported having inadequate or poor support from the father of their child and 25/187 (13.4%) reported having inadequate or poor support from their family. MLHIV were less often supported by the family with caregiving (19/91, 20.9% HEU vs 6/96, 6.3% HUU, P = .003).
Of 187 infants, 27 (14.4%) had low birth weight (<2500 g) and 26/177 (14.7%) were born preterm (gestational age <37 weeks) with no significant difference between groups (Table 2). Infants HEU were less likely to breastfeed at 3 months, 9 months, and 12 months. More than half of all infants (98/177, 55.4%) started solid foods at 3 months with no significant difference between groups. Compared to infants who were HUU, those who were HEU were more likely to be underweight (weight for age Z score, WAZ< −2), stunted (length for age Z score, LAZ< −2) at age 12 months, and hospitalized during infancy. Mothers of infants who were HEU were more likely to report that their child had been sicker than same age peers compared to mothers of infants HUU.
Table 2.
Infant characteristics compared by HIV exposure status.
| Characteristic | Total (N = 187) | Infants HUU (N = 96) | Infants HEU (N = 91) | P-values |
|---|---|---|---|---|
| Birth related | ||||
| Sex—n (%) | .266 | |||
| Male | 97 (51.9) | 46 (47.9) | 51 (56.0) | |
| Female | 90 (48.1) | 50 (52.1) | 40 (43.9) | |
| Multi-gestation (twins or triplets)—n (%) | 5 (2.7) | 2 (2.08) | 3 (3.3) | .676 |
| Birth weight (grams)—mean (SD) | 2998 (501) | 3033 (509) | 2961 (493) | .334 |
| Gestational age at delivery (weeks)—median (IQR) | 39 (38–40) | 39 (38–40) | 39 (38–40) | .650 |
| Preterm birth (<37 weeks)—n/N (%) | 26/177 (14.7) | 15/93 (16.1) | 11/84 (13.1) | .569 |
| Low birth weight (<2500 g)—n (%) | 27 (14.4) | 14 (14.6) | 13 (14.3) | .954 |
| Feeding received | ||||
| Any breastfeeding at 3 months—n/N (%) | 103/180 (57.2) | 64/94 (68.1) | 39/86 (45.3) | .002 |
| Any breastfeeding at 6 months—n (%) | 80 (42.8) | 46 (47.9) | 34 (37.4) | .145 |
| Any breastfeeding at 9 months—n (%) | 58 (31.0) | 37 (38.5) | 21 (23.1) | .022 |
| Any breastfeeding at 12 months—n (%) | 50 (26.7) | 32 (33.3) | 18 (19.8) | .036 |
| Any solid food at 3 monthsa—n/N (%) | 98/177 (55.4) | 46/92 (50.0) | 52/85 (61.2) | .135 |
| Any solid food at 6 monthsa—n (%) | 175 (93.6) | 89 (92.7) | 86 (94.5) | .616 |
| Child’s age at 12-month visit (months)—median (IQR) | 15 (13–18) | 15 (13–18) | 14 (12–19) | .332 |
| Growth and development at 12-month visit | ||||
| Z scores—median (IQR) | ||||
| Weight for age (WAZ) |
|
|
|
.001 |
| Length for age (LAZ) |
|
|
|
.011 |
| Weight for length (WHZ) |
|
|
|
.016 |
| Underweight (WAZ < −2)—n/N (%) | 17/182 (9.3) | 3/95 (3.2) | 14/87 (16.1) | .003 |
| Stunted (LAZ < −2)—n/N (%) | 44/180 (24.4) | 18/95 (18.9) | 26/85 (30.6) | .070 |
| Wasted (WHZ < −2)—n/N (%) | 1/180 (0.5) | 0/95 (0.0) | 1/85 (1.2) | .472 |
| Development as expected at 12 monthsb—n (%) | 169 (90.4) | 88 (91.7) | 81 (89.0) | .538 |
| Mother concerned about child’s development at 12 months—n (%) | 14 (7.5) | 7 (7.3) | 7 (7.7) | .917 |
| Child health | ||||
| Immunization up to date at 12 months—n/N (%) | 142/161 (88.2) | 78/88 (88.6) | 64/73 (87.7) | .850 |
| Child ever reported being sick in first year—n/N (%) | 21/181 (11.6) | 10/94 (10.6) | 11/87 (12.6) | .674 |
| Child ever hospitalized in first year—n/N (%) | 50/181 (27.6) | 20/94 (21.3) | 30/87 (34.5) | .047 |
| Child started on tuberculosis treatment during first year—n/N (%) | 9/180 (5.0) | 5/94 (5.3) | 4/86 (4.6) | 1.000 |
| Mother perceived child’s overall health in first year was not so good or poor—n (%) | 30 (16.04) | 12 (12.5) | 18 (19.8) | .175 |
| Mother perceived child had been more sick than other children in first year—n (%) | 20 (10.7) | 6 (6.3) | 14 (15.4) | .043 |
| Daycare support—Child stays at home—n (%) | 152 (81.3) | 74 (77.1) | 78 (85.7) | .130 |
Any solid food at 3 and 6 months—if child was fed any solids once or more since yesterday morning, other than drinking milk or liquids.
