Caregiver and child factors predicting HIV status disclosure among children enrolled on ART: a cross-regional study in Addis Ababa and Oromia, Ethiopia

ABSTRACT

Introduction: The World Health Organization recommends that for children of school age who are HIV-positive, their HIV status should be disclosed. The purpose of this study was to get a better understanding of disclosure rates among school-aged children in Ethiopia and to investigate the variables amongst caregivers and children predicting disclosure/non-disclosure. Methods: Data from 231 children were collected prospectively through a structured questionnaire after initiation of ART treatment. Data were analysed with χ² test for comparison and logistic regression to verify the prediction of independent variables with the child’s HIV status disclosure. Results: Child age, caregiver marital status and caregiver residence were significantly associated with disclosure at 6 months. Similarly, child age, caregiver marital status and caregiver–child relationship were found to be significantly associated with disclosure at 12 months. Child age greater than nine years at 6 months (aOR 8.5, 95% CI: 2.5–29) and at 12 months (aOR 5.3, 95% CI: 1.8–16) were found to be significantly associated with disclosure. Furthermore, at 6 months, children with caregivers who had a partner they were not married to were more likely to disclose the HIV status to the child. This study suggests that paediatrics HIV care and treatment should consider these issues through contextualized strategies on child HIV disclosure and related challenges. Large-scale studies are required to better generalize on the subject.

KEYWORDS:

• Children
• disclosure
• HIV

Introduction

Background

Ethiopia is located in the northeast of Africa and is the second most populated country on the continent with 112 million people (Ethiopia: Overview, 2019; World Population Prospects, 2019). The country has a fast-growing economy but is still considered very poor, with a per capita income of 790 US dollars (Ethiopia: Overview, 2019). In 2018, 690,000 people in Ethiopia were living with HIV. The incidence of HIV per 1000 people over one year was 0.24 and the prevalence was 1%, with around 11,000 people dying of AIDS-related diseases in the same year (Ethiopia Overview, 2018). In 2018 in Ethiopia, 75% of adults living with HIV knew of their status, with 65% receiving Antiretroviral Therapy (ART). In the same year, 59% of children aged 0–14 years were receiving ART in Ethiopia (Ethiopia Overview, 2018).

In Ethiopia, there were an estimated 62,000 children living with HIV in 2016, of whom about 3800 were children who had been newly infected, and an estimated 2800 died due to HIV (Ethiopia HIV Country Profile, 2016). A study from 2018 where 200 caregivers of school-aged children (6–14 years) with HIV in Ethiopia were interviewed about whether they had disclosed their child’s HIV status to the infected child, showed that only 28% had disclosed the status to the child (Lencha et al., 2018).

Stigma has been identified as one of the major barriers that might prevent a caregiver from telling a child about their HIV status. This stigma may take many forms. It might be internal, referring to a lowered perception of oneself, or received, referring to other people’s changed perceptions of a person because of their HIV status (Krauss et al., 2011). Other reasons not to tell a child include fear that the child might tell others about their status and that there might be other negative consequences for the child (Mahloko & Madiba, 2012). Some caregivers also feel uncertain about how to tell the child (Hayfron-Benjamin et al., 2018), and the process is important, as it has been shown that inadvertent disclosure of HIV status to a child, when done without support, may negatively affect the child’s well-being (Krauss et al., 2011).

Despite these potential risks accompanying disclosure, the World Health Organization (WHO) recommends that children of school age, defined as between 6 and 12 years of age, of normal cognitive and emotional maturity, are informed about their HIV status. Disclosure is a process and should preferably be done in small steps, matching the child’s age and maturity (Krauss et al., 2011). Previous studies have shown that the best person to disclose a child is the caregiver with the support of health-care-providers (Doat et al., 2019). There is a reduced risk of death after disclosure, better adherence to treatment, and little evidence of emotional harm associated with a well-executed disclosure (Krauss et al., 2011). Due to these factors, it must be considered being in the child’s best interest to be disclosed about his/her status.

Aim of the study

To describe characteristics of caregivers with disclosed/undisclosed children 6 and 12 months after the child’s ART-treatment initiation. The further aim was to describe some characteristics of the children.

Specific research questions

• – How many children were informed about their HIV status 6 months/12 months after treatment initiation?

• – Do caregiver sex, age, marital status, education, residence or relationship to child affect status disclosure and if so, how?

• – Do child age, sex, WHO clinical stage, BAZ score, CD4+ count affect status disclosure, and if so, how?

