Article Text

Original research
Socioeconomic inequality and contributors in accepting attitudes toward people living with HIV among adults in Ethiopia from 2005 to 2016: a population-based cross-sectional study
  1. Aklilu Endalamaw1,2,
  2. Charles F Gilks1,
  3. Fentie Ambaw2,
  4. Yibeltal Assefa1
  1. 1School of Public Health, the University of Queensland, Brisbane, Queensland, Australia
  2. 2College of Medicine and Health Sciences, Bahir Dar University, Bahir Dar, Amhara, Ethiopia
  1. Correspondence to Aklilu Endalamaw; yaklilu12{at}gmail.com

Abstract

Introduction The public’s accepting attitude toward people living with HIV is crucial in reducing HIV-related stigma and discrimination, increasing people’s access to HIV service. This study examines the inequalities in accepting attitudes toward people living with HIV in Ethiopia from 2005 to 2016.

Methods This cross-sectional study was based on the 2005, 2011 and 2016 Ethiopian Demographic Health Surveys. A total of 17 075, 28 478 and 25 542 adults were included in the 2005, 2011 and 2016 surveys, respectively. Socioeconomic inequality was investigated using the concentration curve and Erreygers’ concentration index (ECI), which is scaled from −1 (pro-poor) to +1 (pro-rich). The ECI was decomposed to identify the contributors to socioeconomic inequality using generalised linear regression with the logit link function.

Results Accepting attitude toward people living with HIV was 17.9% (95% CI: 16.6%, 19.3%) in 2005, which increased to 33.5% (95% CI: 31.8%, 35.3%) in 2011 and 39.6% (95% CI: 37.6%, 41.9%) in 2016. ECI was 0.342 (p<0.001), 0.436 (p<0.001) and 0.388 (p<0.001), respectively, for 2005, 2011 and 2016. The trend line illustrates socioeconomic inequality seems diverging over time, with an increasing ECI of 0.005 every year (r=0.53; p=0.642; slope=0.005).

Conclusions The current study found that there was pro-rich inequality from 2005 to 2016. People with higher socioeconomic status had a better attitude toward people living with HIV. Comprehensive knowledge about HIV/AIDS, education status, residence, and access to electronic and paper-based media, as well as HIV testing, contribute to a better accepting attitude toward people living with HIV. It is of utmost importance for the country to enhance accepting attitude toward people living with HIV to reduce stigma and discrimination. This requires whole-system response according to the primary healthcare approach toward ending the epidemic of HIV/AIDS in the country.

  • HIV & AIDS
  • Health Equity
  • PUBLIC HEALTH

Data availability statement

Data are available upon reasonable request. Data are available upon reasonable request from Aklilu Endalamaw (yaklilu12@gmail.com).

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

STRENGTHS AND LIMITATIONS OF THIS STUDY

  • This is a population-based study with a large sample size, so its representativeness is valid.

  • Accepting attitude toward people living with HIV was measured based on respondents’ self-report, which may be prone to response set bias.

  • Additionally, the association between independent and dependent variables may not represent cause-and-effect relationships due to its cross-sectional nature.

  • The trend analysis may not be a true reflection of the years from 2005 to 2016 because the survey years were only 2005, 2011 and 2016.

Introduction

The sub-Saharan African region, particularly east and south Africa, is the epicentre of HIV epidemic. It is one of the highest inequalities in the world, caused by structural, social and cultural determinants.1 2 The majority (59%) of the world’s HIV incidence occurred in sub-Saharan African countries in 2021.3 There is also a disparity in HIV/AIDS between sub-Saharan countries. Eswatini was the first country to have 27.9% of its population aged 15–49 years living with HIV in 2021.4 Some other sub-Saharan African countries with high HIV prevalence were Lesotho (23.10%), Botswana (22.20%), Zimbabwe (21.40%), South Africa (13.30%), Namibia (12.70%), Mozambique (12.10%) and Zambia (12.10%) in 2019.5 Ethiopia is one of the countries with the lowest HIV prevalence (0.9% in 2016), but 620 000 people were living with HIV in Ethiopia in 2020,6 and this burden varied by gender7 and geography8 due to differences in HIV/AIDS services.9