Based on Development Screening Questionnaire (DSQ) scores.
Among MLHIV, they were living with HIV for a median of 6 years (IQR 2–10) by the time their child reached 12 months of age. Of 91 mothers, 66 (72.5%) initiated ART before the index pregnancy and 77/91 (84.6%) mothers were on tenofovir–lamivudine–dolutegravir (SDC Table 2). Except for three, all infants born to MLHIV received antiretroviral prophylaxis at birth with the majority receiving nevirapine and zidovudine.
The overall median ZCBS score was 3 (IQR 1–8), median 4 (IQR 1–9) in MLHIV and median 3 (IQR 1–9) in mothers without HIV (Table 1) (Supplementary Results and Fig. S2). More MLHIV perceived moderate-to-high caregiver burden (score > 10; 22/91, 24.2% HEU vs 11/96, 11.5% HUU, P = .023) as compared to those without HIV. Frequency of high burden responses was more in MLHIV for ZCBS item 6, 9, and 12. However, the difference was not statistically significant (Table 3).
Table 3.
High burden responsesa for Zarit Caregiver Burden Scale (ZCBS) individual items.
| Item no. | ZCBS Scale items | Total (N = 187)b | Mothers of infants HUU (N = 96)b | Mothers of infants HEU (N = 91)b | P-values |
|---|---|---|---|---|---|
| 1 | How often do you feel that because of the time you have to spend caring for your child’s healthcare needs, you don’t have enough time for yourself? | 10 (5.3) | 5 (5.2) | 5 (5.5) | 0.931 |
| 2 | How often do you feel that you are stressed between seeking healthcare for the child and trying to meet other responsibilities for your family or work? | 7 (3.7) | 3 (3.1) | 4 (4.4) | 0.715 |
| 3 | How often do you have to be absent from your work/lose wages because of visit to the clinic/hospital for your child’s healthcare needs? | 11 (5.9) | 6 (6.2) | 5 (5.5) | 1.000 |
| 4 | How often do you feel that you have to ignore your own healthcare needs/your health has suffered because of child’s clinic/hospital visits? | 7 (3.7) | 4 (4.2) | 3 (3.3) | 1.000 |
| 5 | How often do you feel that time spent on child’s healthcare visits affects your relationship with family members and friends in a negative way? | 1 (0.5) | 0 | 1 (1.1) | 0.487 |
| 6 | How often do you feel angry/strained because of having to take the child to a clinic/hospital? | 7 (3.7) | 2 (2.1) | 5 (5.5) | 0.269 |
| 7 | How often do you feel that your social life has suffered because of having to care for your child’s healthcare needs and clinic visits? | 1 (0.5) | 0 | 1 (1.1) | 0.487 |
| 8 | How often do you feel that you don’t have as much privacy because of having to care for your child’s healthcare needs? | 4 (2.1) | 2 (2.1) | 2 (2.2) | 1.000 |
| 9 | How often do you feel that you are uncertain about how to manage your child’s healthcare needs? | 4 (2.1) | 1 (1.0) | 3 (3.3) | 0.358 |
| 10 | How often do you feel that you have lost control over your own life due to your child’s healthcare needs? | 0 | 0 | 0 | 0 |
| 11 | How often do you feel that you need to be doing more to meet your child’s healthcare needs? | 20 (10.7) | 11 (11.5) | 9 (9.9) | 0.729 |
| 12 | How often do you feel that you could do a better job in terms of caring for your child’s health? | 22 (11.8) | 8 (8.3) | 14 (15.4) | 0.135 |
High burden response defined as how many mothers reported ‘quite frequently’ (Likert scale response = 3) or ‘nearly always’ (Likert scale response = 4) for the question.
