A comparative analysis of social support, risk behaviors, and HIV health service use among adolescent and young males and females in Lusaka, Zambia

Abstract

Background

Human immunodeficiency virus (HIV) remains the leading cause of death in Zambia. While females are disproportionately affected by HIV, males – especially young males – are vulnerable to the disease due to a variety of risk factors. This study aimed to understand what, if any, sex-related differences exist between young females and males on social support, risk behavior, and HIV healthcare utilization issues.

Methods

Baseline survey responses from an implementation trial (NCT03995953) were examined for 863 females and 302 males affected by HIV between ages 15 and 26. We created summary statistics related to peer and familial support, risk factors (i.e., physical safety, economic security, mental health, substance abuse, and sexual behavior), and HIV healthcare utilization. Summary statistics were evaluated for statistical significance through Pearson Chi-Square testing.

Findings

Females and males, regardless of HIV status, have higher average confidence in familial support (67 %) than peer support (40 %). Across HIV status, females and males had similar rates of physical safety risk. Regardless of HIV status, about half the participants reported worrying about running out of food. Substance abuse risk is higher among males; 15 % of males at risk of HIV and 7 % of males living with HIV report drug usage other than alcohol or marijuana compared to just 1 % of all females. Among individuals at risk of HIV, there are differences in rates of HIV testing by sex: 27.7 % among males vs. 6.7 % among females.

Interpretations

While there are some differences, the many similarities between young females and males suggest that joint interventions which incorporate familial support could be beneficial to address shared risk factors. These joint interventions can be supplemented with sex-specific interventions related to substance abuse for males and HIV testing for females.

Funding

Research reported in this publication was supported by the Eunice Kennedy Shriver National Institute Of Child Health & Human Development of the National Institutes of Health under Award Number UH3HD096908. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Introduction

With over 1.3 million people living with human immunodeficiency virus (HIV), Zambia is one of the countries most affected globally.⁽¹⁾ HIV remains the leading cause of death in the country.⁽¹⁾ While the devastating impact of HIV can be felt across communities, HIV disproportionately affects certain groups. Of the 1.3 million new infections in 2022, more than one-quarter occurred in adolescents and young adults (AYA) between the ages of 15 and 24.⁽²⁾ In addition, HIV disproportionately affects women, as they account for two-thirds of HIV cases.⁽³⁾ In a 2024 report, the Joint United Nations Programme on HIV/AIDS (UNAIDS) found that “the incidence of HIV among adolescent girls and young women aged 15 – 24 years was extraordinarily high in parts of sub-Saharan Africa.”⁽⁴⁾ However, a recent study has shown that, of new infections occurring globally, HIV is predominantly impacting young men.⁽⁵⁾ Adolescent boys and young men (ABYM) are especially vulnerable because their age and sex make them less likely to have access to testing, care, or antiretroviral therapy (ART).^(2,5) Given that AYA face unique challenges with respect to HIV testing and treatment, it is important to understand the differences and similarities male and female AYA experience in their social supports and risk behaviors to improve the efficacy of future HIV-related interventions.

There is limited actionable research examining the impact of familial and general peer support on the utilization of HIV healthcare services and health outcomes. Several studies have investigated the impact of similarly-aged counselors or mentors on an adolescent's utilization of HIV healthcare services and health outcomes, and these studies have found that such interventions have positive impacts on HIV testing, care linkages, ART adherence, and viral load suppression.⁽[6], [7], [8], [9]⁾ However, the current body of research has not explored the influence that an adolescent's own peer group has on his or her utilization of HIV healthcare services and health outcomes. The literature pertaining to the impact an adolescent's family has on HIV healthcare utilization and health outcomes is similarly sparse. One study explored the impacts of parent-child communication on sexual and pregnancy risks in Zambia. The researchers found that fear-based parenting and opposition to sexual education were negatively associated with open parent-child discussions about sexual health. This, in turn, was associated with an increase in sexual risk.⁽¹⁰⁾ There is limited information on the differences in the impact of peer and family support between male and female AYA.

Furthermore, despite there being a strong body of research on the impact of risk behavior on HIV transmission, differences between ABYM and adolescent girls and young women (AGYW) have rarely been examined. One study investigating the impacts of substance usage (i.e., alcohol or tobacco use) and suicidality on sexual behavior among college students in Zambia found that nearly three-quarters of all the participants engaging in sexual behavior were engaging in “risky” sexual behavior; risky behavior was more prevalent in men.⁽¹¹⁾ Risky sexual behavior was defined as engaging in unprotected sex, having more than two partners in a 12-month period, engaging in sex after alcohol or drug consumption, or having sex with strangers.⁽¹¹⁾ Students who were older and reported higher levels of suicidality and alcohol consumption were also more likely to engage in risky sexual behavior.⁽¹¹⁾ Understanding how mental health and substance use affect sexual behavior, and specifically how those risk behaviors differ by sex, can improve interventions meant to promote safe sex practices, and therefore reduce HIV transmission through sexual intercourse.