Method

As part of a longitudinal study, quantitative data were collected prospectively from a cohort of children between the ages of 6 and 14 and their caregivers between 20 December 2014 and 20 April 2015 (Shargie, 2017). Data regarding disclosure 6 and 12 months after ART initiation was analysed.

Study setting

Eight different health facilities in the regions of Addis Ababa, Adama, Bishoftu, and Mojo participated in the study. In these health facilities, patients from both urban and rural areas receive comprehensive ART. The Bishoftu, Mojo and Adama regions are all located in the Oromia region, including the main highway between the city of Addis Ababa and the city of Dijibouti. The highway is the main route for export and import to the capital, and the risk of HIV transmission is considered high in these cities due to all the travellers coming through. The health facilities were chosen based on high numbers of adults and children enrolled on ART.

Study participants

The inclusion of caregivers of children diagnosed with HIV and enrolled on ART was done in multiple steps. Providers of ART to children diagnosed with HIV and enrolled on ART completed a screening log to ensure their eligibility to participate in the study. This was confirmed when the caregiver was added to an enrolment log, and their informed consent/refusal was taken and documented. When the caregivers agreed to participate they were given a SPID (Specific Participant Identification Code) that was registered in the recruitment log. The screening log, enrolment log and the chart review template were continuously reviewed by the principal investigator.

Inclusion criteria included: the family caregiver over age 18, child age between 6 years and 14 years, and the child diagnosed with HIV and enrolled on ART for one month or less. Children enrolled on ART for more than a month and children who had other severe illnesses or were terminally ill were excluded.

The only exception made was if a person under 18 was considered the head of the family, then they were also eligible to participate as caregivers. This resulted in two caregivers in this study being 17 years old. In total, 309 children and their caregivers from the eight health facilities were eligible to participate in the study. One hundred and thirty-one of the caregivers were biological mothers and 30 were biological fathers. Seventy children had caregivers who were not their biological mother or father (see Table 1).

Table 1. Caregiver relationship to child at 6 months.

	Frequency (n)	Percent
Biological father	30	13.0
Biological mother	131	56.7
Brother	4	1.7
Grandfather/grandmother	14	6.1
Legal guardian	4	1.7
Other relative	38	16.5
Sister	10	4.3
Total	231	100.0

After the child had been diagnosed and enrolled on ART, the caregiver of the child was approached by the ART provider. The caregiver was informed about the study and asked for consent. Three caregivers refused to participate in the study. Three hundred and six caregivers with 306 children with HIV were included. Because two caregivers received the same SPID code, one male caregiver (and child) and one female caregiver (and child), they were excluded from the study, giving 304 caregivers and their children who were followed for 12 months.

For this article, only children between the ages of 6 and 14 years were included, resulting in 231 children who were subsequently followed for 12 months.

Ethics

Ethical approvals were obtained from the Swedish Ethics Board (Ref. No. 2013/85), Institutional Review Board of Addis Ababa University (Protocol no. 045/13/sph), Oromia Regional Health Bureau (Ref. no. HIV-AIDS/sup-10/379), Addis Ababa City Administration Health Bureau (Ref. no AAH13/2084/227) and the National Research Ethical Review Committee in Ethiopia (Ref. no. 3.10/322/05).

The caregivers were provided with oral and written information about the study process and the use of the data. For individuals who could not read or write, the information and consent statement was read aloud, and consent was obtained with their fingerprints. The ART clinical focal staff member was asked to add his/her signature as an impartial witness.

Each time data was collected the caregiver and the child were offered fruit, drinks and compensation for their travel.

Data collection

ART providers and data clerks collected the data from the children’s medical records using a chart review template after receiving training and support and supervision twice a month from the principal investigator. Only caregivers were involved in responding to the interview questions.

A pilot of the chart review template used for data collection was performed at two sites. The template was then slightly modified to better suit the purpose of the study.

According to the national guidelines, appointments for the child’s follow-up (National Guidelines For Comprehensive HIV Prevention, Care and Treatment, 2014) were held once every month, and data were collected at 6 and 12 months through the questionnaire.

Variables included in the questionnaire were adherence to treatment, disclosure, socioeconomic and socio-demographic status, baseline and continued CD4+ cell count, baseline WHO clinical HIV staging as well as anthropometric measurements (BMI adjusted for age and sex).