Values shown as n (column %) per group. Percentages are calculated within each group.
MLHIV had higher odds of perceiving moderate-to-high caregiver burden as compared to mothers without HIV (unadjusted odds ratio [OR] = 2.46, 95% CI = 1.12–5.43, P = .025). Mothers with more cohabitating adults and more children in the family had lower odds of perceiving moderate-to-high caregiver burden, whereas those who did not have much support with caregiving from the father of the child, or the family and were not satisfied with healthcare services, had higher odds of perceiving moderate-to-high caregiver burden. Similarly, mothers of infants with low birth weight, started on solid food at 3 months and underweight or stunted at 12 months had higher odds of perceiving moderate-to-high caregiver burden. Further, mothers who perceived that their child’s health was poor during infancy had higher odds of perceiving moderate-to-high caregiver burden. Stratified analyses revealed that while there was no significant difference in perceived caregiver burden by HIV status in preterm and low birth weight infants. Mothers of infants HEU who were term born (20/73, 27.4% HEU vs 8/78, 10.3% HUU, P = .007) and normal birth weight (17/78, 21.8% HEU vs 7/82, 8.5% HUU, P = .019) were more likely to perceive moderate-to-high burden as compared to those of infants HUU who were term born and normal birth weight (Supplementary Tables S3 and S4).
Considering a parsimonious approach to modelling with this relatively small sample, variables determined a priori to be confounders between maternal HIV and caregiver burden (maternal age, parity, income, and depression) and additional variables with the largest effect on the association between HIV status and caregiver burden (cohabitating adults; Supplementary Table S5 for OR of HIV status adjusted for single additional variables) were prioritized in the multivariable model. After considering collinearity between maternal age and parity and between income and employment, ORs adjusted for maternal age, income, depression, and cohabiting adults were calculated. In this model, the adjusted OR (aOR) for HIV status was 2.06 (95% CI = 0.85–4.98, P = .108). A second model excluding depression showed aOR for HIV status of 2.11 (95% CI = 0.88–5.07, P = .094) (Table 4).
Table 4.
Odds ratios for association between maternal or child characteristics and moderate-to-high caregiver burden.
| Exposure variable | Reference category | Unadjusted OR (95% CI) for moderate-to-high caregiver burden | Model 1 aORa (95% CI) | Model 2 aORa (95% CI) |
|---|---|---|---|---|
| Mother’s HIV status (mother with HIV) | No HIV | 2.46 (1.12–5.43)c | 2.06 (0.85–4.98)d | 2.11 (0.88–5.07)e |
| Maternal age (years) | Per year | 1.01 (0.95–1.07) | 0.97 (0.91–1.04) | 0.97 (0.91–1.04) |
| Cohabiting adults in family | Per adult | 0.64 (0.42–0.94) | 0.59 (0.38–0.92) | 0.58 (0.37–0.90) |
| Any income | No income | 0.46 (0.19–1.30) | 0.45 (0.16–1.24) | 0.42 (0.15–1.14) |
| Ever screened positive for depression on EPDSb during pregnancy or child’s infancy | Never screened positive for depression | 1.63 (0.49–5.42) | 1.46 (0.41–5.28) |
aOR—adjusted odds ratio.
EPDS—Edinburgh Postnatal Depression Scale.
