Abstract
HIV testing serves as the gateway to HIV prevention and treatment. However, research examining men’s HIV testing behaviors in the Caribbean remains limited. The Andersen Behavioral Model of Health Services Utilization was used to examine factors associated with HIV testing among 7,354 men who participated in the 2012 Demographic and Health Survey conducted in Haiti. Few men (35%) reported having ever been tested for HIV. Logistic regression analyses revealed that HIV testing increased with education and wealth. Marital status was associated with HIV testing, with married men more likely to have been tested (adjusted odds ratio: 2.57, 95% CI [2.07, 3.19]) than unmarried men. Positive attitudes toward people living with HIV, indicated by willing to care for a relative who has HIV/AIDS, was also correlated with higher odds of having been tested (adjusted odds ratio: 1.28, 95% CI [1.08, 1.51]). Men who reported condom use during last sex were more likely to have been tested (odds ratio: 1.58, 95% CI [1.33, 1.88). The findings indicate that HIV testing rates remain low among men in Haiti and more efforts are needed to increase HIV testing among men who are not married, have low level of education, and engage in unprotected sex.
Introduction
Haiti was one of the first countries to report HIV in 1982, and still has one of the highest rates of HIV in the Caribbean and Latin America (Rouzier et al., 2014). Current estimates indicate that Haiti accounts for 55% of all people living with HIV in the Caribbean (UNAIDS, 2014). Despite the political and economic challenges Haiti has faced, successful health promotion, and education strategies advocating early HIV voluntary counseling and testing (VCT) and treatment approaches have contributed to a steady decline in the prevalence rate of adult HIV from 6.2% in 1993 to 2.2% in 2012 (Rouzier et al., 2014). VCT plays a critical role in detecting and linking people newly diagnosed with HIV to treatment and care, with a reported 53,781 people in Haiti receiving antiretroviral therapy as of December 2013 (Rouzier et al., 2014).
Most of the research on HIV among men in the Caribbean has largely focused on men who have sex with men (Caceres, 2002; De Boni, Veloso, & Grinsztejn, 2014). More research and programs for men in the general population in the Caribbean are needed given the low rate of VCT use reported men in other regions (Bwambale, Ssali, Byaruhanga, Kalyango, & Karamagi, 2008; Conserve, Sevilla, Mbwambo, & King, 2012). The reasons men seek VCT services less compared with women is largely due to the emphasis on VCT for women during prenatal care (Katz et al., 2009) and because men are less likely to use health care services (Bertakis, Azari, Helms, Callahan, & Robbins, 2000; Skovdal et al., 2011; Vaidya, Partha, & Karmakar, 2012). Therefore, it is important to understand the contextual factors that influence VCT use among men, to ensure that more men seek VCT, which can help further slow the rates of HIV in Haiti.
A number of health behavior theories have been employed to examine factors associated with health service utilization. The Andersen Behavioral Model of Health Services Utilization is one example of such theories that provides a useful framework to understand the underlying population characteristics (predisposing, enabling, and need for care factors) that influence health service use (Aday & Andersen, 1974). Predisposing characteristics, such as age, education, marital status, and religion, explain the propensity of individuals to use health services. Enabling factors include income, health insurance, familial resources, and refer to the means individuals have available to them for the use of services. Characteristics related to one’s need for care, such as history of illness, are the most immediate predictors of health services use. In 2000, the model was revised as the Gelberg–Andersen–Leake Behavioral Model for Vulnerable Populations to incorporate the unique health care access challenges faced by vulnerable populations (Gelberg, Andersen, & Leake, 2000). Researchers have used the model to examine health care services utilization, such as HIV testing among young women in Trinidad and Tobago (Andrews, 2013), American women in midlife (Wigfall et al., 2011), rural American cocaine users (Wright et al., 2014), and primary care use among HIV-positive Haitian immigrants in Florida (Saint-Jean et al., 2011). A review of the literature revealed that no previous studies have focused on factors associated with HIV testing among men in Haiti. This study aims to fill this gap by using the Andersen Behavioral Model of Health Services Utilization to identify population characteristics associated with HIV testing among men in Haiti.
