A community-based study of intention to use HIV self-testing among young people in urban areas of southern Ethiopia

Abstract

In sub-Saharan Africa, young people are at high risk of HIV infection, representing nearly 4 out of 5 new infections. HIV self-testing (HIVST), a new and proactive testing scheme that involves self-collection of a specimen and interpretation of results, is deemed potentially helpful for increasing testing amongst population groups like young people who do not frequently use routine testing services. This study assessed young people’s intention to use HIVST in urban areas of southern Ethiopia drawing on the Theory of Planned Behaviour. A community-based cross-sectional study was conducted with 634 participants aged 15–24 years from six town administrations of two zones in January to February 2023. The participation rate was 634/636 yielding 99.7%. The OraQuick HIVST kit was demonstrated to young people recruited in a door-to-door survey with a face-to-face interview using an electronic questionnaire in a mobile phone-based application. Intention to use HIVST was measured from a 6-point Likert scale with scores of agreements ranging from 1 to 6. Descriptive statistics and ordinal logistic regression analysis were done using STATA version 18. Most of the participants agreed that they would use HIVST if it was available (86.3% agreeing or strongly agreeing). Interestingly, young people who perceived themselves at some to high risk were 0.51 times less likely to be in the higher order of intention to use when HIVST is available to them than those who perceived themselves at no to low risk. Intention to use HIVST increased by a factor of 1.29, 1.84 and 2.35 for every one-unit increase on the mean favourable attitude, perceived behavioural control, and acceptability scores, respectively. The majority of young people intended to use HIVST. Young people’s perceived behavioural control, and acceptability of HIVST affected their intention to use. Intention and subsequent use of HIVST can be enhanced through an understanding of the role of risk perception and positive attitude, confidence to perform and acceptance of the test. Implementation studies are required to examine the actual uptake of HIVST among young people.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-77728-5.

Background

Internationally, two-thirds of people living with HIV (PLWH) in 2023 lived in the sub-Saharan Africa (SSA). In 2023, the majority of new HIV infections in SSA occurred among women (over 60%) and younger populations aged 15–34 years (72.5%)¹. Eastern and southern Africa are the most affected regions in SSA with 54% and 44% of global HIV prevalence and new infections in 2019 respectively were in these regions². The joint United Nations Program on HIV (UNAIDS) proposed a second round of its three ambitious targets to end the HIV pandemic in 2030: the so-called 90-90-90 and the 95-95-95 targets set in 2014 and 2020. These aim to test 90/95% of people with HIV infection, enrol 90/95% of those who tested positive into treatment, and achieve 90/95% viral suppression for those on treatment in 2020 and 2025 respectively^3,4. To achieve the first target (testing 95% of people with HIV infection) in the context of widespread stigma and discrimination towards HIV, the expansion of testing modalities that ensure the privacy and confidentiality, as well as reducing the subsequent stigma and discrimination, is essential. HIVST is a new and proactive way of testing oneself for HIV in a private place^5,6 and is one proposed strategy to increase HIV testing by the World Health Organization (WHO)⁶. Apart from improving privacy and confidentiality, HIVST may improve access to HIV testing services because of factors such as convenience in saving time and transport related costs enabling it the test of choice for areas and populations with greatest gaps in testing coverage⁵. Many countries have started to incorporate HIVST policies and strategies into their context, mostly across Eastern and Southern Africa and Western and Central Europe⁷.

As one of the countries in eastern Africa, Ethiopia had a high prevalence of HIV (3%) in 2013, being one of the 10 countries comprising 80% of PLWH in the SSA⁸. Even though there has been impressive progress across the years with a national prevalence of less than 1% in the 2016 Ethiopian Demographic and Health Survey (EDHS), there is a significant disparity in urban-rural residences (with HIV prevalence 7 times higher in urban areas)^9,10. Only 2 regions in the country (Addis Ababa city administration, and Harari regional state) achieved the testing and treatment of the 90-90-90 targets out of 10 regional states, and two city charters in Ethiopia, and all regions have achieved the third, viral suppression 90 target¹¹. However, the country did not achieve the testing 90 target particularly among young people aged 15–24 years and men¹². A recent spatial mapping study based on the 2016 EDHS identified that the Southern Nations Nationalities and Peoples Regional State was one of the regions with a lesser proportion of adolescent girls and young women who had ever tested for HIV, and being in the ages between 15 and 19 was a factor that was negatively associated with the testing rate¹³. In an urban context in Ethiopia, being male, and between 15 and 19 years old were negatively associated with HIV testing among youths 15–24 years old¹⁴. While HIVST has a potential role in increasing HIV testing, there is no existing research in Ethiopia on young people’s likely intention to take up HIVST if it were offered. The present study assesses the intention to use HIVST and factors associated with it among young people residing in urban areas of southern Ethiopia.

Young people, especially adolescent girls and young women, are at a disproportionately high risk for new HIV infection in SSA, accounting for 63% of all new infections in 2022^15,16. These groups are at particular risk because of complex factors that include engaging in multiple and age-disparate transactional sexual behaviours as a result of being at a socio-economic disadvantage, and religious and cultural factors like early marriage and other harmful traditional practices that encourage gender inequitable norms and practices¹⁷. The Ethiopian population-based HIV Impact Assessment (EPHIA) survey identified adolescents and youths at increased risk of HIV infection due to risky sexual behaviors related to early sexual debut and having multiple sexual partners^18–20. Despite this risky sexual behavior, the testing rate remained low with only three out of 10 older adolescent boys²⁰ and one in five young people aged 15–24 years tested for HIV^14,20. The Ethiopian Public Health Institute also identified that youths between 15 and 24 years had the least awareness of their HIV status (72.6%) compared with those in the older age groups [25–34 (82.2%) and 35–49 (89.9%)]¹¹.