Compared to risk behavior, the impacts of education and economic status on HIV transmission are better understood but can be inconsistent across populations. For example, we understand that higher levels of education and higher economic status are protective factors against risky sexual behavior for males and females.⁽¹²⁾ However, evidence exists that females have lower primary education completion rates and secondary school enrollment rates.⁽¹³⁾ These data suggest that females may not be benefitting from the protective effects of education against HIV to the same degree as their male counterparts. The literature also demonstrates sex-based differences in employment, with more females engaged in unpaid care work or in-kind employment.⁽¹⁴⁾ This could encourage females to engage in transactional sex, which has been shown to be associated with HIV transmission for women.⁽¹⁵⁾ It is also understood that older adolescence is associated with a reduction in risky sexual behavior.⁽¹²⁾ A reason for this reduction could be that HIV can delay puberty in perinatally infected youth.⁽¹⁶⁾ Delayed puberty can elicit feelings of shame and frustration for AYA, especially for AGYW, which, in turn, can encourage risky sexual behavior in order for AYA to “prove” themselves to their peers.^(12,16)

This analysis aims to better understand differences and similarities between ABYM and AGYW regarding social support, risk factors and behaviors (i.e., physical safety, economic stability, mental health, substance abuse, and sexual behavior), and HIV healthcare utilization (i.e., HIV testing and ART adherence). We utilized baseline demographic data as well as measures of social support, risk factors and behaviors, and healthcare utilization. A comparison of ABYM and AGYW can reveal important information on how to improve the delivery of future HIV-related interventions. Areas of commonality between the two sexes can be leveraged to build connections between ABYM and AGYW while targeted interventions can support areas of divergence.

Methods

We utilized data from the baseline assessment of the Support for HIV Integrated Education, Linkages to Care, and Destigmatization (SHIELD) intervention to perform our analysis. SHIELD was designed to educate and empower HIV-affected AYA ages 10–26. The intervention was tested in a cluster-randomized controlled trial implemented in six communities and clinics in Lusaka, Zambia reflecting urban and peri-urban populations. For the purposes of this analysis, we restricted the ages of both ABYM and AGYW to 15–26. This study was approved by institutional review boards (IRBs) at RTI International, Population Council, and ERES Converge and has been registered on clinicaltrials.gov under the clinical trial number NCT03995953. While a full description of the methodological approach was published,⁽¹⁷⁾ a brief summary of the study's methods is outlined below.

Study overview

SHIELD recruited female participants aged 10–26 and male participants aged 15–26 who resided in the non-contiguous catchment areas of six clinics in the greater Lusaka area. Study participants included those living with HIV and those who reported their HIV status as negative or unknown. Female participants could not be pregnant at enrollment. For those living with HIV, recruitment was conducted in a manner to maintain confidentiality. Healthcare providers and community outreach staff at selected clinics were asked to identify and approach eligible AYA to ask for their consent to be contacted by study researchers. Utilizing community outreach staff allowed this study to identify and approach AYA living with HIV who were not actively receiving treatment. For those living without HIV, recruitment was conducted by mapping out the six catchment areas and utilizing peer navigators to visit households to recruit eligible AYA. While HIV status was not serologically confirmed for participants not living with HIV at the start of the study, all participants were tested at the end of the study. For AYA below the age of 18, we obtained both written parental consent and assent from the minors. AYA above the age of 18 provided written informed consent in their preferred language.

If study participants met the enrollment criteria, they were asked to complete a baseline survey. For AGYW between the ages of 15 and 24, this survey was conducted between May 2021 and September 2021. For ABYM between the ages of 15 and 24, this survey was conducted between December 2021 and January 2022. Data collectors were trained to sensitively administer the baseline survey on a tablet using REDCap's mobile app. Participants had the option to self-administer portions of the survey focused on sexual behavior. Because engaging in sexual behavior with individuals younger than 16 years is illegal in Zambia, questions related to sexual behavior were only asked to participants who were aged 16 or older.

Randomization and masking

SHIELD was randomized at the community level. There was no masking present in SHIELD.