Behavioural services play a crucial role in HIV prevention by fostering an accepting attitude toward people living with HIV.10 11 Such an accepting attitude encourages individuals with HIV to engage in social interactions free from the burden of HIV-associated stigma and discrimination. As a result, the general population is more likely to access HIV prevention services and adhere to them upon receiving HIV positive diagnosis.12 For instance, individuals who have a positive attitude toward people living with HIV experience less stigma or hold lower stigmatised attitudes, exhibit safer sexual activities such as condom use, undergo HIV testing, volunteer to disclose their HIV positive status and encourage people living with HIV to start antiretroviral therapy.13 14 These effects have been supported by various HIV control and prevention activities.

There have been several strategies and policies to maximise a favourable attitude in the community or eliminate stigma and discrimination, which continue to be part of human rights issues and are included in health policies.15 For instance, the National Institute of Health considers eradicating stigma and discrimination as one of the top priorities.16 There have been organisations, such as Horizon and Partners, that have been effective in reducing stigma and discrimination by supporting institutions.17 Similarly, in 2017, the Global Partnership was formed with the goal of eliminating stigma and discrimination, recognising that without addressing HIV-related stigma and discrimination, the world will not achieve the goal of ending AIDS as a public health threat by 2030.18 Thus, the Joint United Nations Programme for HIV/AIDS (UNAIDS) has prioritised efforts to combat stigma and discrimination.19

Despite several strategies, many people still have a more stigmatised or less accepting attitude toward people living with HIV. For instance, 95% of sexually active adults had a stigmatised attitude toward people living with HIV in 2016.20 Ethiopia set targets to achieve 90% of adults will develop accepting attitudes toward people living with HIV by the end of 2025. However, looking only the national coverage may mask the underlying disparities between social groups. Hence, UNAIDS calls for examining the hidden gaps in HIV responses.21 The Ethiopian national plan also acknowledged the limitation of crude national performance and emphasises the importance of equity-oriented research by applying equity lenses.22 Therefore, it is necessary to investigate socioeconomic inequalities in healthcare and their contributors.23 Rare studies assessed (in)equality in attitudes toward people living with HIV and demonstrated that inequality in attitudes toward people living with HIV was higher among low-income groups.24 25 None of the previous studies investigated the status, contributors or trend of socioeconomic inequality in attitudes toward people living with HIV.

Thus, this study investigated the level, contributors and trend of socioeconomic inequality in accepting attitudes toward people living with HIV in Ethiopia from 2005 to 2016.

Methods

Data sources, study population and study variables

A population-based cross-sectional study was conducted. We used data from the 2005, 2011 and 2016 Ethiopian Demographic and Health Survey (EDHS). The multistage sampling method used by EDHS involves stratification, clustering and selection of the sample over two stages. Stratification and clustering refer to the sampling design used in data collection. Urban and rural areas were grouped into nine regions and two city administrations as the basis for stratification. Regarding clustering, EDHS employed a two-stage cluster sampling technique. The first cluster is called primary sampling units, which are geographical areas such as villages. Then, households are randomly selected within the chosen clusters.

Adults who gave consent and agreed to participate in the study responded to an interviewer-administered questionnaire. The study population included adults aged 15–49 years old. Adults identified as either permanent residents or visitors who stayed in the selected household the night before the survey were recruited and participated in the EDHS. The total sample size was 71 095: 17 075 in 2005, 28 478 in 2011 and 25 542 in 2016 surveys after handling missed observations and affirmative responses to the HIV/AIDS awareness question. EDHS data quality was assured with the provision of training for data collectors, supervisors and field editors; ongoing supervision, using standardised and translated questionnaires into several local languages; and appropriate software for data entry. Throughout this phase, systematic bias was handled. Proper data management includes appending women’s and men’s data, handling missed observations by missing them completely at random, recoding and conducting variable categories. In handling missing data, participants who did provide complete responses to all questions were excluded from analysis. For example, in 2005, a total of 19 542 adults aged 15–49 years participated in the survey. After excluding participants who did not provide responses to questions about employment status, watching television, reading newspaper and listening to the radio, a total of 18 818 participants remained. No incomplete observations were found in other variables. Among 18 818 respondents, a total of 1743 respondents indicated being unaware of HIV/AIDS, and they were subsequently excluded from the attitude analysis. Ultimately, the final analysis for the 2005 finding included 17 075 adults. In a similar fashion, in 2011, out of the 29 383 adults aged 15–49 years, 29 264 were included in the analysis after excluding participants with incomplete data. Among them, 28 478 respondents provided information about HIV/AIDS and were included in the final analysis. Regarding the 2016 survey, no incomplete observations were found. Therefore, this study included adults aged 15–49 years who had complete data.

