Condom Use During Last Sexual Contact and Last 30 Days in Two Samples of Caribbean Military Personnel

Abstract

Condom use during last sexual contact is a survey measure that may be used to inform monitoring and evaluation indicators of recent condom use in populations at risk for HIV infection and other sexually transmitted infections, such as the uniformed services. The authors tested for differences in condom use measures that were fielded within separate Biological and Behavioral Surveillance Surveys conducted in the armed forces of two separate nations: the Dominican Republic and Belize. Both surveys included measures of condom use during last sexual contact with specified partners and both surveys included the Risk Behavior Assessment (RBA), which measures specific sexual acts and condom use frequency during a specified time period. In both samples, more than 40% of respondents who reported condom use during last sexual contact with a regular partner also reported engaging in unprotected sex when screened with the RBA. Furthermore, more than 60% of respondents who reported condom use during last sexual encounter with a commercial sex worker also reported engaging in unprotected sex when screened with the RBA. The results carry implications for monitoring and evaluation indicators of large-scale HIV prevention programs. The authors recommend that, when feasible, more in-depth instruments such as the RBA be considered to measure recent condom use in populations of uniformed services personnel.

Introduction

Biological and Behavioral Surveillance Surveys (BBSS) of HIV infection can provide policy makers with information for evaluation, program planning, and advocacy (UNAIDS/WHO Working Group, 2004). BBSS endeavors can provide HIV prevention efforts with important information regarding the prevalence of HIV as well as the prevalence and correlates of risk behaviors that are targeted by HIV prevention efforts. Furthermore, BBSS endeavors can supply programs with indicators of particular risk behaviors in populations and can illustrate how those risk behaviors change over time. Oftentimes, the behavioral component of a BBSS endeavor involves questionnaires that include measures of sexual risk behavior, including whether a condom was used during last sexual contact with a specified partner (Anand et al., 2009; Anastario et al., 2011; Anastario, Tavarez, & Chun, 2010; Tavarez, Chun, & Anastario, 2011). Condom use during last sexual contact can be viewed as an efficient measure of recent condom use in a population, as it does not require ample survey space and may be easier for respondents to recall. As this measure is one method of indicating the relative prevalence of sexual risk behavior in a population, it may be used as a program monitoring and evaluation (M&E) indicator to measure changes over time in a population that is being targeted with HIV prevention efforts. Condom use during last sexual contact is often used to inform M&E indicators of recent condom use in populations at risk for HIV and other sexually transmitted infections (STIs), such as the uniformed services. This measure may be misleading, as condom used during last sexual contact does not capture the variation in condom use consistency during a recent period of time.

The uniformed services are increasingly being recognized as an at-risk population for HIV/AIDS. This is in part due to the ways in which security institutions and contextual insecurity shapes patterns of HIV risk and that HIV prevention efforts have been poorly integrated with components of security institutions, such as security sector reforms, humanitarian assistance, and peace-keeping and peace-building efforts (de Waal, Klot, & Mahajan, 2009). In this brief report, we test for differences in condom use measures using data derived from identical measures fielded in two BBSS endeavors conducted with the armed forces of two Caribbean military populations: the Dominican Republic and Belize. Both populations comprise a majority of male personnel, both surveys included measures of condom use during last sexual encounter with a commercial sex worker (CSW) and regular sexual partner, and both surveys included the Risk Behavior Assessment (RBA; National Institute on Drug Abuse, 1993), an instrument measuring the frequency of specific sexual acts and corresponding condom use relative to a discrete time period—in this case, the 30 days preceding the interview. This analysis has direct implications for standard measures used in the monitoring and evaluation of HIV prevention programs.

Method

Belize

Data were drawn from a recent study conducted with the Belize Defense Force in which serological data and structured behavioral interviews were collected from a representative sample of defense force personnel in 2009-2010. The larger study has been described in detail elsewhere (Anastario et al., 2011). A total of 334 audio computer-assisted self-interviews were completed by male and female personnel. As this study is concerned with addressing measures of male condom use, and given the relatively low prevalence of females in each population, the analytic sample excluded female personnel to yield a sample size of 304 male personnel.

Dominican Republic

Data were also drawn from a BBSS conducted in the armed forces of the Dominican Republic between May and June 2010. Active duty male and female personnel who were at least 18 years of age were randomly sampled using a systematic sampling strategy to obtain participants for the study. A total of 653 male and female personnel completed questionnaires across 14 bases nationwide. The analytic sample consisted of 607 male personnel.