Data analysis

Many of the questions used in the baseline survey for the SHIELD intervention are derived from harmonized measures created by the Prevention and Treatment through a Comprehensive Care Continuum for HIV-affected Adolescents in Resource Constrained Settings (PATC³H) consortium.⁽¹⁸⁾ This consortium spent two years creating validated measures for future research related to AYA living with HIV (AYAHIV). The result is a set of harmonized instruments covering three domains: the AYAHIV prevention continuum, the AYAHIV care continuum, and structural and cross-cutting topics related to AYAHIV.⁽¹⁸⁾ Questions related to demographics, social support, economic stability, sexual behavior, mental health, substance abuse, HIV testing, and ART adherence were all a part of the PATC³H harmonization and were utilized in the baseline survey.⁽¹⁸⁾ We created the summary statistics presented in this analysis from these questions.

We analyzed responses to the baseline survey for 302 ABYM and 863 AGYW. We divided our sample into four groups: (1) ABYM not living with HIV (NLHIV), (2) AGYW NLHIV, (3) ABYM living with HIV (LHIV), and (4) AGYW LHIV. We utilized demographic data as well as responses related to social support, risk factors and behaviors, and healthcare utilization to create our summary statistics.

Summary statistics

For the purposes of this analysis, social support was split into two categories: familial support and peer support. The SHIELD baseline survey contained a “social support” section that included questions from the Multidimensional Scale of Perceived Social Support.⁽¹⁹⁾ This validated measure consists of 12 questions asking about family, friends, and other “special” individuals from whom the respondent derives support.⁽¹⁹⁾ We selected comparable survey questions that explicitly mentioned “family” to create the familial support summary statistic and questions that explicitly mentioned “friend(s)” to create the peer support summary statistic. Table 1 presents the questions selected to compare familial and peer support. For all questions, respondents selected from a range of five response options from “strongly agree” to “strongly disagree” as shown in Table 1. The social support summary statistics were created condensing the five-point Likert scale into a binary scoring system (i.e., has support vs. does not have support). The proportion of respondents who “strongly agreed” with the survey questions presented below was considered to have social support, and all other responses were considered to not have social support. We did not include those who responded “moderately agree” in the “has social support” category because this answer choice could suggest that the individual did not receive all the support they needed.

Table 1.

Social support summary statistics.

Summary Statistic Category & Survey Questions	Responses options
Familial Support
My family really tries to help me. I get the emotional help and support I need from my family. I can talk about my problems with my family	Strongly agree Moderately agree Neutral Moderately disagree Strongly disagree
Peer Support
My friends really try to help me. I can count on my friends when things go wrong. I have friends with whom I can share my joys and sorrows.

We split risk factors and behaviors into five categories for our assessment: physical safety risk, economic stability risk, mental health risk, substance abuse risk, and sexual behavior risk. A full list of survey questions and their responses that were used to create these summary statistics can be found in Table 2. The physical safety and economic stability summary statistics were created by analyzing the proportion of participants who responded “yes” to the questions presented in Table 2. The mental health risk summary statistics looked at the proportion of participants who selected either “more than half the days” or “nearly every day.” The substance abuse risk summary statistics examined the proportion of respondents who drank four or more times a week and the proportion of respondents who had ever used drugs other than alcohol or marijuana for recreational purposes. Binge drinking, as defined by the World Health Organization, is the consumption of four alcoholic drinks for women and five alcoholic drinks for men on three or more days in a week.⁽²⁰⁾ We simplified this definition for our analysis. Finally, the sexual behavior risk summary statistics looked at the proportion of respondents who used a condom less than half the time with a casual partner in the past three months and the proportion of respondents who had never used a contraceptive method. This was the definition we utilized to measure a low level of condom usage.

Table 2.

Risk factor and behavior measures.

Measures & Survey Questions	Response Options
Physical Safety Risk
In the past 6 months has anyone punched, slapped, kicked, bit, or caused you any type of physical harm? In the past 6 months, has anyone insulted, ignored, humiliated, yelled at, or made you feel ashamed or bad about yourself? In the past 6 months, has anyone made you feel afraid, unsafe, or in danger?	Yes/No
Economic Stability Risk
In the past 3 months, were you worried you would run out of food before you or someone in your household could get money to buy more? In the past 3 months, did you worry about shelter or having a stable place to live? In the past 3 months, did you worry about losing work, whether a formal job or other work? In the past 3 months, have you searched for work but been unable to find any?	Yes/No
Mental Health Risk
In the past 2 weeks, how often have you been bothered by feeling nervous, anxious, or on edge? In the past 2 weeks, how often have you been bothered by not being able to stop or control worrying? In the past 2 weeks, how often have you been bothered by little interest or pleasure in doing things? In the past 2 weeks, how often have you been bothered by feeling down, depressed, or hopeless?	Not at all Several days More than half the days Nearly every day
Substance Abuse Risk
How often do you have a drink containing alcohol?	Never Monthly 2 to 4 times a month 2 to 3 times a week 4 or more times a week
Have you ever used drugs other than alcohol or marijuana (not for medical/treatment purposes)?	Yes/No
Sexual Behavior Risk
How often in the last 3 months did you use condoms with casual partners during sex?	Never Less than half the time About half the time More than half the time Always
Have you ever used a contraceptive method?	Yes/No