Accepting attitude toward people living with HIV was an outcome variable. To measure the outcome variable, respondents who have heard of HIV/AIDS were asked two consecutive questions. These questions are: (1) Would you buy fresh vegetables from a shopkeeper or vendor who had HIV/AIDS? with response options ‘yes’, ‘no’ and ‘do not know’; and (2) Should children with HIV be allowed to attend school with children without HIV? with response options ‘yes’, ‘no’ and ‘do not know’. In DHS 2005 and 2011, the second question was different: should women with HIV be allowed to teach HIV-negative children? with response options ‘yes’, ‘no’ and ‘do not know'. Each question has three options to be answered, including: yes (coded as 1), no (coded as 0) and do not know (also coded as 0). Those who answered ‘yes’ to all questions were considered to have accepting attitude toward people living with HIV, otherwise considered as not having accepting attitude toward people living with HIV. Generally, those who answered positively to two attitude questions were considered to have an accepting attitude toward people living with HIV, while those who responded negatively to either of the questions were coded as having no accepting attitude toward people living with HIV.10 Independent variables are based on the PROGRESS framework group.26 27 The current study incorporated sociodemographic variables, including age category, sex, residence, region, education status, mass media exposure (reading a newspaper, watching television, listening to the radio), marital status, sex of household head, comprehensive knowledge about HIV/AIDS and ever been tested for HIV. To illustrate, the PROGRESS framework stands for social determinants of health such as age, gender, residence, religion, region, marital status, education status and household wealth index rank. Wealth index was used to estimate socioeconomic inequality. The wealth index in DHS is typically estimated using a methodology called principal component analysis using items such as household appliances, ownership of land or property, access to utilities and other indicators of wealth, which are described in the DHS report.28 Comprehensive knowledge about HIV/AIDS was a composite variable from five questions after participants responded ‘yes’ to having ever heard of AIDS; those who answered ‘yes’ for always using a condom and having only one faithful sexual partner can reduce the risk of acquiring HIV; and a healthy-looking person can have HIV; and responding ‘no’ to two questions: Can you get HIV from mosquito bites? and Can you get HIV by sharing food with someone who has AIDS?

Statistical analysis

All the analyses were conducted using STATA V.17.0 (Stata Corp, Texas, USA, 2021). The multistage sampling design effect was controlled in all the analyses. Weighting was applied for frequency distribution; analytical weighting was applied for the concentration curve (CC) and Erreygers’ concentration index (ECI) analysis. Findings were presented using tables and figures. The CC and ECI were employed to assess wealth-based inequality.29 The decomposition of ECI was followed to evaluate the percentage contribution of and factors to socioeconomic inequalities in accepting attitudes toward people living with HIV30 using the generalised linear model (GLM) with a binomial distribution and a logit link function.31 The concentration index (CI) was analysed to see the degree of socioeconomic inequity. It is a common statistical method to assess healthcare inequality when the dependent variable has binary outcomes.29 CI is understood to be twice the area between CC and the line of equality. The value of CI is between −1 and 1; the exact value of −1 and 1 denotes absolute inequality, and the zero CI value represents equitable service distribution. A negative value indicates absolute inequality when there is a disproportionate concentration of accepting attitudes toward people living with HIV among the poor, and a positive value of CI denotes the reverse. It is estimated from the covariance between the accepting attitude toward people living with HIV and the fractional rank of the study participant by the wealth index. This is explained as:

Embedded Image(1)

where CI is the concentration index for accepting attitude toward people living with HIV; μ is the mean of accepting attitude toward people living with HIV; yi is the dummy variable indicating whether i participant has an accepting attitude toward people living with HIV; cov is the covariance with the sampling weights; and Ri is the rank of the individual in the wealth distribution.32

The standard CI may wrongly estimate inequality when the health variable takes a binary outcome.33 The bounds of the CI for a binary outcome are not −1 and 1 because the outcome variable has a binary outcome. The empirical bound of CI could be between m−1+(Embedded Image) and 1−m+(Embedded Image).30 In the analysis based on a large sample size, the value of (Embedded Image) is close to zero. Therefore, the lower bound and the upper bound of CI will be m−1 and 1−m, respectively.32 To accommodate the bounded nature of health variables with a binary outcome (accepting attitude toward people living with HIV in this study), ECI is applied.29 The ECI can be estimated as follows:

Embedded Image(2)

where ECI is the Erreygers’ corrected concentration index, CI is the generalised concentration index and m is the mean of the health variable, the proportion of accepting attitudes toward people living with HIV in this study. The ‘Conindex’ command was used to estimate ECI.34 The ECI was decomposed into the contributions of each determinant to income-related inequalities.

The contribution of each individual factor to the overall wealth-related inequality depends on two things. First, its impact on the level of accepting attitude toward people living with HIV (elasticity), and second, the degree of unequal distribution across different socioeconomic groups (CI). A factor that has a high impact but little variation across different wealth quintiles will contribute minimally to the overall inequality. Finally, we decomposed the CI to obtain the contribution of covariates to the overall wealth-related inequality in accepting attitudes toward people living with HIV using methods suggested by Erreygers and others.30 Because the decomposition of ECI depends on the assumption that healthcare is a linear function of the dependent variable, a suitable statistical analysis is required; this allows categorical variables to be understood in a linear approach. This used a discrete change from 0 to 1 to use marginal or partial effects (dh/dx) estimates as:

Embedded Image(3)

Where βkm is the marginal effects (dh/dx) of each xk and e denotes the error term generated by the linear approximation. The regression-based decomposition was employed to examine the degree to which each covariate contributed to the inequality in accepting attitude toward people living with HIV. The coefficients of contributors were included in the regression-based decomposition analysis. The GLM with a binomial distribution and a logit link function was analysed to get the elasticity, ECI for each variable and contribution. GLM is the best approach to decomposing health variables with binary outcomes as a non-linear regression model.31 Therefore, based on equation 3, to determine absolute contributions, decomposition was conducted as:

Embedded Image

Where Embedded Image stands for the mean of independent variables, βkm represents the partial effect on independent variable xk (dy/dxk), CIk is the CI for determinant xk and GCIe is the generalised CI for the error term.

By applying the generalised linear regression model of the health-dependent variables (accepting attitude toward people living with HIV) for all xk, the elasticity of the health variables was subsequently estimated for each x (xk), which is the sensitivity of accepting attitude toward people living with HIV to changes in the determinants (Embedded Image). It denotes the change in the accepting attitude toward people living with HIV associated with a one-unit change in the independent variables. Then, the CIs are estimated for accepting attitude toward people living with HIV and each independent variable (CIk). Finally, the contribution of each independent variable to the overall CI is estimated by multiplying the elasticity of each determinant by its CI (Embedded Image)CIk. Overall, from the decomposition analysis, we described the elasticity, ECI of each variable category and percentage contributions. Multicollinearity of independent variables was checked by variance inflation factor, and we found that there was no significant collinearity between independent variables.

Patient and public involvement

No patients were involved.