There is limited research examining intention to perform HIV self-testing and existing literature has found varying degrees of intention for different population groups and contexts, albeit there was variation in definitions used to measure intention and factors from study to study. For example, 40.9% of young black transgender women and men who have sex with men (MSM) in the USA²¹, 38.5% of French people including MSM²², 23.8% of male clients of female sex workers²³, 23.7% of migrant male factory workers²⁴, 40% of MSM in China²⁵, 83% of MSM in Spain²⁶, 49.1% of MSM in Brazil²⁷, 86% of women in Nigeria²⁸, and 100% of female sex workers in Benin²⁹ intended to use HIVST. Regarding intention among young people specifically, 90.4% of Columbian youths³⁰, 66% of young university students in Zimbabwe³¹ and 100% of a predominantly young population (18–24 years) of MSM in South Africa³² were intending to use HIVST. Some of the studies also documented factors associated with participants intention to use HIVST^{21,22,24–26,31}. In studies among key populations and/or adults, intention to test was higher among: young transgender women in the USA who were comfortable in testing at home with the assistance of a friend or partner, and who reported stigma or fear as reasons for not testing²¹; male clients of female sex workers in China who had a positive attitude, perceived that HIVST uptake is under their control, and perceived at high risk for HIV²³; and MSM in China who were older, having a higher monthly income, being informed about HIVST, perceived accuracy of HIVST, and perceived ability to reduce fear during testing²⁵. HIVST intention was lower for young transgender women with high social support and having had a health insurance²¹; French MSM²²; and MSM in China having had higher scores on HIV knowledge²⁵. In terms of young people specifically, young Zimbabwean university students with prior history of HIV testing and who noted the absence of a counselor during HIVST as a distressing situation were more likely to intend to test than those who had no previous history of HIV testing or who did not consider the absence of counselor as distressing situation respectively³¹. There are no studies examining predictors of HIVST intention in Ethiopia, including for young people.

According to Ajzen’s Theory of Planned Behavior (TPB), intentions are self-instructions to perform behaviors or to obtain certain outcomes^33,34.This theory asserts a given behavior like HIVST is driven by the intention to perform that behavior³⁴. Although inconsistencies between intention and behavior can be expected, especially when one alters their intention before performing the behavior³⁴, there is strong evidence that intention explains much of the variance of behavior^35,36. Thus, to make important decisions on behaviors required for novel situations, as in the situation with HIVST, intention-driven action, and processing have been indicated as relevant³⁷. Even though not among young people, evidence among women attending antenatal care in Ethiopia³⁸, and other key populations elsewhere^39–41 have also signified that HIV testing intention is related with later HIV testing. The TPB is a potentially useful approach to considering HIVST intention for it is a very helpful theoretical model to examine intention for a given behaviour. It is an extension to the theory of reasoned action, considering also perceived behavioural control or self-efficacy as antecedent variables that directly affects a given behaviour as well as indirectly through behavioural intention⁴².

In the African context broadly and in Ethiopia in particular, young people are at most risk of new HIV infection and report low HIV testing rates. Ethiopia has endorsed an HIVST implementation guideline and optimizing strategy that focuses on accessing HIV testing services to key and priority populations including female sex workers, long truck drivers, and vulnerable young people^43–45. While the 90-90-90 targets review called for the application of HIVST as a promising strategy to improve testing rates¹¹, young people more broadly (outside those included in the priority groups like young daily laborers, assistants of long truck drivers, prisoners, etc.) are not yet included in the HIVST guidelines even though assisted HIVST is allowed for those opting in to test based on risk assessment at health facilities and the community. As the final phase of HIVST service delivery will be to move to the provision of unassisted HIVST in open access through social marketing^44,46, there is a need to investigate the intention of young people who are generally regarded as at high risk of HIV infection. Thus, before wide scale implementation of the HIVST among young people and the general community in Ethiopia, it is important to assess their intention to use HIVST and the factors associated with intention as there is a dearth of such evidence among young people in the country. Therefore, this study drew on the TPB and aimed to assess the degree of intention of urban young people to self-test for HIV and to identify factors associated with their intention in the south Ethiopian region through employing the constructs of the TPB and other potential factors identified from the literature. The findings from this study offer insights into how to optimise HIVST uptake of young people in the Ethiopian context and will also provide a baseline for future implementation research.

Methods

Overview

We conducted a community-based cross-sectional study aimed to address the intention of HIVST among young people aged 15 to 24 years. After three weeks of piloting the data collection instrument and recruitment of participants, data were collected between 24 January 2023 to 10 February 2023. This study was approved by the Flinders University Human Research Ethics Committee (HREC 5535) and is reported guided by the STROBE guidelines⁴⁷. The STROBE checklist is given in Supplementary Material 2.

Study setting

The study was conducted in urban administrations of Gamo and Gofa zones in the South Ethiopia Region, which are respectively located 409 and 550 km southwest of Addis Ababa, the capital city of Ethiopia. The Ethiopian population in the 2007 population and housing census (the most recent census) was 73,750,932⁴⁸, and the population growth rate is estimated at 2.5% in 2022⁴⁹.The main administrative centers for Gamo and Gofa zones are Arba Minch and Sawla towns which have populations of 123,446 and 85,704 respectively based on the 2007 Ethiopian census projected for the year 2022. Gamo and Gofa zones have six and four town administrations respectively. The total number of young people aged 15 to 24 years living in the six towns of the study area was estimated to be 61,834⁵⁰. In the Gamo zone, a total of 14 public health facilities were providing anti-retroviral treatment services and 3 were providing HIVST services. There is an increase in the total number of people who tested positive for HIV (278 (0.4%) to 315(0.5%)) irrespective of age in the Gamo zone from the year 2021 to 2022⁵¹.