The final category that we examined was HIV healthcare utilization. For this, we examined questions related to HIV testing and ART adherence. For those living without HIV, we selected a question measuring the rate of HIV testing within the past six months. For those living with HIV, we looked at questions measuring the proportion of participants who were on ART, whether they experienced any side effects, and how regularly they adhered to treatment.

When performing the analysis, participants who declined to answer a question were removed from the denominator. Participants who were not asked questions due to the survey's skip patterns were also removed from the denominator. For example, some of the economic stability questions related to employment were only asked to those who had indicated that they were currently working. Questions related to sexual behavior risk were only asked to participants 16 years and older who indicated that they were sexually active. We conducted statistical testing to assess differences between ABYM and AGYW cohorts by HIV status. We performed the analysis of the data in Stata. Pearson Chi-Square tests were used for significance testing between sex and categorical/ordinal variables. Statistical significance was defined at p ≤ 0.050.

Role of the funding source

The research reported in this manuscript was supported by the Eunice Kennedy Scriver National Institute of Child Health and Human Development (NICHD) of the National Institutes of Health (NIH). The NIH were not involved in any aspect related to study design, data collection and interpretation, analysis, the writing of this manuscript, or where to submit this manuscript for submission. The content expressed in this manuscript is the sole responsibility of the authors and does not necessarily represent the official views of the NIH.

Results

Demographics

Among the NLHIV cohort, we observed differences in age distribution among AGYW, with 71.2 % of this cohort aged 15–18 compared to only 22.0 % of ABYM. The NLHIV cohort also exhibited differences in educational level; for example, a higher proportion of AGYW NLHIV (57.5 %) reported being in school at the time of the baseline survey compared to ABYM NLHIV (26.8 %). In both HIV cohorts, a higher proportion of ABYM were working (NLHIV 39.3 %, LHIV 44.1 %) than AGYW (NLHIV 10.9 %, LHIV 22.0 %). Additionally, more AGYW were married (NLHIV 5.9 %, LHIV 13.6 %) compared to their male counterparts (NLHIV 1.3 %, LHIV 2.0 %). A complete demographic summary can be found in Table 3.

Table 3.

Demographic characteristics of those not living with HIV and those living with HIV by sex (2022).

	NLHIV			LHIV
	ABYM	AGYW	p	ABYM	AGYW	p
	(n = 150) %	(n = 340) %		(n = 152) %	(n = 523) %
Age			<0.0001			0.2
15–18	22.0	71.2		25.0	20.7
19–22	50.7	28.8		52.6	49.3
23–26	27.3	0		22.4	30.0
Education Level			<0.0001			0.1
No schooling	2.0	1.2		0.7	1.3
Some primary school	14.7	33.1		13.2	20.5
Some secondary school	79.3	65.4		77.6	72.4
Some college or university	4.0	0.3		8.6	5.7
Currently in school			0.03			0.06
Yes	26.8	57.5		35.5	27.6
Currently Working			<0.0001			<0.0001
Yes	39.3	10.9		44.1	22.0
Relationship status			0.03			<0.0001
Single	98.7	94.1		98.0	86.4
Married	1.3	5.9		2.0	13.6

Abbreviations: NLHIV Not living with HIV, LHIV Living with HIV, ABYM Adolescent boys and young men, AGYW Adolescent girls and young women.

Social support

As shown in Table 4, a higher proportion of AGYW NLHIV strongly agreed that they received sufficient emotional help and support from family (65 %) compared to their male counterparts (51 %). Regarding peer support, more ABYM LHIV strongly agreed that they had friends with whom they could share their joys and sorrows (58 %) than AGYW LHIV (43 %).

Table 4.

Social support of those not living with HIV and those living with HIV by sex (2022).