Method
This study used secondary data from the 2012 Demographic and Health Survey (DHS) conducted in all 10 departments of Haiti from January to June 2012 by the Haitian Childhood Institute (Ministry of Public Health and Population, Haitian Childhood Institute, & ICF International, 2013). The DHS data collection was administered using a two-stage sampling strategy to select nationally representative households for inclusion. Funding for the DHS was provided by the U.S. Agency for International Development, the United Nations Children’s Fund, the Canadian International Development, the Global Fund to Fight AIDS, Tuberculosis, and Malaria, and the United Nations Development Program, and the United Nations Population Funds. A total of 14,287 women aged 15 to 49 years and 9,493 men aged 15 to 59 years were interviewed. For the purpose of this study, the analyses included only 7,354 men who had complete data for the variables of interest. The topic of interest was men’s history of previous HIV testing. Permission to use the data was obtained from Measure DHS. Additional details about survey design, data collection, and management can be found in the 2012 Haiti DHS final report (Cayemittes et al., 2013).
Variables
The main outcome variable was self-reported previous HIV testing (yes/no). The main independent variables were selected based on the components of the Andersen Behavioral Model of Health Services Utilization (Figure 1). Predisposing factors were categorized as follow: age (15-19/20-29/30-39/40 years or older), educational level (none/primary/secondary or higher), religious affiliation (none/Catholic/Protestant/Voudousant), marital status (never in a union/married/living in union/separated or widowed), willing to care for a relative with HIV/AIDS (no/yes), condom use during last sex (no/yes), ever paid for sex (no/yes), alcohol use (every day/from time to time/rarely/never). Attitudes toward gender norms and intimate partner violence were also included: belief that a wife is justified to ask her husband to use a condom (no/yes) and belief that a husband is justified to hit or beat his wife if she refuses to have sex with him (no/yes). The enabling factors were categorized as follow: wealth index (a composite measure of household’s cumulative living standard was calculated using a posteriori and divided into categories poorest/poorer/middle/richer/richest), health insurance (no/yes), and the need for care factor as follows: genital discharge in the past 12 months (no/yes).
The Andersen Behavioral Model of Health Services Utilization.
Statistical Analysis
Analyses were restricted to participants with complete data for the variables of interest. Frequencies and percentages were calculated for all variables. Cross-tabulation analyses between HIV testing, sociodemographic, and behavioral variables was performed. Last, the variables that had statistically significant associations with previous HIV testing (as determined using Rao–Scott’s χ2, p < .05) in the bivariate analyses were all used in the final multiple variable logistic regression model to examine correlates of previous HIV testing. The analyses were conducted using SAS statistical software version 9.3 (SAS Institute Inc., Cary, NC) and accounted for the sampling design. Weights provided by DHS were used in all analyses to account for the sampling design (i.e., unequal selection probabilty). Sampling weights are adjustment factors applied to each case to adjust for differences in probability of selection between cases in a sample, either due to design or happenstance (The DHS Program, n.d.).
Results
The final sample consisted of 7,354 men, with 2,710 (35%) having a primary education (Table 1). Approximately, 3,653 (46%) reported being Catholic and 1,610 (20%) were in the middle wealth index. Half 3,720 (50%) of the men were married and the majority 7,058 (95%) of them did not have health insurance. In the 12 months prior to the survey, 524 (7%) of the men reported they had a genital discharge. Regarding attitudes toward people living with HIV/AIDS, 1,629 (22%) reported they would not be willing to care for a relative if he or she had HIV/AIDS. More than half 4,699 (63%) reported they did not use a condom during their most recent sex activity and only 2,516 (35%) had been tested for HIV.
Demographic Characteristics of Men Included in 2012 Demographic Health Survey.
Note. Unweighted frequencies and weighted percentages.
As reported in Table 2, older men were more likely to have been tested for HIV than men aged 15 to 19 years (p < .0001). The likelihood of having been tested increased with education, being lowest, 173 (15%), among men without any formal education and highest, 1,693 (47%), among men with secondary education or more (p < .0001). Compared with men who reported having no religion, 173 (23%), men who were either Catholic, 1,198 (34%), or Protestant, 1,116 (39%), were more likely to have been tested (p < .0001). Regarding marital status, the highest proportion of men who had been tested were men living with their partners, 172 (42%, p < .0001). When compared with their counterparts who did not report these attributes or behaviors, men who had used a condom during their last sexual encounter, 1,078 (41%, p <.0001), were willing to care for a relative with HIV/AIDS, 2,115 (38%, p < .0001), were more likely to have been tested for HIV).