Youth friendly centres located in the health facilities and community youth centres in Ethiopia provide voluntary HIV counselling and testing service, and promotive and preventive activities targeted in risk reduction and awareness raising on HIV⁵².

Participants

Eligibility criteria

The study population were young people between the ages 15–24 years who were permanent residents (lived for at least six months) in the study area. The definition of young people varies by organization and country. According to the United Nations Population Fund and the WHO, young people is defined as those individuals in the ages of 10 to 24 years^53,54. The African Charter defines youths as those in the age group of 15 to 35 years⁵³ and Ethiopia defines its youths as those between the ages 15 to 29 years⁵⁵. In this study, we have targeted young people between the ages of 15 to 24 years and all genders, to align the lower age boundary with the age of consent for HIV testing in Ethiopia⁵⁶ and the upper age boundary with global recommendations by the United Nations and WHO⁵³, and the national level classification the EDHS and the EPHIA reports on^10,20.

Youths were eligible to participate if they did not have an HIV diagnosis and were willing or able to provide consent, were not critically ill and did not have a disability that prevented participation in an interview.

Recruitment procedure

The study participants were randomly selected from four town administrations of the Gamo zone (Arba Minch, Kamba, Selamber, and Birbir) and two from the Gofa zone (Sawla and Beto) using a systematic, multistage sampling technique with households as sampling units. At first, towns in the two zones were stratified by their size as small (less than 50,000 population) and large towns (greater than or equal to 50,000 population). The main towns in the two zones (Arba Minch and Sawla) were included in the study for they were the only ones to fulfill the large town definition in the two zones. Out of five small-town administrations in the Gamo zone, we selected three towns (Birbir, Kamba and Selamber) using a lottery method of randomly drawing the names of towns on a coiled paper put on a dish. Likewise, one town (Beto) was selected from three small towns (Bulki, Beto and Laha) in the Gofa zone. In the second stage, ‘kebeles’/’ketenas’ (the smallest administrative units) from each of the included towns were selected based on the same lottery method. Households within these ‘kebeles’/’ketenas’ were included using a systematic random sampling allocated by proportional sampling based on the number of young people in each kebele with sampling intervals calculated for each town from the total sample size by weighing to these proportions. Whenever there was more than one young person in a household, one was selected randomly using a lottery method. For closed houses or households with no one at home, before considering it as a non-response, data collectors visited the same house for the second time on the next day of data collection, and if there was still no one at home, they checked the immediate next household for eligible participants. Likewise, when a household eligible by the random allocation using the sampling interval was found to have no eligible participants, the next households in order were approached to recruit eligible participants. Further details on the number of households and sampling intervals for each town is available in the Supplementary Material 3.

Data collection techniques and instrument

Data were collected via face-to-face interviewing of participants at their homes using 18 trained data collectors (five for Arba Minch town, three for each of Sawla, Birbir, and Selamber towns, and two for each of Kamba and Beto towns) who were aged 21–39 years and received a day long training from the research team. The training was aimed to familiarize the data collectors with interviewing skills, demonstration of HIVST kits and adherence to ethical standards. Data collectors were selected purposively based on criteria of prior experience in survey methods, having a smartphone, being female, being a peer navigator trained on HIVST in the main towns, and being a member of youth-friendly center in the respective small town. Four supervisors with a Master of Public Health degree and with prior experience in related field research practice were recruited to monitor and evaluate the data collection process.

The interviewers used a smart-phone based Kobo Toolbox application to record data without the need to have an internet connection or mobile carrier service, with data submission to an online server processed at a later time⁵⁷. The questionnaire template used for data collection is provided in the Supplementary Material 4.

Because the pilot study found that few of young people had ever heard about HIVST, the attitude of young people toward HIVST was assessed after demonstrating OraQuick HIVST kits to them and describing the procedure for performing the test. The OraQuick HIV-1/2 kit is an innovative HIVST kit which is an immunochromatographic and visually read test that is used to detect antibodies of HIV 1/2 and is one of the recommended test kits by the WHO⁵⁸. A diagnostic accuracy test conducted in public health facilities of Addis Ababa, Ethiopia showed a high diagnostic performance of the OraQuick HIV self-test and indicated that it has a high potential for use in Ethiopia⁵⁹.

Variables and Measurement

Outcome variable

The outcome variable of the study is the intention to self-test for HIV. The intention to use HIVST was measured using the item: “If HIVST is available to me, I would use it”. The item is on a six-point Likert scale ranging from strongly disagree (‘1’) to strongly agree (‘6’). We have used a single item to measure intention to use HIVST based on original recommendations by Icek Ajzen⁶⁰ and later by Fishman et al., 2020⁶¹. The latter study indicated that a single item used to measure intention demonstrated a strong predictive validity, and little additional value was added with the addition of one or more items⁶¹.

Independent variables

The latent independent variables considered in this study include constructs of the TPB (attitude towards HIVST, subjective norms about HIVST, and Perceived Behavioral Control), knowledge of HIVST, discriminatory attitude about HIV, and acceptability of HIVST. According to Seckhon et al.’s theoretical framework of acceptability⁶², acceptability is defined distinctly from intention, with its constructs proposed as predictors of intention to use an evidence-based healthcare practice. Advancing on this, a recent theoretical framework that has taken into account the Seckhon et al.’s framework and developed to theorize acceptability of health and social interventions among young people in Africa hypothesized acceptability as a predictor of intention to perform a given health behavior⁶³. Thus, in this study, we assessed the prospective acceptability of HIVST and considered the acceptability of HIVST as one of the potential predictors of intention to use HIVST.