Unadjusted OR for association between mother's HIV status and moderate to high caregiver burden.
OR adjusted for maternal age, cohabiting adults, income and depression.
OR adjusted for maternal age, income and cohabiting adults (excluding depression).
For mothers of infants who were HEU, none of the HIV related characteristics were associated with perception of moderate-to-high caregiver burden (Supplementary Table S6).
Discussion
This study shows that MLHIV had almost 2.5 times higher odds of perceiving moderate-to-high caregiver burden related to healthcare needs of their infants who were HEU as compared to those with infants HUU. Even though overall ZCBS median score was quite low and only 17.6% of mothers perceived moderate-to-high burden, the unadjusted odds ratio, taking overall maternal experience into account, was higher for MLHIV. Despite the wide CI including null (likely attributed to limited statistical power), the attenuated odds ratio observed after adjusting for maternal age, income, number of cohabitating adults, and maternal depression highlights that part of the observed association is explained by contextual factors—particularly financial constraints and cohabitating adults (proxy for structural support with caregiving).
By applying a standardized measure of caregiver burden (ZCBS), this study provides novel insights on how caregiving demands affect mothers of infants who are HEU. A high burden of caregiving could potentially affect mothers’ physical and mental health, compromise the quality of caregiver–child interaction and influence children’s well-being in the long term [9, 26]. Thus, it is important to measure this burden and investigate its drivers, to identify which mothers are battling to cope and need targeted support.
Firstly, known maternal factors affecting caregiver burden include age, marital status, HIV related illness, ART handling, mental health, stigma, and fear of disclosure of HIV status [27–29]. Amongst all these factors, one of the consistent moderators is caregiver depression [9, 30, 31]. Our finding that MLHIV were more likely to experience symptoms of depression and report feelings of stress and inadequacy related to infants’ needs, could be attributed to their psychological distress, indicating that additional psychosocial support during early motherhood could be beneficial [18, 19, 32]. Another known determinant is caregivers’ ill health, in the context of HIV [19, 20, 28, 33]. However, contrary to evidence and our assumption, illness or hospitalizations of MLHIV during pregnancy or infancy were not significantly associated with caregiver burden in our study. This could be attributed to advances in ART regimens which have improved AIDS-related morbidity in MLHIV.
Secondly, a child who is HIV affected can impose additional needs on a mother. While some of these needs may be common to all children HIV affected, others may differ according to the child’s HIV status—living with HIV, HEU or HIV affected (without exposure). Mothers of CLHIV, in particular, are known to perceive a high caregiver burden as they need to manage their child’s ART medication, stigma attached to the diagnosis and frequent illness leading to missed school attendance [27, 34, 35]. Our study adds novel understanding of caregiver burden perceived by mothers of children HEU and demonstrates that although these children are HIV negative, the perceived caregiver burden can be substantial.
Children who are HEU are being increasingly recognized as vulnerable, as they face the adverse effects of being born to MLHIV [14–16, 36, 37]. Infants HEU in our study were more likely to be underweight, stunted and hospitalized and their mothers were less likely to breastfeed, all of which could increase caregiving demands—concerns about HIV transmission, extra clinic visits, hospital stays, and extra care at home during episodes of sickness or growth faltering. While most of these factors were individually associated with moderate-to-high caregiver burden, mediation analysis with larger samples would be required to quantify their contribution to perceived caregiver burden.
Thirdly, presence of HIV in the household increases caregiver burden and affects child behavioural outcomes [9, 38]. In our study, MLHIV were more likely to live under socio-economic constraints and lacked caregiving support from their partner and family, thus reflecting an unfavourable environment for caregiving. While these factors affect many mothers in low-income settings regardless of HIV status, they are disproportionately experienced by mothers living with HIV at the population level [39]. Financial constraints hinder mothers’ ability to arrange for child’s food, clothes, formula milk, and money for transport to health facility [19, 27, 28, 40]. In our study population, many mothers had temporary employment on farms and struggled with their children’s basic needs once the harvest season was over. Although in South Africa, child support grants are available for low-income caregivers, uptake remains inconsistent due to administrative and logistic issues [41]. This calls for a need to strengthen and expand social protection for MLHIV from low-income communities especially during early childhood [20, 42]. Similarly, support from a partner or family has been shown to improve MLHIV’s mental health and reduce their caregiver burden by sharing childcare and household responsibilities [20, 27, 43, 44]. In South Africa, interventions such as mentor mothers and post-natal clubs have shown potential in assisting MLHIV through peer support and informational access, in a non-stigmatizing environment [45, 46].