Results

Characteristics of participants and accepting attitude toward people living with HIV

Among the study participants, 44.8% and 63.5% responded ‘yes’ to the question ‘Should a person with AIDS be allowed to continue teaching?’ in 2005 and 2011, respectively. In 2016, 55.3% responded that children with HIV should be allowed to attend school with children without HIV. Additionally, 21.9%, 38.7% and 46.9% agreed to buy vegetables from a vendor with AIDS in 2005, 2011 and 2016, respectively. Accepting attitude toward people living with HIV was 17.9% (95% CI: 16.6%, 19.3%) in 2005, which increased to 33.5% (95% CI: 31.8%, 35.3%) in 2011 and 39.8% (95% CI: 37.6%, 41.9%) in 2016. Given the pooled population (n=71 095), 32.0% (95% CI: 30.5%, 33.6%) of adults aged from 15 to 49 years old exhibited accepting attitude towards people living with HIV. The background characteristics of study participants and the percentage of accepting attitude towards people living with HIV in each social category are presented (table 1). Online supplemental file 1 displayed the pooled result’s frequency distribution across each cell, which is not necessarily to be the sum of the distributions from each year’s table because it is calculated after appending the data from all three study periods (online supplemental file 1).

Table 1

The background characteristics of study participants and percentage of accepting attitude across variables in Ethiopia in 2005, 2011 and 2016

Wealth-related inequality in accepting attitude toward people living with HIV

Concentration curve

Figure 1 displays that a better attitude toward people living with HIV is concentrated among the richest. In all 3 years, the line of equality lies below the line of equality, displaying that the area between the two lines seemed wider in 2005 and became narrower in 2011 and 2016.

Figure 1

The concentration curves present the cumulative percentage of accepting attitude toward people living with HIV on the y-axis against the cumulative percentage of the population ranked by living standards (wealth index), beginning with the poorest and ending with the richest on the x-axis in Ethiopia (2005, 2011, 2016), and combination of three periods.

CI for accepting attitude toward people living with HIV

The magnitude of ECI was positive in each year, with a significant difference between the poor and the rich (p<0.001). ECI was 0.342 (p<0.001), 0.436 (p<0.001) and 0.388 (p<0.001) among adults aged 15–49 years, respectively, for 2005, 2011 and 2016. These indicate adults with a higher socioeconomic status (SES) have developed a better attitude toward people living with HIV than people with a low SES. The trend line illustrates socioeconomic inequality, which seems to be diverging over time with an increasing ECI of 0.005 every year (r=0.53; p=0.642; slope=0.005) (figure 2). The pooled ECI for all 3 years having a sample size of 71 095 was 0.402 (p<0.001).

Figure 2

Trend of Erreygers’ concentration index (ECI) for attitude toward people living with HIV between 2005 and 2016.

Decomposition of ECI

Table 2 illustrates the marginal effect of each covariate for 2005, 2011 and 2016. Gender, residence, education status, religion, household wealth rank, mass media exposure, having ever been tested for HIV and knowledge about HIV were persistent determinants in all 3 years. The individuals with a lower accepting attitude toward people living with HIV were women, rural residents, Catholic or Protestant, were not exposed to mass media (except listening to the radio in 2016), did not undergo HIV testing and did not have comprehensive knowledge about HIV/AIDS.

Table 2

The marginal effects following decomposition of ECI of accepting attitude toward people living with HIV in Ethiopia (2005, 2011 and 2016)

Online supplemental file 2 presents the elasticity, ECI and percentage contribution of covariates to socioeconomic inequality in 2005, 2011 and 2016. The value of elasticity for urban residents was 0.046, 0.075 and 0.085 in 2005, 2011 and 2016, respectively, indicating that a change in place of residence from a rural area to an urban area would result in a 4.6% increment in pro-rich socioeconomic inequality of accepting attitude toward people living with HIV in 2005 that increased to 7.5% and 8.5% in 2011 and 2016, respectively. The positive ECI indicates the contributors are concentrated among adults with a higher SES, favouring the richer to have a better attitude toward people living with HIV. For example, urban residents (ECI=0.547 in 2005, ECI=0.643 in 2011 and 0.601 in 2016). Other covariates that have positive ECI persistently are secondary and higher education status, richer and richest household rank, mass media exposure, those who undergo HIV testing and having comprehensive knowledge about HIV/AIDS. This implies a need to work on these variables to be distributed to the poor groups. Household wealth rank (23.0%), residence (13.2%), education status (10.0%), watching television (4.1%), marital status (3.3%), reading a newspaper (3.2%), geographical region (3.1%), religion (0.04%) and listening to the radio (−2.3%) were contributors to socioeconomic inequality. Also, ever having been tested for HIV (7.5%) and having comprehensive knowledge of HIV/AIDS (7.4%) were contributors over time. Over time, household wealth rank, residence, education, marital status and watching television increased socioeconomic disparity, while religion, reading a newspaper and listening to the radio decreased socioeconomic inequality (online supplemental file 2 and figure 3). The pooled or combined sample size’s ECI, elasticity, absolute contribution value and percentage contribution of each variable are presented in the online supplemental file 3.