	NLHIV			LHIV
	ABYM (n = 150)	AGYW (n = 340)	p	ABYM (n = 152)	AGYW (n = 523)	p
	n	n		n	n
	%	%		%	%
Familial support
My family really tries to help me.	103	254	0.2	105	355	0.8
	68.7	74.7		69.1	67.9
I get the emotional help and support I need.	77	221	0.004	92	316	1.0
	51.3	65.2		60.5	60.4
I can talk about my problems.	81	207	0.2	100	354	0.7
	54.4	60.9		65.8	67.8
Peer support
My friends really try to help me.	57	156	0.1	51	146	0.2
	38.3	46.0		33.6	27.9
I count on my friends when things go wrong.	51	135	0.2	46	140	0.4
	34.0	39.8		30.7	26.8
I have friends I can share my joys and sorrows.	80	207	0.1	87	225	0.002
	53.3	61.1		57.6	43.1

Abbreviations: NLHIV Not living with HIV, LHIV Living with HIV, ABYM Adolescent boys and young men, AGYW Adolescent girls and young women.

Risk factors and behaviors

Our cohort did exhibit risks to their physical safety and there were no differences in physical safety risk by sex. Our cohort also exhibited economic stress; for example, around half of all respondents had worried about running out of food in the past three months. For both those living with and without HIV, in general ABYM experienced greater economic stress than AGYW. A higher proportion of ABYM NLHIV worried about shelter (28.2 %) and an inability to find work (72 %). AGYW NLHIV experienced notably lower economic stress with only 19.1 % worried about shelter and 40.9 % worried about not finding work. The greater economic stress experienced by ABYM was evident in the LHIV cohort as well; 61.2 % of ABYM LHIV worried about losing work whereas only 36.5 % of AGYW LHIV had the same concern.

When examining mental health risk in the NLHIV cohort, we saw no sex-based differences. For those in the LHIV cohort, we observed a higher mental health burden among AGYW. AGYW LHIV experienced higher rates of nervous or anxious feelings (22.7 %) than their male counterparts (12.6 %). The proportion of AGYW LHIV who had little interest or pleasure in doing things (25.5 %) was double that of ABYM LHIV (12.5 %).

Substance abuse risk did exhibit a sex-based difference among both those living with and without HIV. A larger proportion of ABYM have used drugs other than alcohol or marijuana than AGYW. For those living without HIV, 14.7 % of ABYM had experimented with substances compared to 0.6 % of AGYW. For those living with HIV, these values are 6.7 % and 0.8 %, respectively.

Sexual behavior risk also demonstrated a sex-based difference that was present across disease status. For both those living with and without HIV, AGYW reported low levels condom usage in the past 3 months with a casual partner (58.9 % NHLIV and 52.8 % LHIV) at rates more than two times higher than ABYM (26.3 % NLHIV and 20.4 % LHIV). The complete summary of this analysis as it relates to risk factors and behaviors can be found in Table 5.

Table 5.

Risk factors and behaviors of those not living with HIV and those living with HIV by sex (2022).

	NLHIV			LHIV
	ABYM (n = 150)	AGYW (n = 340)	p	ABYM (n = 152)	AGYW (n = 523)	p
	n	n		n	n
	%	%		%	%
Physical Safety Risk In the past six months, has anyone…
caused you physical harm?	33	85	0.5	23	88	0.6
	22.0	25.1		15.2	16.8
made you feel insulted, ignored, or ashamed?	91	203	0.8	93	274	0.05
	61.5	59.7		61.2	52.5
made you feel afraid, unsafe, or in danger?	43	107	0.5	33	131	0.4
	28.7	31.5		21.7	25.1
Economic Stability Risk In the past 3 months, did you worry about…
running out of food?	80	157	0.2	78	283	0.5
	53.3	46.3		52.0	54.3
shelter or having a stable place to live?	42	65	0.03	46	151	0.7
	28.2	19.1		30.7	28.9
losing work?⁎	41	15	0.05	41	42	0.006
	69.5	40.5		61.2	36.5
being unable to find work?	54	79	<0.0001	44	205	0.7
	72.0	40.9		65.7	68.3
Mental Health Risk In the past two weeks, how often have you been bothered by…
feeling nervous, anxious, or on edge?	30	58	0.4	19	118	0.007
	20.0	17.2		12.6	22.7
not being able to stop or control worrying?	40	72	0.2	35	145	0.3
	26.7	20.9		23.1	27.9
little interest or pleasure in doing things?	30	76	0.6	19	133	0.0008
	20.0	22.5		12.5	25.5
feeling down, depressed, or hopeless?	32	73	1.0	29	130	0.1
	21.4	21.7		19.3	24.8
Substance Abuse Risk
How often do you have an alcoholic drink?
4 or more times a week	1	1	0.6	3	5	0.3
	0.7	0.3		2.0	1.0
Drug use other than alcohol or marijuana?
Yes	22	2	<0·0001	10	4	<0.0001
	14.7	0.6		6·7	0·8
Sexual Behavior Risk
Condom usage with casual partners in the past 3 months
Less than half the time	15	30	0.0005	9	65	0.0002
	26.3	58.9		20.4	52.8
Have you ever used a contraceptive method?
No	25	30	0.9	12	54	0.7
	21.7	21.3		12.6	14.4