Our study has limitations. We could not determine the relative contribution or timing of maternal, child, and environmental factors on caregiver burden—an opportunity for future research. By measuring caregiver burden at 12 months, we may have missed the peak burden during early infancy when concerns about HIV transmission, breastfeeding, and anxiety of HIV testing were likely highest. As caregiving demands of mothers evolve, it is critical to assess temporality in terms of what contributes to their burden across stages of early motherhood. A small sample size and non-probability sampling limits the generalizability of our findings. Selection bias is possible as all our participants were already enrolled in CHERISH, may have been more health conscious, and had better access to psychosocial support. Future research should include more diverse populations of caregivers across different geographies.
In conclusion, caring for infants who are HEU increases mothers’ caregiver burden during the critical period of early motherhood, reinforcing the need to strengthen and expand caregiver support interventions that include economic security and psychosocial support for MLHIV, simultaneously addressing additional vulnerabilities specific to children who are HEU. At a service delivery level, while it is important to identify high-risk children who are HEU [36] and ensure that they receive routine child health services, it is equally necessary to quantify their mothers’ caregiver burden and identify those who need help and may benefit from early intervention [47]. There is scope for VHT prevention programmes to incorporate simple screening tools like ZCBS during routine visits by healthcare practitioners to measure caregiver burden and facilitate responsive assistance.
Supplementary Material
Acknowledgements
We thank the whole team of CHERISH cohort study based at Worcester, South Africa, and deeply appreciate their assistance with logistics, overall coordination, and insightful feedback during the data collection process. Our Research Assistants, Noluthando Sithole and Tiffany Schroeder, deserve special thanks for their assistance with the interviews. Our appreciation also extends to Heinrich Cupido and Andisiwe Bovu for their kind assistance with data management and analysis. Last but not the least, we are sincerely grateful to all the mothers who agreed to participate in this study and contributed their valuable time to answer our questions.
Contributor Information
Himani Pandya, Department of Global Health, Faculty of Medicine & Health Sciences, Stellenbosch University, Western Cape, Cape Town, 8000, South Africa.
Peter S Nyasulu, Division of Epidemiology & Biostatistics, Department of Global Health, Faculty of Medicine & Health Sciences, Stellenbosch University, Western Cape, Cape Town, 8000, South Africa; School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, 2193, South Africa.
Xanthe Hunt, Africa Health Research Institute, Somkhele, KwaZulu Natal, 3935, South Africa; Department of Global Health, Institute for Life Course Health Research, Stellenbosch University, Western Cape, Cape Town, 8000, South Africa; Mental Health, Alcohol, Substance Use and Tobacco Research Unit (MAST-RU), South African Medical Research Council, Western Cape, Cape Town, 7505, South Africa.
Amy L Slogrove, Department of Paediatrics and Child Health, Faculty of Medicine & Health Sciences, Stellenbosch University, Western Cape, Cape Town, 8000, South Africa.
Author contributions
Peter S. Nyasulu (Conceptualization [equal], Formal analysis [supporting], Writing—review & editing [supporting]), Xanthe Hunt (Conceptualization [supporting], Writing—review & editing [supporting]), and Amy L. Slogrove (Conceptualization [equal], Formal analysis [supporting], Writing—review & editing [equal])
Supplementary material
Supplementary material is available at Journal of Tropical Pediatrics online.
Conflicts of interest
The views represented in this article are our own and do not necessarily reflect those of the funding agency. ALS is employed by Elsevier for The Lancet Child and Adolescent Health. This work was done in her capacity as an Extraordinary Associate Professor at Stellenbosch University and not as a representative of Elsevier. None of the other authors have any conflicts of interest, financial or non-financial interests to disclose.