Data analysis

Descriptive statistics were used to analyse the data regarding caregiver/child characteristics. The explanatory variables used were socio-demographic for caregivers (age, sex, residence, education, marital status and biological/non-biological relationship to child) and clinical markers for the child; haemoglobin measured in g/dl, WHO clinical staging (measured at baseline), CD4+ count measured in cells/mm³, BMI adjusted for sex and age of the child (z-score) as well as adherence dichotomized to good (>95% adherent) vs poor (≤95% adherent). The outcomes at the end of 6 and 12 months were categorized as disclosed or not disclosed.

Statistical analysis was undertaken using the Stata program version 14. Variables found to have a statistically significant association (χ² test) with disclosure were analysed using logistic regression. Statistical significance was described as 95% confidence interval (CI) and statistical significance was set at p ≤ 0.05.

Findings

Description of data collected about caregivers after 6 months

After 6 months, there was a majority of female caregivers (82%), most of whom (40%) were between 35 and 44 years of age. Almost half (42%) were married, lived in urban areas (94%), and had completed grades 1 to 8 (32%) (see Table 2). All biological parents of the children were HIV-positive as well. Thirty-two of the children had no living biological parents and 109 children had two living parents (see Table 3).

Table 2. Socio-demographic characteristics of caregivers at 6 and 12 months after child ART enrolment.

Caregivers characteristics	At 6 months		At 12 months
	N	%	N	%
Caregivers
Sex
Female	189	82	188	82
Male	42	18	41	18
Age (years)
17	2	1	–^a	–^a
18-24	15	6	11	5
25–34	77	33	78	34
35–44	93	40	95	41
>44	44	19	45	20
Marital status
Married	96	42	99	43
Divorcer/separated	35	15	33	14
Never married partner	37	16	37	16
Widower/widow	49	21	46	20
Cohabited	14	6	14	6
Residence
Rural	15	6	16	7
Urban	216	94	213	93
Education
College/university completed	26	11	26	11
High school completed	65	28	64	28
Grade 1–8 completed	74	32	78	34
Able to read and write	27	12	27	12
Illiterate	39	17	37	16

^a At 12 months, all caregivers were aged 18 years and above.

Table 3. Vital status of biological parents of children at 6 months.

	Frequency (n)	Percent (%)
Both mother and father died	32	13.9
Both father and mother alive	109	47.2
Father died	45	19.5
Mother died	45	19.5
Total	231	100.0

Description of data collected about children after 6 months

Most children were male (52%), aged nine or above (77%), were in WHO status I or II (76%), had haemoglobin levels ≥ 10 (97%) and CD4 + count ≥ 200 (97%). A majority also had good adherence to treatment (96%) and a biological relationship to their caregiver (82%).

Description of data collected about adherence to study after 6 months

Regarding adherence at 6 months, 13 of the participants were registered as non-response (NR), making the total number of answers 213 (see Table 4). Data regarding the sex of one child were missing. At 6 months three children were marked as outliers in the BAZ score, and their data are not described in the table or further measured. Two caregivers (one male one female) with their children were marked as lost to follow-up (LTFU) between the data collections at 6 and 12 months, resulting in 229 caregivers/children being followed up after 6 months.

Table 4. Descriptive data of children at 6 months and 12 months after child ART enrolment.

Characteristics of child on ART	At 6 months		At 12 months
	N	%	n	%
Sex
Female	112	48	112	49
Male	118	52	117	51
Age (years)
6–8	54	23	54	24
≥9	177	77	175	76
WHO status
I–II	176	76	175	76
III–IV	55	24	54	24
Haemoglobin level (g/dl)
<10	6	3	15	7
≥10	225	97	178	93
CD4+ count (cells/mm^3)
<200	8	3	47	21
≥200	223	97	179	78
BAZ score
z-score < −2	31	14	31	14
2 > z-score > −2	197	86	195	86
Caregiver–child relationship
Biological family	189	82	187	82
Others	42	18	42	18
Adherence
Good	208	98	220	96
Poor	5	2	9	4
Disclosure of the child
Disclosed	66	29	97	42
Not disclosed	165	71	132	58

Description of data collected about caregivers and children after 12 months

At 12 months, three children were still marked as outliers in the BAZ score. Data were missing for 36 children regarding haemoglobin levels and for three children regarding CD4+ count. No biological parents had died at this time compared to the data collection at 6 months (see Table 5).

Table 5. Vital status of biological parents at 12 months.