Characteristics Associated With Previous HIV Testing for Men.
Note. Unweighted frequencies followed by weighted percentages in parentheses.
Logistic Regression Analyses of Previous HIV Testing.
Logistic regression revealed that predisposing factors, such as age, education, religion, and marital status, remained strongly associated with HIV testing after controlling for other variables. Men in the 30- to 39-year-old category were approximately four times (odds ratio [OR]: 3.82, 95% confidence interval [CI: 2.85, 5.12]) more likely to have been tested for HIV than men 15 to 19 years old. Education was positively associated with previous HIV testing, as the odds of having been tested for HIV was 1.63 (95% CI [1.27, 2.08]) for men with a primary education and 3.25 (95% CI [2.48, 4.26]) for men with a secondary or higher education. Catholic and Protestant men were more likely to have been tested than nonreligious men (OR: 1.56, 95% CI [1.22, 1.99]; OR: 1.74, 95% CI [1.36, 2.23], respectively). Middle and richer level men were approximately two times more likely (OR: 1.87, 95% CI [1.50, 2.34]; OR: 1.92, 95% CI [1.52, 2.41], respectively), while richest men were approximately three times (OR: 2.83, 95% CI [2.24, 3.58]) more likely to have been tested than poorest men. Being married increased the odds of having been tested (OR: 2.57, 95% CI [2.07, 3.19]) in comparison with men who had never been married.
Other predisposing factors, such as willingness to care for a relative who has HIV/AIDS and condom use, were also related to previous HIV testing. Men who would be willing to care for a relative with HIV/AIDS were also more likely to have been tested (OR: 1.28, 95% CI [1.08, 1.51]). Having used a condom during last sex with most recent sexual partner was associated with an increased odds of having been tested (OR: 1.58, 95% CI [1.33, 1.88]). The only other variable related to HIV testing was an enabling factor; men who had health insurance were more likely to have been tested (OR: 2.46, 95% CI [1.75, 3.45]).
Discussion
To our knowledge, this is the first article to examine correlates of previous HIV testing among men in Haiti using a nationally representative sample. Although the benefits of early HIV testing are widely recognized, studies examining factors that are associated with HIV testing among men in Haiti and throughout the Caribbean region are limited. This study is a necessary and significant step toward understanding the contextual factors that serve as correlates of HIV testing and developing initiatives for men in the Caribbean.
Overall, 35% of the men in this study had been tested. The HIV testing rate is consistent with other studies conducted in other low-income countries such as Tanzania and South Africa (Conserve et al., 2012; Mhlongo et al., 2013; Snow, Madalane, & Poulsen, 2010). The rate of HIV testing reported in the current study is higher, however, than the 23% of men reported to have received HIV testing in Uganda (Bwambale et al., 2008). Furthermore, in agreement with other investigation (Gage & Ali, 2005; Sambisa, Curtis, & Mishra, 2010; Weiser et al., 2006), this study identified that early middle age (30-39 years), having secondary or higher education, being of Catholic or Protestant religion, greater household wealth, having health insurance, marriage, willingness to care for a relative with AIDS, and recent condom use with a sexual partner were associated with having tested for HIV.
The differences in previous HIV testing among middle-aged men versus younger men may be influenced by the fact that younger men have limited sexual experiences and may be less informed about sexual issues than middle-aged men (Peltzer & Matseke, 2013). Furthermore, the fact that secondary or higher education influenced likelihood of HIV testing is important for the promotion of HIV testing. Previous studies have suggested that increasing men’s education works synergistically to affect HIV testing via prolonged exposure to HIV prevention messages that are more readily available in the school system than in the community (Hargreaves et al., 2008). As a result, the promotion of education may be a potential means to increase HIV testing among men in Haiti.