Funding
Financial support for this study was received from the US National Institute of Health Grant R33HD103093.
References
- 1. Macedo EC, da Silva LR, Paiva MS et al. Burden and quality of life of mothers of children and adolescents with chronic illnesses: an integrative review. Rev Lat Am Enfermagem 2015;23:769–77. 10.1590/0104-1169.0196.2613 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Peter R, Kumar KA. Mothers’ caregiving resources and practices for children under 5 years in the slums of Hyderabad, India: a cross-sectional study. WHO South East Asia J Public Health 2014;3:254–65. 10.4103/2224-3151.206748 [DOI] [PubMed] [Google Scholar]
- 3. Sarikhani Y, Najibi SM, Razavi Z. Key barriers to the provision and utilization of maternal health services in low-and lower-middle-income countries; a scoping review. BMC Womens Health 2024;24:325. 10.1186/s12905-024-03177-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Lobo E, Babu GR, Mukherjee D et al. It takes a village: unpacking contextual factors influencing caregiving in urban poor neighbourhoods of Bangalore, South India. Healthcare (Basel) 2025;13:1459. 10.3390/healthcare13121459 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Bédard M, Molloy DW, Squire L et al. The Zarit burden interview: a new short version and screening version. Gerontologist 2001;41:652–7. 10.1093/geront/41.5.652 [DOI] [PubMed] [Google Scholar]
- 6. Brannan AM, Heflinger CA, Bickman L. The Caregiver Strain Questionnaire: measuring the impact on the family of living with a child with serious emotional disturbance. J Emot Behav Disord 1997;5:212–22. 10.1177/106342669700500404 [DOI] [Google Scholar]
- 7. Fauk NK, Gesesew HA, Mwanri L et al. Understanding coping strategies of mothers living with HIV who care for children living with HIV: a qualitative study in Indonesia. BMC Womens Health 2023;23:172. 10.1186/s12905-023-02299-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Rochat T, Netsi E, Redinger S et al. Parenting and HIV. Curr Opin Psychol 2017;15:155–61. 10.1016/j.copsyc.2017.02.019 [DOI] [PubMed] [Google Scholar]
- 9. Sherr L, Skeen S, Hensels IS et al. The effects of caregiver and household HIV on child development: a community‐based longitudinal study of young children. Child Care Health Dev 2016;42:890–9. 10.1111/cch.12387 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Pandya H, Hunt X, Carstens T et al. Caring for a young child who is HIV exposed and whose HIV status is not yet known: a scoping review of mothers’ experiences. AIDS Care 2025;37:179–201. 10.1080/09540121.2024.2441251 [DOI] [PubMed] [Google Scholar]
- 11. Slogrove AL, Goetghebuer T, Cotton MF et al. Pattern of infectious morbidity in HIV exposed uninfected infants and children. Front Immunol 2016;7:164. 10.3389/fimmu.2016.00164 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Alvarenga WA, Dupas G. Experience of taking care of children exposed to HIV: a trajectory of expectations. Rev Lat Am Enfermagem 2014;22:848–56. 10.1590/0104-1169.3607.2489 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Cohen C, Moyes J, Tempia S et al. Epidemiology of acute lower respiratory tract infection in HIV exposed uninfected infants. Pediatrics 2016;137:e20153272. 10.1542/peds.2015-3272 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Prendergast AJ, Evans C. Children who are HIV exposed and uninfected: evidence for action. AIDS 2023;37:205–15. 10.1097/QAD.0000000000003409 [DOI] [PubMed] [Google Scholar]