In the context of this study, the basic constructs of the TPB were measured: attitude (perception of how important HIVST is), subjective norms (SN: perceived support of performing HIVST by others valued to the young person), and perceived behavioral control (PBC: the young person’s confidence in their ability to perform HIVST). In the TPB, PBC affects actual behavior directly or through affecting one’s intention³⁴.

SN and PBC were measured from 4 and 2 items that are presented in the questionnaire template (Supplementary material 4). The items used to measure SN and PBC were developed based on the TPB questionnaire originally suggested by Icek Ajzen, 2006⁶⁰. Internal consistency among items used to measure SN was assessed using Cronbach’s alpha (α = 0.63) which is in the acceptable range^64,65. The mean score of the four items is used to calculate a SN score for each participant. The minimum mean score was 1.00 and the maximum was 5.75 from the sum of scores given for each of the questions from points 1 to 6 for “strongly disagree”, “disagree”, “somewhat disagree”, “somewhat agree”, “agree” and “strongly agree” respectively yielding total scores ranging from 4 to 24 for the four items. Higher scores indicate more favorable views. The reliability of the two items used to measure PBC was calculated based on suggestions by Eisinga⁶⁶ using Spearman correlation which showed a positive association between the items (Spearman’s rho = 0.319, p-value less than or equal to 0.001). The total scores ranged from 2 to 12 for the two items.

Attitude questions presented in the questionnaire template (Supplementary Material 4) were adapted from a previous study in a similar setting⁶⁷ and rated based on a six-point Likert scale response. Principal component analysis (PCA) and reliability analysis were undertaken to check whether the 11 attitude items were measuring the same construct or not. Before conducting the principal components analysis, items 7–11 (Q.3.8–11, Supplementary Material 4) were reverse scored so that high scores on attitude could indicate higher attitude scores. However, at the first iteration step of the PCA, the items showed two distinct factors: the first six items together made a single factor with eigenvalues greater than one and factor loadings greater than 0.4, and the reverse-scored items likewise indicated a distinct second factor. Thus, we considered the attitude variable as two separate variables, favorable and unfavorable attitude towards HIVST. In the latter case, we used the items in their original form (not reverse-scored). The reliability of the positive items measured using Cronbach’s alpha was acceptable (α = 0.78), and the negative items were also found to be acceptable (α = 0.68)^64,65. Mean favorable attitude, and mean unfavorable attitude were computed as two distinct variables. Higher scores on favorable attitude indicate a more favorable attitude towards HIVST. Likewise, higher scores on unfavorable attitude indicate a more unfavorable attitude towards HIVST. An additional description of the the PCA analysis and measurement of other latent variables is provided in the Supplementary Material 5.

The additional factors considered to predict the outcome variable were socio-demographic factors (age, sex, average annual income, educational status, occupation, marital status); HIV test and risk perception-related factors (history of HIV test, and perceived risk to acquire HIV infection); and sexual and reproductive health-related factors comprising risky sexual behavior (such as condomless sex, and multiple sexual behaviors), being sexually active, age of sexual debut, and having recent (last 12 months) history of sexually transmitted infection. These items were measured from a single item that asked the participants to list or rate about the specific attribute. A conceptual framework depicting hypothesized relationships between the outcome and explanatory variables is presented in Fig. 1.

Fig. 1.

A conceptual framework depicting the relationship between intention to use HIVST and explanatory variables among young people in the urban areas of southern Ethiopian region.

Bias

Social desirability bias was anticipated in this survey context as the contents of the survey instrument contain sensitive matters related to one’s sexual and reproductive characteristics, and health service choices. We have taken this into account and designed the chronological order of the questions from less sensitive information about socio-demographic characteristics and knowledge matters to the more sensitive aspects of sexual practice. Data collectors were all female and were peer navigators trained on HIVST and working in the same community to provide HIVST service to key and priority populations in the two main towns and had a rich experience in working with youths at youth-friendly centres in the smaller towns where peer navigators were not yet in practice. The decision to involve only female data collectors was based on a reluctance of female young people to participate in the study when male data collectors approached them during the piloting of the survey instrument. In addition, data were collected in a private place in the participants’ homes. The use of Kobo Toolbox reduced missing of important data to the bare minimum for most alternatives were set as required information before scrolling down and asking the next question/s. As HIVST kits were a new technology to the community, the likelihood of information contamination was also high. Data collectors were advised during initial training to approach young people in private and advise them not to influence their peers.

Study sample size

The sample size was calculated using a single population proportion formula based on assumptions of 95% confidence interval (z α/2 = 1.96 ), 5% degree of precision,, and 75% willingness for future HIVST among young adults in Kenya⁶⁸. The formula and the calculation are presented as follows.

Where n = sample size, zα/2 = reliability coefficient, p = proportion and D = degree of precision⁶⁹.

A design effect of 2 was used for compensation of error arising from multistage cluster sampling. Moreover, to compensate for potential non-response, a 10% additional number of participants was added. The final sample size was calculated by multiplying the calculated sample size by the design effect and adding the additional 10% samples considered for potential non-response. Hence, based on these assumptions and considerations, the final sample size was 636.

Data processing and analysis

Data were transferred to the STATA software version 18 for cleaning and statistical analysis. Descriptive statistics were calculated to show the frequency of categorical variables in proportions or percentages, to explore the presence of outliers and to transform data as needed, to compute means and standard deviations for continuous variables, and to compute scores for composite variables. Ordinal logistic regression was used to identify independent factors associated with intention to use HIVST.

We checked multicollinearity and goodness-fit-of the model using variance inflation factor (VIF ≥ 5) criteria and a likelihood ratio chi-square test that checks whether there is a significant improvement in the fit of the final model relative to the intercept-only model (p-value < 0.001). The proportional odds assumption of ordinal logistic regression was checked using a test of parallel lines (non-significance of the − 2LL chi-square) (p-value = 0.54) using the omodel logit command⁷⁰. Further details about data processing and analysis are given in the Supplementary Material 5.