	Frequency (n)	Percent
Both parents alive	109	47.6
Both parents died	32	14.0
Father died	44	19.2
Mother died	44	19.2
Total	229	100.0

Description of data collected about disclosure after 6 months and 12 months

After 6 months, 66 of the children (29%) were informed about their HIV status, 165 children (71%) had not been informed. After 12 months, another 31 children were told, making the overall number of informed children at 12 months 97 (42%). One hundred and thirty-two children (58%) had still not been told at that time.

Variables associated with disclosure at 6 months

Children aged nine years and above were informed significantly more (63 children) than children between 6 and 8 years of age (3 children) (p < 0.001). The marital status of the caregiver and caregiver residence (urban/rural) was also associated with disclosure (p = 0.05 respectively).

Variables associated with disclosure at 12 months

At 12 months, a statistically significant association was found between the marital status of the caregiver and disclosure, p = 0.044, child age p = 0.005, where children ≥9 years were informed at a higher rate (83 children) than younger children (14 children). There was also a statistically significant association between caregiver–child relationship and disclosure, with p = 0.05. Children with a non-biological caregiver were informed at a higher rate than those whose caregivers were biological family. Caregiver residence, which showed a slight statistical association at 6 months (p = 0.052), still showed some association with disclosure at 12 months (p = 0.051) of ART follow-up.

Findings using bivariate and multivariable analysis

At both 6 and 12 months, children aged ≥9 years and caregivers who had a partner they were not married to were independently associated with disclosure in bivariate analysis (see Table 6). However, in adjusted multivariate analysis, only children aged ≥9 years (aOR 8.5, 95% CI: 2.5–29) and (aOR 5.3, 95% CI: 1.8–16) were found to be significantly associated with disclosure at 6 and 12 months respectively (see Tables 7 and 8).

Table 6. Differences in child disclosure at 6 and 12 months after child ART enrolment.

Characteristics of the participants	At 6 months			At 12 months
	Not disclosed n (%)	Disclosed n (%)	P-value	Not disclosed n (%)	Disclosed n (%)	P-value
Child age (years)			<0.001			0.005
6–8	51 (94)	3 (6)		40 (74)	14 (26)
≥ 9	114 (64)	63 (36)		92 (53)	83 (47)
Caregiver marital status			0.05			0.04
Cohabited	13 (93)	1 (7)		10 (71)	4 (29)
Currently married	67 (70)	29 (30)		54 (54)	45 (45)
Divorced/separated	29 (83)	6 (17)		26 (79)	7 (21)
Never married partner	21 (57)	16 (43)		17 (46)	20 (54)
Widowed/Widower	35 (71)	14 (29)		25 (54)	21 (46)
Caregiver residence			0.05			0.05
Rural	14 (93)	1 (7)		106 (61)	67 (39)
Urban	151 (70)	65 (30)		26 (46)	30 (54)

Table 7. Predictors of child disclosure at 6 months after child ART enrolment.

Characteristics of the participants	Child not disclosed n (%)	Child disclosed n (%)	cOR (95% CI)	P-value	aOR (95% CI)	p-value
Child age
6–8 years	51 (94)	3 (6)	Reference		Reference
≥9 years	114 (64)	63 (36)	9.4 (1.8, 31)	<0.001	8.5 (2.5, 29)	0.001
Caregiver marital status
Cohabited	13 (93)	1 (7)	Reference		Reference
Married	67 (70)	29 (30)	5.6 (0.70, 45)	0.10	4.4 (0.52, 37)	0.17
Divorced/separated	29 (83)	6 (17)	2.7 (0.29, 25)	0.38	1.9 (0.20, 18)	0.58
Never married partner	21 (57)	16 (43)	9.9 (1.2, 84)	0.035	7.3 (0.82, 65)	0.076
Widower/widowed	35 (71)	14 (29)	5.2 (0.62, 44)	0.13	4.4 (0.50, 39)	0.18
Caregiver residence
Rural	14 (93)	1 (7)	Reference		Reference
Urban	151 (70)	65 (30)	6.0 (0.78, 47)	0.086	4.2 (0.50, 35)	0.19

Table 8. Predictors of child disclosure at 12 months after child ART enrolment.