^⁎The denominator for this measure only included individuals who had responded “yes” to “currently working” in the demographic section; Abbreviations: NLHIV Not living with HIV, LHIV Living with HIV, ABYM Adolescent boys and young men, AGYW Adolescent girls and young women.

HIV Healthcare Utilization.

Rates of HIV testing within the past six months were consistently higher among ABYM NLHIV, as shown in Table 6. For all ages, ABYM NLHIV tested for HIV at a rate over four times greater (27.7 %) than their female counterparts. Removing the younger age groups from this measure did not alter the observed difference in HIV testing rate between the sexes; for those aged 18 and older, 28.3 % of ABYM NLHIV tested for HIV within the past six months compared to 10.3 % of AGYW NLHIV.

Table 6.

Healthcare utilization of those not living with HIV and those living with HIV by sex (2022).

	NLHIV			LHIV
	ABYM	AGYW	p	ABYM	AGYW	p
	(n = 150) n %	(n = 340) n %		(n = 152) n %	(n = 523) n %
HIV Testing
Have you tested for HIV in the past 6 months?
Yes (All ages)	36	17	<0.0001
	27.7	6.7
Yes (Ages 18+)	34	14	0.0002
	28.3	10.3
ART Adherence
Are you currently on ART?
Yes				149	520	0.1
				98.0	100.0
Do you experience any side effects or discomfort?
No side effects or discomfort				177	390	0.4
				79.1	75.0
How good a job did you do at taking your medicines?
Good, very good, or excellent				141	507	0.08
				94.6	97.5

Abbreviations: NLHIV Not living with HIV, LHIV Living with HIV, ABYM Adolescent boys and young men, AGYW Adolescent girls and young women, ART Antiretroviral therapy.

Discussion

This analysis aimed to understand what, if any, are the sex-related differences between AYA living with and without HIV in Lusaka, Zambia. We found that while there are some differences, there are many similarities between ABYM and AGYW with respect to their views on social support, their risk factors and behaviors, and their utilization of HIV healthcare resources. Our analysis did not identify many differences between those living with and without HIV.

Our analysis revealed that participants had greater confidence in familial support compared to peer support; this was true across sex and HIV status. Over half of all respondents strongly agreed with all the familial support measures. In contrast, only one peer support measure (“I have friends with whom I can share my joys and sorrows”) experienced similar response rates. Prior studies highlight that peer and family support can be positive influences for AYA regarding HIV testing and treatment, and some suggest that peer support is more influential. One study looking at the effects of social support on AYA HIV testing in Nairobi, Kenya found that 12 % of their respondents reported being influenced by a parent to test while 22 % reported being influenced by a peer.⁽²¹⁾ Another study examining the role of family dynamics on peer adherence support among the LHIV population in South Africa found that the effect of peer support is dependent on an individual's family dynamics.⁽²²⁾ This study found that peer support only had a positive effect on health outcomes in well-functioning families; among dysfunctional families, peer support had a negative effect.⁽²²⁾

Our study was able to compare the two types of support and revealed that respondents had higher levels of confidence in family support than peer support. Our finding is supported by previous research conducted in South Africa related to daily oral pre-exposure prophylaxis for HIV (PrEP) which found that parental buy-in was a very influential factor regarding treatment uptake and adherence.⁽²³⁾ AGYW with parents who felt excluded from the consultation process or who had received misinformation regarding PrEP were more likely to discontinue their medication due to parental pressure.⁽²³⁾ Another study, also conducted in South Africa, found that parental involvement and support were key factors related to the success of an HIV testing and counseling program implemented at 17 high schools.⁽²⁴⁾