- 15. Sudfeld CR, Lei Q, Chinyanga Y et al. Linear growth faltering among HIV exposed uninfected children. J Acquir Immune Defic Syndr 2016;73:182–9. 10.1097/QAI.0000000000001034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Wedderburn CJ, Weldon E, Bertran-Cobo C et al. Early neurodevelopment of HIV exposed uninfected children in the era of antiretroviral therapy: a systematic review and meta-analysis. Lancet Child Adolesc Health 2022;6:393–408. 10.1016/S2352-4642(22)00071-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Anderson K, Rabie H, Eley BS et al. Mother‐child dyads living with HIV in the Western Cape, South Africa: undetectable = undetectable? J Int AIDS Soc 2025;28:e26418. 10.1002/jia2.26418 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Inamdar SA, Kosambiya J, Modi A. Caregiver’s burden of children living with HIV on antiretroviral therapy at an urban setup. Indian J Community Med 2021;46:744–7. 10.4103/ijcm.IJCM_49_21 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Kidman R, Thurman TR. Caregiver burden among adults caring for orphaned children in rural South Africa. Vulnerable Child Youth Stud 2014;9:234–46. 10.1080/17450128.2014.929493 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Maddocks S, Erratum CV. Burden of caring for children living with human immunodeficiency virus in a semi-rural South African community. S Afr Fam Pract 2021;62:a5379. 10.4102/safp.v62i1.5379 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Slogrove AL, de Beer ST, Kalk E et al. Survival and health of children who are HIV exposed uninfected: study protocol for the CHERISH (Children HIV Exposed Uninfected—Research to Inform Survival and Health) dynamic, prospective, maternal-child cohort study. BMJ Open 2023;13:e070465. 10.1136/bmjopen-2022-070465 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Anderson K, Kalk E, Heekes A et al. Factors associated with vertical transmission of HIV in the Western Cape, South Africa: a retrospective cohort analysis. J Int AIDS Soc 2024;27:e26235. 10.1002/jia2.26235 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Slogrove AL, Burmen B, Davies MA et al. Standardized definitions of in utero human immunodeficiency virus and antiretroviral drug exposure among children. Clin Infect Dis 2022;75:347–55. 10.1093/cid/ciab845 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Dean AG, Sullivan KM, Soe MM. OpenEpi: Open Source Epidemiologic Statistics for Public Health. http://www.openepi.com/Menu/OE_Menu.htm
- 25. Harvey L. REDCap: web-based software for all types of data storage and collection. Spinal Cord 2018;56:625. [DOI] [PubMed] [Google Scholar]
- 26. Webster KD, de Bruyn MM, Zalwango SK et al. Caregiver socioemotional health as a determinant of child well‐being in school‐aged and adolescent Ugandan children with and without perinatal HIV exposure. Trop Med Int Health 2019;24:608–19. 10.1111/tmi.13223 [DOI] [PubMed] [Google Scholar]
- 27. Mujjuzi I, Mutegeki P, Nabuwufu S et al. Care burden and coping strategies among caregivers of paediatric HIV/AIDS in Northern Uganda: a cross‐sectional mixed‐method study. AIDS Res Treat 2021;2021:6660337. 10.1155/2021/6660337 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Kalomo EN, Liao M. Burden of care among caregivers of persons living with HIV/AIDS in rural Namibia: correlates and outcomes. Soc Work Public Health 2018;33:70–84. 10.1080/19371918.2017.1376132 [DOI] [PubMed] [Google Scholar]
- 29. Chauhan RC, Rai SK, Kant S et al. Burden among caregivers of children living with human immunodeficiency virus in North India. N Am J Med Sci 2016;8:129–33. 10.4103/1947-2714.179939 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Sherr L, Macedo A, Tomlinson M et al. Parenting in adversity: effects of older caregivers, biological carers and troubled carers on child outcomes in high HIV-affected communities. J Cross Cult Gerontol 2023;38:155–71. 10.1007/s10823-022-09474-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Asadullah M, Kamath R, Pattanshetty S et al. Psychological impact on caregivers of HIV-infected children in Udupi district, Karnataka. AIDS Care 2017;29:787–92. 10.1080/09540121.2016.1257090 [DOI] [PubMed] [Google Scholar]