Figure 3

Percentage contribution of sociodemographic variables to socioeconomic inequality in attitude toward people living with HIV in Ethiopia between 2005 and 2016.

Discussion

The study analysed socioeconomic inequalities in accepting attitudes toward people living with HIV among the general population in Ethiopia from 2005 to 2016. This study found significant pro-rich inequalities in accepting attitude towards people living with HIV. Household wealth rank, residence and education were the major contributors to socioeconomic-related inequality between 2005 and 2016. Knowledge about HIV/AIDS and HIV testing also contributes to socioeconomic-related inequality in accepting attitude toward people living with HIV.

There was a large socioeconomic-related inequality in the accepting attitude toward people living with HIV. This might be due to people in the richest households having better access to HIV/AIDS services. For instance, people in the highest household rank had better HIV services than people with lower household income.35 36 Moreover, a difference in social interaction between community members could have a contribution. Wealth inequality loosens social cohesion.37 People with a higher wealth status had better social cohesion,38 which reduced stigma (maximise attitude) among people with chronic diseases.39 Thus, the higher the SES, the more social cohesiveness leads to better HIV service utilisation,40 which can improve attitude toward people living with HIV. Socioeconomic-related inequality was higher between women and men, but women had better comprehensive knowledge persistently. This might be due to the large gap between men and women in education, media exposure, internet usage, employment status and household head.8 The contribution of gender to socioeconomic inequality declined over time.

Likewise, those in urban areas had a better attitude toward people living with HIV than rural residents. In the Ethiopian context, many people with low income are situated in rural settings and are struggling to fulfil necessities (eg, food), which takes over their health-seeking behaviour.41 Additionally, the rural resident population rarely gets media exposure (unable to buy a newspaper, radio or television)8 and could not attend the health system frequently due to several barriers (eg, transportation).42 Moreover, those who had frequent exposure to watching television and reading newspapers had a better attitude toward people living with HIV in the current study. All these lead to wide and deep inequalities among disadvantaged groups of a population. Furthermore, the wealthiest are more likely to reside in urban areas, attend better-resourced schools or colleges, and give more time to health advice due to the healthcare infrastructure being better in urban areas.43 Attending healthcare allows them to acquire more information and knowledge that change their attitude. Similarly, those who had ever been tested for HIV and had more knowledge about HIV/AIDS had a better attitude toward people living with HIV in the current study. Therefore, effective behavioural interventions should better be implemented for rural residents through house-to-house walking services or using social or mass gathering. In Ethiopia, for instance, education at the coffee ceremony increased healthcare service uptake.44

Socioeconomic inequality in accepting attitude toward people living with HIV was observed between individuals with above primary education and non-educated backgrounds. Education status caused a large difference in living standards.45 Accordingly, in this study, education status shares a large contribution to socioeconomic inequality besides the overall healthcare disparity.46 47 An intervention that considers level of education could be more effective because level of understanding, perception and the way to see things are different depending on education status. People who are engaging in HIV prevention services and HIV/AIDS strategic documents should consider this difference. However, in the HIV/AIDS national strategic plan draft for Ethiopia from 2021 to 2025, the plan and strategies to accomplish interventions did not accommodate and appreciate the presence of attitude differences based on education status.22

Other contributors to socioeconomic inequality were geographical region, religion, marital status and household headship. This variation between different social groups reflects the underlying structural and health service inequalities in behavioural interventions. It is appreciated that widowed and divorced men and women, and people in high HIV/AIDS risk geographical areas have priority to get HIV prevention services, and it is planned to engage religious fathers in HIV prevention services.22 It would be better to consider and deliver more behavioural interventions to women who live in households with male heads.