Over half of the ABYM and AGYW respondents reported food insecurity, and a sizable portion of respondents who were working reported feelings of job insecurity. Many of the geographic regions that are regularly impacted by food insecurity overlap with regions experiencing high rates of HIV transmission. For those living without HIV, risky coping strategies (i.e., exchange of sex for food or money), nutritional deficiencies, and compromised immunity – which all increase the risk of contracting HIV – can be the result of food insecurity.⁽²⁵⁾ Similarly, food insecurity can have devastating consequences for those living with HIV. HIV-associated wasting, as a result of nutritional deficiencies, is closely tied with mortality.⁽²⁵⁾ A strong body of evidence demonstrates that food insecurity is negatively associated with ART adherence.^(25,26) A lack of food can encourage patients to skip ART, since it is recommended that the medication be taken with food to reduce side effects. Recognizing the impact of food and job insecurity on HIV-related health outcomes is important when creating interventions to prevent the transmission and progression of HIV. Our findings demonstrated some sex-related differences regarding economic stability. This suggests that providing vocational support for ABYM living with and without HIV could be beneficial. Empowering AYA economically could go on to reduce HIV transmission (due to risky sex and poor nutrition) and increase ART adherence.

There were no significant differences between the sexes in the NLHIV cohort related to mental health risk. However, among the LHIV cohort, there were two categories of mental health risk in which AGYW experienced a higher burden. The data suggest that differences in mental health may vary more across HIV status than they do by sex. Although the literature on this issue remains sparse, there is some evidence to support our findings.^(27,28) Studies from Kenya and Rwanda show that AYAHIV experience depression at higher rates compared to their uninfected counterparts.^(29,30) Mental health is an often overlooked or neglected aspect of healthcare that has serious long-term impacts on economic development and healthcare expenses for countries.⁽³¹⁾ Existing literature emphasizes the importance of mental healthcare with regard to HIV-related health outcomes and our findings demonstrated that those living with HIV, especially AGYW, experience a greater mental health burden. It is, therefore, imperative to integrate mental health support in any intervention aiming to reduce HIV transmission and mortality. Increasing access to individual and group counseling can provide AYAHIV with a safe space to discuss pertinent issues and act as another support system for these individuals.

Substance use and sexual behavior risks were two categories that did exhibit statistically significant differences between sexes. ABYM exhibited a higher rate of substance abuse risk, as measured by frequent alcohol consumption and drug use other than alcohol and marijuana, than AGYW. This is not surprising, as there is a well-established body of literature that illustrates that substance abuse is higher among males in Zambia. In a 2017 national survey comprised of over 4000 Zambian adults between the ages of 18 and 69, men were found to engage in binge drinking at rates three or more times greater than that of women.⁽²⁰⁾ A 2022 study that examined substance use among AYAs from Kalulushi district, Zambia, found that males had increased odds of drinking than girls.⁽³²⁾ Our analysis confirms these previous findings. This suggests that targeted substance abuse interventions for ABYM could be valuable.

Our results show that sexual behavior risk, as measured by low/lack of condom usage with casual partners in the past 3 months, was significantly higher for AGYW than ABYM. This finding, as well as its magnitude, is true across HIV status. Our findings are supported by existing literature. That females exhibit a lower rate of condom usage demonstrates the reality for AGYW in Zambia who are often pressured into unprotected sexual activity with male partners.⁽³³⁾ An intervention targeted towards 10,000 Zambian AGYW found that transactional sex was the most cited barrier against sexual protection.⁽³⁴⁾ Transactional sexual relationships are not uncommon and are more prevalent in communities of low economic stability; the prevalence of transactional sex in sub-Saharan Africa can be as high as 16 %.⁽³⁵⁾ Shared spaces between AGYW and ABYM to discuss power differences in relationships could be beneficial for both groups, since condom usage among AGYW is influenced by their partners. These spaces, which can occur in the form of mixed-gender support groups that take place in adolescent-friendly community or clinic spaces in the style of the group education sessions that were a part of the SHIELD intervention, could encourage both males and females to advocate for increased condom use in their sexual lives.⁽¹⁷⁾

The findings demonstrated that healthcare utilization exhibited sex-related differences in behavior, but only for those living without HIV. Our results showed that ABYM had tested for HIV within the last six months at rates significantly higher than that of AGYW. We know, from existing literature, that younger adolescents test for HIV at lower rates compared to their older counterparts.⁽³⁶⁾ For this reason, we decided to see if this large gap in HIV testing rates between ABYM and AGYW would remain when looking at older respondents only. This difference, as well as the magnitude, persisted when we removed ages 15–17 from the response pool. We believe that stigma, both perceived and experienced and in multiple settings, could explain the notable differences in testing rates between ABYM and AGYW. In Zambia, HIV is still perceived as a sexually transmitted infection.⁽³⁷⁾ At the individual level, AGYW experience higher levels of stigma, and may refrain from testing because they do not want to be perceived as promiscuous by healthcare workers.⁽³⁸⁾ Evidence exists that these anxieties are not entirely unwarranted; healthcare workers have seen colleagues label AGYW with terms such as “prostitute” or “indulgent” and have also reported observing scolding behavior.⁽³⁷⁾ Finally, fears related to involuntary disclosure or a lack of privacy can also discourage AGYW from testing.⁽³⁹⁾ This suggests that AGYW require more targeted interventions related to increasing the uptake of HIV testing.