Results

Overview

Of the proposed sample size of 636, we recruited 634 participants into the study yielding a response rate of 99.7%. Two participants did not proceed with providing information after withdrawing consent to participate in the study with complaints of time shortage and perceived inability to provide correct answers. In 147/634 (23.2%) households there was more than one eligible young people (and the lottery method was used) and in 79/634 (12.5%) households, young people were not present at their home on two occasions of consecutive days of data collection.

Participant characteristics

Out of the 634 participants, 350 (55.2%) were females. The mean age of the study participants was 20.17 (SD = 2.51). The average monthly household income was 5,546.01 Ethiopian Birr (SD = 6,013.25) or 103.4 USD. The average number of people living in the household who live on the income was 4 (range 1-11) people. Nearly two-thirds, 382/634 (60.3%) of the participants were single and more than a third of them 252/634 (39.7%) were in a relationship. Regarding their occupation, 402/634 (63.4%) of the participants were students, and 67/634 (10.6%) were self-employed. Nearly half of the study participants, 302/634(47.6%) were attending in secondary education in the high schools, and 127/634 (20%) were enrolled in colleges, and vocational schools.

Intention to self-test for HIV

Young people’s intention to use HIVST was assessed from the question “I would use HIVST if it were available to me”. The majority of the participants (86%) agreed or strongly agreed that they will use HIVST if it is available (Table 1).

Table 1.

Intention, subjective norm, and perceived behavioral control of young people towards HIVST in urban areas of Gamo and Gofa Zones, southern Ethiopia, February 2023 (n = 634).

Variable		Category, frequency (%)
		Strongly disagree	Disagree	Somewhat disagree	Somewhat agree	Agree	Strongly agree
Intention	I would use HIVST if it were available to me	11 (1.7)	36 (5.7)	7 (1.1)	35 (5.5)	441 (69.6)	104 (16.4)
Subjective norm	Approval or permission from family is needed for HIVST	45 (7.1)	334 (52.7)	24 (3.8)	51 (8.0)	171 (27.0)	9 (1.4)
	Approval or permission from friends is required for HIVST	44 (6.9)	441 (69.6)	19 (3.0)	45 (7.1)	84 (13.2)	1 (0.2)
	Approval or permission is required from a partner for HIVST	35 (5.5)	265 (41.8)	26 (4.1)	48 (7.6)	237 (37.4)	23 (3.6)
	Knowing about HIV status of myself is expected from me	6 (0.9)	15 (2.4)	11 (1.7)	9 (1.4)	508 (80.1)	85 (13.4)
Perceived behavioural control	I am confident to test for HIVST	8 (1.3)	17 (2.7)	14 (2.2)	36 (5.7)	439 (69.2)	120 (18.9)
	Deciding whether to use HIVST or not is up to me	6 (0.9)	269 (4.1)	10 (1.6)	28 (4.4)	470 (74.1)	94 (14.8)

HIV testing history and perception of HIV risk

Slightly over half, 327/634 (51.6%) of young people in the study had not ever tested for HIV and 15/634 (2.4%) of participants were not sure whether they had ever tested or not. Nearly two-thirds 384/634 (60.6%) believed that they were not at risk of HIV infection. The remaining 250/634 (39.4%) varied in their level of HIV infection risk perception; 23/634 (3.6%) high risk, 43/634 (6.8%) at moderate risk, 66/634 (10.4%) somewhat at risk, and 118/634 (18.6%) at low risk. Among those ever tested for HIV, the number of times they tested ranged from 1 to 8, with a mean testing frequency of 2.23 (SD = 1.410). The more recent time of the test was 1 month and the longest time since the last test was 7 years and 4 months, with the average time after the test 12.01 months (SD = 14.98).

Sexual risk behavior and STI history

More than a third of the participants 236/634 (37.2%) responded that they have ever had a sexual intercourse and about 1 in 10 were not willing to respond to their experience on sexual affairs. Out of those who have had any sexual intercourse history, 208/634 (32.8%) had been sexually active in the six months before the survey. Among those who had a recent sexual intercourse history, 71/634 (11.2%) had reported that they had more than one sexual partner. Regarding their condom use during recent sexual practice, only 51/634 (8%) regularly used condoms.

Knowledge about HIVST

Of the 634 young people who participated in the study, 177 (27.9%) had ever heard about HIVST and this was from various sources. Health professionals accounted for the largest source of information (18.8%) followed by the mass media (6.9%). Based on the overall knowledge score from the five reliable items, 23.9% participants had a high level of knowledge while nearly two-thirds (62.6%) had a low level of knowledge (Table 2).

Table 2.

Awareness and knowledge about HIVST among young people in urban areas of Gamo and Gofa Zones z, southern Ethiopia, February 2023 (n = 634).