The association between religion and HIV testing is also encouraging as it suggests that membership in some form of organized religion may be protective against HIV/AIDS (Trinitapoli & Regnerus, 2006) by influencing HIV testing among men (Sambisa et al., 2010). In the past, efforts have been made by USAID to promote HIV testing via religious leaders in Haiti, which may explain partially the relation between religion and HIV testing in the country (Faulkner & Maynard-Tucker, 2002). The current findings also differ from previous studies that suggest that religion does not play a role in men’s HIV testing behavior (Mbago, 2004).
Available evidence has also suggested that socioeconomic status is important in determining whether men seek HIV testing (Gage & Ali, 2005). This study revealed that household wealth, particularly wealth in the upper quintiles and having some health insurance were associated with increased likelihood of having been tested for HIV, a finding that is consistent with previous studies conducted in Uganda and Zimbabwe (Gage & Ali, 2005; Sambisa et al., 2010). Alternatively, lack of financial resources and health insurance may serve as barriers to HIV testing among men in Haiti, a finding that is also consistent with research conducted in Mozambique (Agha, 2012). Overall these findings indicate that HIV prevention strategies to increase uptake of HIV testing should include interventions that target socioeconomic insecurity. For example, the promotion of interventions such as microfinance interventions or cash transfers may increase HIV testing among men in Haiti (Dworkin & Kim, 2009; Heise, Lutz, Ranganathan, & Watts, 2013).
Being married was associated with an increased likelihood of previous HIV testing in the current sample. It may be that marriage and fathering children motivate changes in sexual behavior, which in turn may influence HIV testing among these men (Iwelunmor, Sofolahan-Oladeinde, & Airhihenbuwa, 2014; Siu, Wight, & Seeley, 2014). More research on the association between HIV testing, marriage, and fatherhood among men in Haiti is needed. Furthermore, the personal experience of knowing or caring for someone with HIV/AIDS may be an important driver of being tested for HIV among men in Haiti. Previous studies conducted in sub-Saharan Africa have documented how AIDS illness within family systems or social networks motivates changes in sexual behavior (Agha, 2012; Iwelunmor, Airhihenbuwa, Okoror, Brown, & BeLue, 2006). Knowing an HIV-positive person has also been reported to be associated with HIV testing (deGraft-Johnson, Paz-Soldan, Kasote, & Tsui, 2005). The current findings indicate that involving people living with HIV in intervention programs may be fundamental for effective promotion of HIV testing among men in Haiti. In addition, interventions targeting men’s social networks can also be successful as research from Tanzania reports that men are more likely to have been tested for HIV if one of their friends had encouraged HIV testing (Yamanis et al., 2016).
Finally, HIV testing was higher among men who used condoms during their last sexual encounter. It may be that males who are sexually active and who have multiple partners are more likely to test for HIV and use condoms often (Mhlongo et al., 2013). The alternative is also plausible, where HIV testing and condom use are common among men in monogamous relationships, or even men with multiple partners, not only in order to prevent HIV but also to prevent other sexually transmitted infections or unwanted pregnancy. Other studies have also reported that men who use condoms consistently were more likely to get tested than those who do not (Conserve et al., 2012; Masters et al., 2014). Further research is needed to investigate condom use and HIV testing in Haiti, paying close attention to correlates of HIV testing and consistency of condom use among men.
The findings from the current study should be interpreted in the context of a number of limitations. First, as the DHS is a cross-sectional survey, causality cannot be determined from the findings. Second, the study did not account for some potentially important variables such as willingness to seek VCT in the future. While the DHS is a nationally representative survey, the data are self-reported, thereby potentially introducing social desirability bias in participants’ responses. Finally, the use of country-specific data from Haiti makes it impossible to generalize the findings to the Caribbean.
In light of the declining trend of the HIV/AIDS epidemic in Haiti, efforts have been made to ensure early HIV testing among high-risk groups such as sex workers, and men who have sex with men. However, given that the primary mode of HIV transmission in the Caribbean is via heterosexual contact (Cock & Weiss, 2000; Rouzier et al., 2014), concerted efforts should be made to scale up HIV testing service use among men in the general poluation. These efforts should be accompanied by appropriate frameworks that highlight the historical, social, religious, and cultural factors that influence men’s health and HIV services utilization.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by a training grant from the National Institute of Health—National Institute of Allergy and Infectious Diseases (T32 AI007001).