This analysis had a few limitations which would affect the generalizability of our results. Firstly, our analysis had far fewer ABYM respondents; there were roughly two to three times as many AGYW respondents living without and with HIV, respectively. Small sample sizes were also a limiting factor for measures that contained questions with skip patterns, such as our economic stability questions. The small sample size of ABYM, compared to AGYW, could have exaggerated observed differences between the sexes. Secondly, and on a related note, the skewed age distribution of AGYW NLHIV made it difficult to attribute sex-related differences as the reason for some of the differences. Between the three age groups, the other cohorts had roughly a quarter of their populations in the youngest group, half in the middle, and a quarter in the oldest. This was not true for AGYW NLHIV. This skewed age distribution could have affected our findings on questions related to school, work, and economic stability risk. Lastly, utilizing data from the baseline survey of an implementation trial that took place during the COVID-19 pandemic may make it difficult to generalize the results of our analysis to a wider population. This analysis did contain small amounts of missing data for specific responses (<2 % of the total sample size for each cohort). When encountered, the missing data was removed from the numerator and denominator.

While there are some differences between males and females related to risk factors and behaviors, overall, the groups have many similarities. The commonalities between the sexes suggest that joint interventions, such as education session that include both males and females, could be beneficial. A meta-analysis examining the effectiveness of 14 educational HIV prevention programs for adolescents found that peer-led education improved adolescents' HIV prevention knowledge, behavioral intentions, and attitudes.⁽⁴⁰⁾ These proposed sessions could serve as an opportunity to address common risk factors and provide forums for discussing sex-related power imbalances in relationships. Facilitated discussions between males and females can foster greater gender equality. Opening this type of dialogue can serve as a space for males and females to discuss sensitive issues related to condom and contraceptive usage. Our findings also suggest that males would benefit from targeted interventions related to substance abuse while females would benefit from targeted HIV testing interventions. For both males and females, regardless of HIV status, finding novel ways to integrate families into interventions could result in improved testing uptake and treatment adherence. Future research in this area should examine whether the observed differences and similarities persist with larger cohort of young people and in different geographic locations.

Data sharing

Deidentified data will be available for download from a data repository that will be specified by the study funder. This data will be freely available to all those interested in using the data and the specific terms of the data use will be provided once the data has been uploaded. NICHD is currently in the process of finalizing the data repository and we will share this information as soon as it is available.

Evidence before this study

The data used in this study comes from an implementation science trial. The trial was designed based on a prior study that was conducted in Zambia which analyzed barriers and facilitators. This study clearly highlighted the dual burden of cervical cancer and HIV in Zambia which led to the design of the implementation trial which addressed the dual burden through HIV testing and HPV vaccination.

Added value of this study

This study provides specific information on differences between adolescent girls and boys and young women and men that highlights differences and similarities in risk factors. Understanding the risk factors is important to design interventions to improve health outcomes for this vulnerable group.

Implications of all the available evidence

This study provides details on key risk factors that should be addressed in future interventions designed for adolescents and young adults.

CRediT authorship contribution statement

Sanjana Batabyal: Writing – review & editing, Writing – original draft, Visualization, Validation, Formal analysis, Data curation, Conceptualization. Ronald Mungoni: Writing – review & editing, Supervision, Resources, Project administration, Methodology, Investigation. Drosin Mulenga: Writing – review & editing. Nachela Chelwa: Writing – review & editing. Michael Mbizvo: Writing – review & editing, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization. Laura Nyblade: Writing – review & editing, Funding acquisition, Conceptualization. Yevgeniya Kaganova: Writing – review & editing, Software, Funding acquisition, Formal analysis, Data curation. Sonja Hoover: Writing – review & editing. Sujha Subramanian: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Sources of funding

This study was funded by the National Institute of Child Health and Human Development (grant number 1 UG3 HD096908-01).

Declaration of competing interest

None of the authors have any conflicts of interest to declare.