Characteristics of the participants	Child not disclosed n (%)	Child disclosed n (%)	cOR (95% CI)	P-value	aOR (95% CI)	P-value
Child age (years)
6–8	40 (74)	14 (26)	Reference		Reference
≥9	92 (53)	83 (47)	2.6 (1.3, 5.1)	0.006	5.3 (1.8, 16)	0.003
Caregiver marital status
Married	54 (55)	45 (45)	Reference		Reference
Divorced/separated	26 (79)	7 (21)	0.32 (0.13, 0.81)	0.016	0.36 (0.11, 1.1)	0.08
Never married partner	17 (46)	20 (54)	1.4 (0.66, 3.0)	0.37	0.83 (0.34, 2.03)	0.69
Widower/widowed	25 (54)	21 (46)	1.0 (0.50, 2.0)	0.98	1.2 (0.55, 2.8)	0.59
Cohabited	10 (71)	4 (29)	0.48 (0.14–1.6)	0.24	0.59 (0.12, 3.0)	0.52

Caregivers who were divorced/separated were found to inform their children to a lesser extent than caregivers who were married at 12 months.

Discussion

The findings showed that of the total 42% of children were informed at 12 months, 29% of them were informed during the first 6 months after ART initiation. The fact that more children were informed during the first 6 months might have to do with that being when the children’s treatment regime first started. This could have resulted in questions from the children regarding why they went to the hospital so many times as well as why they had to take medication, as suggested in earlier studies (Mahloko & Madiba, 2012).

Previous studies have shown that one major reason why caregivers choose not to tell their children is that they feel the child is too young to know about its status (Lencha et al., 2018). The older the children are, the greater is also the likelihood they have been informed. Previous studies have also come to this conclusion (Negese et al., 2012). The risk of the child being informed by mistake is always present, and the older they get the more situations with inadvertent disclosure they will have been exposed to.

One hundred and thirty-two of the children (58%) had still not been informed after 12 months but this does not mean that their caregivers had not started to talk to them about HIV or had started preparing to disclose the child’s status. As earlier mentioned, disclosure is a process. Research suggests that factors for non-disclosure might be that the caregivers are afraid that the child will tell other people about his/her status, and a fear that the child might be rejected socially (Mahloko & Madiba, 2012). Previous studies have also shown that some caregivers feel that the child cannot handle the psychological impact of the diagnosis (Aderomilehin et al., 2016). It has also been shown that some caregivers have chosen not to inform the child as they do not feel confident about the disclosure process, and needs guidance (Hayfron-Benjamin et al., 2018).

Our result showed that the child was less likely to be informed if the family caregiver was divorced/separated than if married. Being a single caregiver might include less social support. Good social support has previously been proven to be positively associated with disclosure (Lencha et al., 2018). This suggests that caregivers who are divorced/separated may need additional support from the health-care system and society.

Regarding adherence, our findings showed that there were no significant associations between disclosure and adherence even though previous research has shown an association. However, it seems as if this is linked to social support, which leads to both disclosure and adherence, rather than only disclosure resulting in better adherence (Krauss et al., 2011).

After 6 months, there was a statistically significant association between caregiver residence and disclosure in the present study. Even though not statistically significant, the likelihood of the child’s status being disclosed increased if the caregiver residence was urban.

The present study indicated that a greater proportion of the children and caregivers were from urban areas, with only 16 from rural areas. This might have added to selection bias since the results presumably can mainly be applied to an urban population. Also, the vast majority of the family caregivers were female. However, it is uncertain to what extent this adds to selection bias since most of the caregivers of children with HIV are their mothers (Shargie, 2017).

Of the 231 who were first enrolled in the study, two were marked as LTFU after 12 months, the sample size was therefore considered intact. As the participants in the study were able to fill out the questionnaire when they went for their children’s monthly check-up, non-response bias was minimized. Recall bias was unlikely in this study as it was conducted prospectively, with data collected monthly.

Conclusions

In conclusion, the results of this study can contribute to the knowledge of disclosure rates among children with HIV in Ethiopia and help predict variables for non-disclosure and disclosure. Child age, caregiver marital status and residence all seem to have an association with disclosure. However, more research is needed to be able to generalize the results. Research regarding HIV disclosure from other countries to compare with would also be of interest, as stigma and other social factors may differ between countries.

This study along with other similar studies helps to show the need for an understanding of the difficulty and complexity of disclosure. Caregivers of children with HIV need support and preferably clear guidelines on how to best disclose their children’s status to them. Giving this support to the caregivers will further enable them in supporting their children in return.

Acknowledgements

We would like to thank Tsegay Hailu for his support with the data analyses.

Disclosure statement

No potential conflict of interest was reported by the author(s).