- 32. Bass JK, Nakasujja N, Familiar-Lopez I et al. Association of caregiver quality of care with neurocognitive outcomes in HIV-affected children aged 2–5 years in Uganda. AIDS Care 2016;28:76–83. 10.1080/09540121.2016.1147010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Govender K, Reardon C, Quinlan T et al. Children’s psychosocial wellbeing in the context of HIV/AIDS and poverty: a comparative investigation of orphaned and non-orphaned children living in South Africa. BMC Public Health 2014;14:615. 10.1186/1471-2458-14-615 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Atanuriba GA, Apiribu F, Boamah Mensah AB et al. Caregivers’ experiences with caring for a child living with HIV/AIDS: a qualitative study in Northern Ghana. Glob Pediatr Health 2021;8:2333794X211003622. 10.1177/2333794X211030186 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Van Deventer C, Wright A. The psychosocial impact of caregiving on the family caregivers of chronically ill AIDS and/or HIV patients in home-based care: a qualitative study in Zimbabwe. South Afr J HIV Med 2017;18:718. 10.4102/sajhivmed.v18i1.718 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Slogrove AL, Cotton MF. Special issues and management of HIV exposed uninfected infants and children. Pediatr Oncall J 2016;13:31–5. 10.7199/ped.oncall.2016.27 [DOI] [Google Scholar]
- 37. Brennan AT, Bonawitz R, Gill CJ et al. A meta-analysis assessing all-cause mortality in HIV exposed uninfected compared with HIV-unexposed uninfected infants and children. AIDS 2016;30:2351–60. 10.1097/QAD.0000000000001216 [DOI] [PubMed] [Google Scholar]
- 38. Oyeyimika OA, Muheezat OO, Adenike OF et al. Caregiver burden, quality of life and coping mechanisms among caregivers of children living with HIV in Lagos, Nigeria. Afr J Health Sci 2020;33:68–77. 10.4314/ajhs.v33i5 [DOI] [Google Scholar]
- 39. Alizadeh M, Chehrzad MM, Mirzaee M, et al. Caregiver burden and related factors in parents of children with thalassemia. J Adv Pharm Educ Res. 2019;9:S2. 10.51847/0Zc0i6jB5k [DOI] [Google Scholar]
- 40. Rahmani F, Ranjbar F, Hosseinzadeh M et al. Coping strategies of family caregivers of patients with schizophrenia in Iran: a cross-sectional survey. Int J Nurs Sci 2019;6:148–53. 10.1016/j.ijnss.2019.03.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Slemming W, Biersteker L, Lake L (eds). South African Child Gauge 2024: Enhancing Early Childhood Development. Cape Town: University of Cape Town; Children’s Institute, 2024. [Google Scholar]
- 42. Begnel ER, Ojee E, Adhiambo J, et al. The Linda Kizazi study: a comparison of morbidity and mortality from birth to 2 years between children who are HIV-unexposed and HIV-exposed, uninfected in the era of universal antiretroviral therapy. BMJ Global Health. 2025;10:e015841. 10.1136/bmjgh-2024-015841 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Toledano-Toledano F, Domínguez-Guedea MT. Psychosocial factors related with caregiver burden among families of children with chronic conditions. Biopsychosoc Med 2019;13:6. 10.1186/s13030-019-0147-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Casale M, Wild L. The relationship between social support and the health of HIV-positive caregivers of children: a review of the empirical literature. Vulnerable Child Youth Stud 2012;7:260–82. 10.1080/17450128.2012.681801 [DOI] [Google Scholar]
- 45. Wynn A, Rotheram-Borus MJ, Leibowitz AA et al. Mentor mothers program improved child health outcomes at a relatively low cost in South Africa. Health Aff 2017;36:1947–55. 10.1377/hlthaff.2017.0239 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Nelson A, Lebelo K, Cassidy T et al. Postnatal clubs: implementation of a differentiated and integrated model of care for mothers living with HIV and their HIV exposed uninfected babies in Cape Town, South Africa. PLoS One 2023;18:e0286906. 10.1371/journal.pone.0286906 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Chersich MF, Blaauw D, Scorgie F et al. Safeguarding maternal and child health in South Africa by starting the child support grant before birth: design lessons from pregnancy support programmes in 27 countries. S Afr Med J 2016;106:1192–210. 10.7196/SAMJ.2016.v106i12.12134 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.