Variable	Category	Frequency	Percentage
Ever heard about HIVST	Yes	177	27.9
	No	406	64.0
	Not sure	51	8.0
Ever seen HIVST	Yes	131	20.7
	No	464	73.2
	Not sure	39	6.2
HIVST is legal in Ethiopia	Yes	98	15.5
	No	323	50.9
	Not sure	213	33.6
HIVST can be performed from blood samples	Yes	354	55.8
	No	121	19.1
	Not sure	159	25.1
HIVST can be done at home	Yes	236	37.2
	No	157	24.8
	Not sure	241	38.0
HIVST can be done by the assistance of health care professionals	Yes	461	72.7
	No	48	7.6
	Not sure	125	19.7
HIVST can give a negative result if tested in 3 months of HIV infection	Yes	299	47.2
	No	107	16.9
	Not sure	228	36.0
A negative HIVST should be re-tested after 3 months	Yes	447	70.5
	No	36	5.7
	Not sure	151	23.8
A positive HIVST shows a definite HIV infection	Yes	386	60.9
	No	89	14.0
	Not sure	159	25.1
Overall knowledge	Low	397	62.6
	Moderate	85	13.4
	High	152	24.0

Discriminatory attitude towards people living with HIV

Most of the participants did not report discriminatory attitudes towards PLWH, with 554 (86%) reporting that they would buy fruit and vegetables from a person known to be HIV positive and 605/634 (95.4%) agreed that children should learn together with HIV-positive children. Taken together, 610/634 (96.2%) did not express agreement with either of these two indicators of discriminatory attitude towards PLWH.

The attitude of participants towards HIVST

About two-thirds of young people (70.5%) agreed that HIVST would increase public awareness about HIV status, increase testing frequency (68.0%), increase the chance of early treatment (63.7%), and save time spent at clinics or hospitals (63.4%). Around half of the participants also agreed that HIVST ensures privacy (55.5) and helps for prevention of HIV infection (49.1%). On the other hand, about half disagreed that instructions on the OraQuick HIVST kit are difficult to read and understand (50.3%). More than a third (37.2%) and about half (46.4%) respectively disagreed that HIVST might lead to partner and family member disapproval. Young people were not sure to rate about their attitude in all the items with 1.6–30.8% of the participants in the study responding respectively as they were not sure whether HIVST saves time from clinic or hospital visits and results can be incorrectly interpreted or not. The mean favorable and unfavorable attitude scores for each participant ranged from 1.17 to 7.00, and 1.40 to 7.00 respectively. The total mean computed from all the mean scores of the participants was 5.66 (SD = 0.88) and 4.02 (SD = 1.15) respectively for favorable attitude and unfavorable attitude (Table 3).

Table 3.

Attitude towards HIVST among young people in Urban areas of Gamo and Gofa Zones, southern Ethiopia, February 2023 (n = 634).

Variables	Category, frequency (%)
	Strongly disagree	Disagree	Somewhat disagree	Somewhat agree	Agree	Strongly agree	Not sure
Ensures privacy	19 (3.0)	33 (5.2)	19 (3.0)	61 (9.6)	352 (55.5)	111 (17.5)	39 (6.2)
Help for prevention of HIV infection	10 (1.6)	94 (14.8)	15 (2.4)	94 (14.8)	311 (49.1)	71 (11.2)	39 (6.2)
Increases public awareness about HIV status	6 (0.9)	23 (3.6)	14 (2.2)	37 (5.8)	447 (70.5)	89 (14.0)	18 (2.8)
Saves time spent at clinic or hospitals	11 (1.7)	24 (3.8)	14 (2.2)	32 (5.0)	402 (63.4)	141 (22.2)	10 (1.6)
Increases testing frequency	5 (0.8)	16 (2.5)	11 (1.7)	40 (6.3)	431 (68.0)	116 (18.3)	15 (2.4)
Increases the chance of early treatment	5 (0.8)	12 (1.9)	10 (1.6)	60 (9.5)	404 (63.7)	122 (19.2)	21 (3.3)
Instructions are difficult to read and understand	50 (8.2)	319 (52.3)	29 (4.8)	325.2	96 (15.7)	24 (3.8)	84 (13.8)
Result can be incorrectly interpreted(N = 633)a	12 (1.9)	150 (23.7)	19 (3.0)	48 (7.6)	176 (27.8)	33 (5.2)	195 (30.8)
People may harm themselves if not counselled and test positive	12 (1.9)	63 (9.9)	17 (2.7)	116 (18.3)	302 (47.6)	60 (9.5)	64 (10.1)
HIVST might lead to partner disapproval	23 (3.6)	236 (37.2)	21 (3.3)	63 (9.9)	174 (27.4)	34 (5.4)	83 (13.1)
HIV might lead to family member disapproval (N = 633)a	18 (2.8)	294 (46.4)	24 (3.8)	59 (9.3)	146 (23.0)	28 (4.4)	64 (10.1)

Note: a = One participant declined to answer this question.

Subjective norm and perceived behavioral control

The mean SN calculated from all the means of each participant from the 4 items used to measure SN was 3.87 (SD = 0.87) while the mean PBC score calculated from the mean of each participant’s score was 4.93 (SD = 0.69) with a minimum mean score of 1.5 and maximum 6 (Table 3).

Factors associated with intention to use HIVST

A chi-square test did not show a statistically significant difference in degrees of intention between male and females (Pearson Chi-square p-value = 0.822), younger youths (15-17 yeas old) and older youths (at or above 18 years old) (Pearson Chi Square p-value = 0.995), those having moderate to high knowledge, and those with low knowledge about HIVST (Pearson Chi-square P-value = 0.082), and those with marital or partnership status and were single (Pearson Chi Square = 0.802). However, a significant difference was observed between those who attended to a higher education and those with secondary education or below secondary education (Pearson Chi-square p-value = 0.032). Of 16 variables considered in the initial model, four variables, namely perceived risk for acquiring HIV infection (p-value = 0.03), perceived behavioral control (p-value < 0.001), attitude (p-value < 0.001), and acceptability (p-value < 0.001) showed a statistically significant association with young people’s intention to self-test for HIV in the final ordinal logistic regression model.

The odds of being in the higher order of intention to use HIVST when it is available for those perceiving themselves at some to high risk for HIV infection was 0.51 times that of those who perceived themselves at no to low risk. On the other hand, the odds of being in the higher order of intention to use HIVST increased by a factor of 1.71, 2.62 and 3.93 for every one-unit increase in favorable attitude, perceived behavioral control, and acceptability mean scores. This indicates that young people with a higher mean score on PBC, high mean favorable attitude, and high mean acceptability were more likely to intend to use HIVST (Table 4).