The findings from this study have crucial implications for the Ethiopian National HIV/AIDS Prevention and Control Plan from 2021 to 2025.22 Moreover, as Ethiopia is a UNAIDS-targeted country, the sustainable development plan to reach ‘zero HIV-related stigma and discrimination’ will not be achieved by 2030 with the current disparities in attitude toward people living with HIV in Ethiopia.48 Hence, equity-oriented, inclusive and persistent interventions may be important. This requires multipurpose strategical, tactical and targeted approaches to enhance attitude toward people living with HIV. In general, culturally sensitive interventions are preferable.49 Education, connections between people, peer service and continuous advocacy may increase an individual’s attitude toward people living with HIV.50 Furthermore, there is evidence of the effectiveness of social marketing and social contact in reducing stigma and discrimination.51 Another key intervention to increasing accepting attitudes toward people living with HIV might include strengthening community-led stigma reduction efforts with people living with HIV leading the effort. The work by Pretorius et al found that community-based stigma reduction can reduce stigma and discrimination.52

Limitations

When interpreting this study’s results, there are relevant limitations. First, the study’s cross-sectional nature reflected that the association between independent and dependent variables may not represent cause-and-effect relationships. This requires further follow-up study to exactly determine whether SES causes difference in accepting attitude toward people living with HIV. Second, while this study represents the largest ever study of Ethiopian population on socioeconomic analysis in accepting attitude toward people living with HIV, it is worth noting that data were collected by survey based on respondents’ self-report. Consequently, we cannot assure the absence of response set bias. Moreover, the trend analysis may not be a true reflection of the years from 2005 to 2016 because the survey years were only 2005, 2011 and 2016. Finally, one of the questions used to operationalise accepting attitude toward people living with HIV was different between 2005, 2011 and 2016. The question ‘would you buy fresh vegetables from a shopkeeper or vendor who had HIV/AIDS?’ was applied for all the three surveys. However, the second question used in 2005 and 2011 (‘should women with HIV be allowed to teach HIV-negative children?’) was replaced by ‘should children with HIV be allowed to attend school with children without HIV?’ in 2016.

Conclusions

The status of accepting attitude toward people living with HIV is alarmingly heterogeneous among social determinants over time. There is a large socioeconomic-related inequality in accepting attitude toward people living with HIV and there was pro-rich inequality over time. Our trend analysis showed that inequality is persistent and diverging over time. Comprehensive knowledge about HIV/AIDS, education status, residence, and access to electronic and paper-based media as well as HIV testing contribute to a better accepting attitude toward people living with HIV. It is of utmost importance for the country to enhance accepting attitude toward people living with HIV to reduce stigma and discrimination. This necessitates a comprehensive response from the government and society as a whole. Adopting this approach will facilitate the progress path toward ending the HIV/AIDS epidemic in the country.

Data availability statement

Data are available upon reasonable request. Data are available upon reasonable request from Aklilu Endalamaw (yaklilu12@gmail.com).

Ethics statements

Patient consent for publication

Ethics approval

Ethical approval was obtained from DHS (https://dhsprogram.com/). The University of Queensland Institutional Ethical Review Board also exempted the ethical issue of this research (approval project number: 2022/HE001760). Other ethical issues, like consent and confidentiality, were responsibly managed during data collection by the EDHS staff.

Acknowledgments

The authors acknowledge the University of Queensland Australia for supporting AE through the University of Queensland Research and Training Scholarship scheme. Additionally, they would like to thank the Demographic Health Survey for giving the dataset to conduct this research.

References

Supplementary materials

Footnotes

  • Contributors AE conceptualised the project and conducted ethical request process. AE analysed the data, wrote the first draft and conducted subsequent revisions. AE accepts full responsibility for the work and/or the conduct of the study, had access to the data, and controlled the decision to publish. YA supervised the whole research process, checked data and edition of the manuscript, and revised the manuscript. CFG supervised the whole research process and revised the manuscript. FAG revised the manuscript. AE, YA, CFG and FA approved the final manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.