Risk Behavior Assessment

The RBA was used to measure sexual and HIV risk behaviors (National Institute on Drug Abuse, 1993). The RBA items included 30-day recall measures of insertive vaginal sex, insertive and receptive oral sex, and insertive and receptive anal sex. Respondents reporting ≥1 instance of any sexual behavior during the past 30 days prompted the question, “Of these times, how many times did you use a condom, latex, or other barrier protection?” For the purposes of this analysis, inconsistent condom use was defined for individuals who did not use condoms during every act of insertive vaginal sex, insertive anal sex, and receptive anal sex during the previous 30 days.

Condom Use During Last Sex

In addition to the RBA, both surveys included measures that screened participants for whether they had a regular sexual partner, sex with CSWs during the past 12 months, and whether a condom was used during the last sexual contact for each specified partner. Respondents who did not have a regular sex partner and who did not report sex with a CSW were treated as missing observations for the purpose of this analysis.

Measurement discordance was defined for individuals who reported using a condom during last sexual contact, but who screened positive for inconsistent condom use using the RBA. In the context of this study, measurement discordance is an important variable as it indicates the degree of “disagreement” that the condom use during last sexual contact measure provides as an overall indicator of recent condom use in a population.

Data Analysis

Data were analyzed using STATA 10 statistical software (Stata Corporation, 2007). We examined distributions of self-reported condom use during last sexual contact and during the past 30 days by country. Post hoc analyses were conducted to test for relationships between measurement discordance and potential demographic covariates, and two-tailed t tests and logistic regression were used to examine these relationships.

Results

Discordant Measurement for Condom Use With a Regular Sexual Partner

Dominican Republic

In the Dominican Republic sample (N = 596), 71.5% of the respondents reported having a regular sexual partner. Among respondents with a regular sexual partner (n = 426), 20.9% reported using a condom during last sexual contact with their regular sexual partner, of which 40.5% reported inconsistent condom use during the 30 days preceding the interview (Figure 1).

Figure 1.

Self-reported condom use during last sexual contact with a regular sexual partner/commercial sex worker (CSW) and inconsistent condom use during the previous 30 days among men in the armed forces of Belize and the Dominican Republic

Belize

In the Belize sample (N = 299), 84.6% reported a regular sexual partner. Among respondents with a regular sexual partner (n = 253), 25.3% reported using a condom during last sexual contact, of which 43.8% reported inconsistent condom use during the 30 days preceding the interview.

Discordant Measurement for Condom Use With a CSW

Dominican Republic

Among men responding to the question regarding whether they had sex with a CSW during the past 12 months (N = 597), 30.8% reported having sex with a CSW during the past 12 months. Among respondents who had ever had sex with a CSW (n = 184), 84.4% reported using a condom during last sexual contact with a CSW, of which 49.3% reported inconsistent condom use during the 30 days preceding the interview (Figure 1).

Belize

Among men responding to the question regarding whether they had sex with a CSW during the past 12 months (N = 304), 31.6% reported sex with a CSW. Among respondents who had sex with a CSW (n = 96), 77.8% reported using a condom during last sexual contact with a CSW, of which 61% reported inconsistent condom use during the 30 days preceding the interview.

Discussion

We documented a high prevalence of measurement discordance regarding recent condom use among male military personnel in two nations in the Caribbean region. We documented that condom use during last sex did not capture the variation evident for inconsistent condom use. Condom use during last sexual encounter may not be the most representative indicator of recent or consistent condom use in a population at risk for HIV and other STIs. Differences in recall periods (last time vs. last 30 days) and partner specificity allowed us to examine the possibility of discordantly qualifying a respondent as a condom user when the respondent was not a consistent condom user, indicating discordance in the measurement of recent condom use. However, there were a negligible number of respondents who reported no condom use during last sexual encounter and consistent condom use during the last 30 days (not reported), further indicating the problematic nature of asking respondents about last condom use with a specified partner. Furthermore, a post hoc analysis revealed that respondents’ likelihood of being discordantly measured was significantly associated with demographic factors, such as single respondents being less likely to be discordantly measured and recently deployed respondents being more likely to be discordantly measured—indicating that subgroups may be inappropriately documented as successfully using condoms in M&E efforts. It is therefore the recommendation of the authors that measures of condom use in M&E efforts be closely scrutinized and that, when feasible, more in-depth instruments be used.