Table 4.

Factors associated with intention to use HIVST among young people living in urban areas of Gamo and Gofa Zones, southern Ethiopia, February 2023.

Variable	Category	B	Sig	OR	95% CI
Intention to use HIVST					Lower	Upper
Perceived risk	Low to no risk	0a		1
	Some to high risk	-0.68	0.03	0.51	0.27	0.93
PBC		0.96	< 0.001	2.62	1.84	3.73
Favourable Attitude		0.54	< 0.001	1.71	1.29	2.25
Acceptability		1.37	< 0.001	3.93	2.35	6.56

Note: a = reference group.

Discussion

This community-based cross-sectional study examined urban young people’s intention to use HIVST in Southern Ethiopia in a door-to-door interview after a demonstration of HIVST kits. Our analysis revealed a high intention for HIVST use among young people and participants’ attitudes, perceived behavioral control, perception of risk and acceptance of the test were significantly associated with their intention to use. The Ethiopian government is currently implementing assisted and unassisted HIVST in two successive phases among key and priority populations including female sex workers, long truck drivers and their assistants, prisoners, and daily laborers which consisted of people over a range of ages including many vulnerable young people. They are considering extending HIVST to all community members in an open access scheme through social marketing in the next third phase in the future^44. These findings are therefore of note for broader HIVST implementation amongst young people as the study included a general population of young people from various backgrounds including those in and out of schooling and indicated the need to consider intrinsic factors related to their attitude and perception.

The high intention to use HIVST in this study is similar to previous reports among youths in other parts of Africa – for example young university students in Zimbabwe³¹, and young MSM in South Africa³² and other key populations in Nigeria²⁸, and Benin²⁹. However, the finding is higher than most studies conducted among young and adult key populations in high- and upper-middle-income countries^21–25,27. This disparity in intention with most of the studies from high- and upper-middle-income countries and low-income countries like ours might be related to the higher preference of young people to test in private so that they can avoid discrimination from health care practitioners, parents, and the community in the latter context. Stigma and discrimination are some of the commonest barriers to accessing health facility-based reproductive health services including HIV testing among young people as supported by a recent review in SSA⁷¹. In addition, in countries with middle to high economies, the routine HIV testing service will likely be confidential, accessible, and of quality that might not necessitate consideration of HIVST⁷². Study variations may also relate to the timeframe measured for intention. Some studies have examined intention within a short period (6 months to 1 year)^21–25,27 and others, such as our own, have left it open or asked about participants’ intentions at any time when HIVST becomes available to them^26,28–32. Our findings, and those of other studies seem to indicate that there is a higher reported intention to use HIVST when the time frame is not fixed in advance. In what follows, we present the factors shown to influence the intention to use HIVST among the participants.

In our study, young people with high favorable attitude scores indicating more positive views towards HIVST showed greater intention than those who were less positive. This concurs with what the TPB³⁴ would predict, and the findings of previous studies which considered the attitude of young people^73–76 and other population groups^24,77–79 as predictor of their intention to use HIV testing services. The independent association of attitude in predicting intention to self-test suggests that young people’s readiness to take up HIVST rests on their intrinsic beliefs about the HIVST. This was further strengthened by the low knowledge about HIVST among the target groups. Although most of the participants had a high attitude score based on the on-site demonstration of kits, high knowledge remains the basic antecedent factor for a positive attitude, intention and subsequent take-up of HIVST. The insignificant association with unfavourable attitudes might imply that young people who had misconceptions also had better intentions to self-test given their high interest and curiosity to take up HIVST.

In terms of the other elements of the TPB, perceived behavioral control was also positively associated with intention. Measured from 4 items that assessed whether young people needed approval or permission from their family, friends, or partner, or they have a normative belief that knowing their own HIV status is something expected of them or not, the subjective norm of the participants did not show a significant association with their intention in this study. Perceived behavioral control has been significantly associated with the intention for HIV testing in various previous studies across the globe^24,74,79,80. While a global systematic review conducted in 2017 indicated that there was inconsistent or insufficient evidence to prove the relationship between perceived behavioral control, subjective norm and HIV testing intention, it noted that perceived behavioral control did show a positive association with intention to test in the African context specifically⁸¹. Measured from two items in our study which asked about young people’s ability to test and make autonomous decisions to conduct HIVST, young people’s perception that HIVST is under their control might be an essential step for HIVST uptake.

Beyond TPB constructs, a number of other factors were associated with intention. Young people’s perception of risk for HIV infection was negatively associated with HIVST intention in the present study. This surprising finding suggests that young people who perceived themselves as low to no risk were more likely to intend to use when HIVST is available to them than those who perceived themselves at a some to high risk. This contradicts previous literature that documented a positive association between risk perception and intention to use HIVST or HIV testing more generally in the USA, Thailand, and China^21,80,82, and no association with behavioral intention to self-test and HIV testing in another study among migrant factory works in China²⁴. The perceived high risk for infection, and its negative relationship with intention to use HIVST, might be related to fear of positive test results. This was a common reflection in the HIVST literature^83–88 where young people themselves and other stakeholders had concerns about social harm arising from testing positive in the absence of appropriate counseling. In other words, those who perceived themselves at some or high risk of HIV infection would be frightened of testing positive and those who perceived themselves at low risk to no risk would have a heightened confidence to intend to test out of the expectation that they will test negative for HIV.

Acceptability of HIVST was found to be the strongest predictor of intention to use HIVST in the current study. The association between acceptability and intention to use a health care intervention or service is only a recent suggestion in the public health literature when Seckon et al., 2017 distinctly defined acceptability as a multifaceted construct that can be taken as antecedent to intention to act on a given health care intervention or behavior⁶², built on by Casale⁶³. Since acceptability was measured as a composite from questions that assessed participants’ perception of taking HIVST (as comfortable, not demanding effort, fair to all, improves knowledge of HIV, clearly beneficial to know their HIV status, and that they can perform it confidently), it is plausible that those having a high mean score of acceptability (who were in more agreement with) would likely intend to use HIVST.

As intention is a strong predictor of prospective behavior (in our case actual HIVST), there is a promising prospect to adopt HIVST among young people in the study region and to the Ethiopian urban young people broadly to increase HIV testing and early linkage to treatment. Thus, the high level of intention in our study is an important finding for policymakers in Ethiopia to consider as they roll out HIVST more generally among young people in the country. The independent association of factors like attitude, perceived behavioral control, and acceptability of HIVST shows young people’s intrinsic feelings about HIVST outweighed any external or non-modifiable factors related to their socio-demographic background. The paradoxical association between low to no perceived risk of acquiring HIV and high intention would suggest the importance of tailoring HIVST promotion on how young people perceived themselves and raising awareness among those who perceived themselves at high risk that HIVST is not confirmatory test or testing positive means not necessarily having the infection. Overall, the factors shown to be associated with young people’s intention to use HIVST are intrinsic to young people’s perceptions and beliefs. This would indicate the importance of the creation of public awareness and clearance of misconceptions through wide-scale promotion of the testing kits using public outlets including the mass-media, and social media, and engaging young people in peer and community conversations locally in advance to the expansion of the HIVST service among young people in the general community. No association with subjective norms of the participants might indicate that young people were not submissive to their role models or significant others, or the lack of culture-motivated advocacy of HIV testing in the study area. This highlights the likely important role of cultural, resource and HIV-specific contexts for these constructs in relation to HIV testing in general and HIVST specifically.

Study strengths and limitations

There were several strengths in our study. This was the first study in Ethiopia to examine HIVST intention amongst young people, as well as explore the predictors of this intention. We used the TPB to hypothesize cognitive variables related to one’s beliefs and perceptions and found a partly consistent finding with what has been formulated in the theory. Our data collectors conducted a demonstration of HIVST at each face-to-face interview, which meant people had an accurate understanding of what was involved with the test. We used data collectors who are trained in HIV self-testing and working as peer navigators and as members of youth-friendly centers in the community. This meant they had highly developed skills in building rapport with young people and discussing sensitive matters and there was a high response rate for the survey. However, there are a few limitations to consider. As noted in “Methods” section, social desirability and information bias were possible because of the sensitive nature of HIV matters. While we sought to address this bias by including only women in the data collection team and through placement of questions to the latter part of the survey, the refusal of some young people to disclose information about their sexual experience might have underestimated the real magnitude of risky sexual behavior in the study area. Although we selected one young people per household for an interview, there were households with complex families of tenants and household heads’ families living together in a single compound. With priority given to permanent residence in the specific town, randomly selecting the young people of the household heads’ family alone might have introduced selection bias. While our purposive selection of two of the main towns for the study was undertaken to balance competing priorities of resource limitation and widening the scope of the study over a range of geographic locations (to ensure recruitment of a wide range of young people with a heightened risk of being engaged into risky sexual behavior and increased demand for HIVST), this might also have introduced selection bias. Regarding the design of the study, its cross-sectional nature limits the causality of association or temporal relationship between the outcome and independent variables, suggesting the need for stronger implementation research on HIVST among the study targets. The other limitation of the study was recruitment of YP based on availability in the randomly selected household while ideally a census of eligible participants would have been done. This was not possible because of resource limitations, and as for absence of participants at the home, we have recruited participants from the neighboring households in order.

Conclusion

The study showed a high intention to use HIVST among young people in the urban areas of Southern Ethiopia. Intention to self-test for HIV was associated with perceived low to no risk of HIV, attitude, PBC, and acceptability of HIVST. Interventions targeting HIVST among young peoples in the general community need to work particularly closely with young peoples with high-risk perception, negative attitude, perceived least control of HIVST, and low acceptance of HIVST.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Abbreviations

EDHS

Ethiopian Demographic and Health Survey

EPHIA

Ethiopian Population-Based HIV Impact Assessment

HIV

Human Immunodeficiency Virus

HIVST

HIV self-testing

MSM

Men who have sex with men

PBC

Perceived Behavioural Control

PLWH

People Living with HIV

SN

Subjective Norm

SSA

Sub-Saharan Africa

TPB

Theory of Planned Behaviour

WHO

World Health Organization

Author contributions

E.A.Z., J.S., H.A.G., and A.Z. contributed to the study conception and design. E.A.Z. wrote the draft proposal, and obtained ethical clearance through the guidance of A.Z., J.S., and H.A.G. Material preparation, and data collection were performed by E.A.Z., B.M.G., and K.T.W. E.A.Z. conducted data management and analysis with K.T.W. through the guidance of A.Z., J.S., and H.A.G. The first draft manuscript was written by E.A.Z., and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This study was funded by Flinders University, College of Medicine, and Public Health.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Human Research Ethics Committee (HREC) of Flinders University (Project ID:5535) and the Institutional Research Ethics Review Board (IRB) of Arba Minch University (Reference Number: IRB/1320/2022).

Consent to participate

Written informed consent was obtained from the participants and from their parents or guardians for minors. This was done using a translated version of a format initially provided by the Human Research Ethics Committee of Flinders University.

Standards of reporting

The STROBE checklist for reporting observational studies was used and is attached in the Supplementary Material.