HIV infection,sexual risk behaviour and condom use in the Belize defense force

Abstract

This study is the first Biological and Behavioral Surveillance Survey to be conducted among personnel in the Belize Defense Force. The purpose of the study was to understand the prevalence of HIV infection and risk behaviours, and to identify key correlates of sexual risk behaviours. A representative sample of personnel underwent serological testing and an Audio Computer-Assisted Self Interview. Of those sampled, 351 completed a blood test and 334 completed a behavioural interview. The prevalence of HIV was 1.14%. Twelve percent had ever reported being diagnosed with a sexually transmitted infection (STI) or screened positive for HIV infection. The odds of ever having an STI/HIV were higher among those who had less education, those who had sex with a commercial sex worker (CSW), those who ever engaged in receptive anal sex and those with post-traumatic stress disorder (PTSD). Alcohol abuse and PTSD were prevalent and associated with HIV risk behaviours. These results are being used to inform current prevention efforts.

Keywords

sexual risk behaviour military personnel condom use mental health HIV Belize

INTRODUCTION

It is difficult to conduct rigorous Biological and Behavioral Surveillance Surveys (BBSS) of military personnel given the limitations on researcher access to military bases, concerns about security and increased sensitivity regarding data and reporting. For this reason, HIV programmes are often implemented in military populations without baseline data. When studied, military personnel have been documented as frequently engaging in sexual risk behaviour, such as transactional sex and inconsistently using condoms with casual sex partners during peacetime and deployment.^1,2 In Central America, military personnel are increasingly being treated as most at-risk population for HIV infection.

Belize is a Central American country where the prevalence of HIV has been estimated at 2.1% (range 1.2–3.1%) in the general population,³ with heterosexual contact as the predominant modality of transmission in adults.⁴ The prevalence of HIV, risk behaviours and condom use has never before been studied in the Belize Defense Force (BDF). There are approximately 1000 active duty personnel stationed at four bases throughout the nation, and their primary task is to defend the nation's borders and provide support to civil authorities. Here, we report on the results of the first BBSS to ever be conducted in the BDF.

MATERIALS AND METHODS

Serological data and structured behavioural interviews were collected from a representative sample of BDF personnel. We used the UNAIDS national HIV infection estimate of 2.1% to inform the power analysis.³ To estimate the unknown value of the infection rate with 95% confidence and precision = 0.015, we aimed to conduct serological tests on 351 military personnel to achieve adequate statistical power.

Participant selection

The study population was comprised of all active duty BDF personnel. This included male and female uniformed service members who were at least 18 years of age. To collect serological and behavioural data, we used a systematic random sampling strategy to select participants from an alphabetized list of active duty BDF personnel. A computer generated a random number to identify the index subject from which to begin selecting potential participants using a sampling interval. The Western Institutional Review Board and one local ethics review board convened by the medical corps of the BDF approved this study.

All study participants provided informed consent, in private, prior to participation. Following consent, participants were asked to provide their names and contact information to the study team. Each respondent was given a numbered card which was used to code the blood sample and questionnaire data. Respondents were provided optional refreshments, in the form of food and beverage, for participating in the study. All serological testing and interviews took place in private locations on each base, and participants were administered the behavioural questionnaire using Audio-Computer Assisted Self Interview (ACASI). Of the 362 respondents who were approached to participate in this study, 351 consented to testing and participation (97% response rate). Among the 351 subjects who took an HIV test, 334 (95%) completed an ACASI.

HIV testing

Subjects were assigned a number that linked their blood sample with their questionnaire data, and which corresponded to the contact information held by the principal investigator in a password-protected file. Approximately 5 mL of blood was drawn per subject. Blood samples were centrifuged prior to testing, and then refrigerated. According to the national HIV-testing protocol in Belize, every positive HIV-1/2 DETERMINE test (Abbott Diagnostics, Abbott Pk, IL, USA) was followed by a Reveal Rapid HIV Antibody Test (Medmira Inc, Halifax, Canada) test, which was then followed by an ELISA if the Medmira test was reactive. For quality control purposes, every 10th negative Abbott test was also followed by a Medmira test to confirm test specificity.

Structured interviews

ACASI was used as the modality by which to collect quantifiable behavioural data. The interview was administered using QDS^TM Version 2.6 Software installed on Windows platform Netbook computers. The use of ACASI can be more efficient than paper-based surveys in that it eliminates the cost of printing and data entry, and may be a more effective method for collecting highly sensitive information,⁵ making it a suitable questionnaire administration modality to use in the context of BBSS for military personnel. Respondents were provided headphones to self-administer the interview. Questions were posed to the respondent in text on the computer screen and through audio via headphones, with a 500-millisecond pause programmed between questions/screens. Respondents were given the option of skipping any question on the ACASI, as well as to return to a previous question. If respondents encountered difficulties with a question and desired assistance, a study team member provided assistance.

Measures

The ACASI included measurements of respondent characteristics and behaviour in domains typical of a knowledge, attitudes and practices (KAP) survey with additional modules measuring sexual coercion and mental health.

Basic knowledge on modes of HIV transmission, characteristics of HIV infection, knowledge about HIV testing and knowledge of condoms were measured using a series of closed-ended questions with the response set: yes; no; I don't know.

Measurement for HIV risk behaviours was drawn from the Risk Behavior Assessment (RBA)⁶ and The Client Diagnostic Questionnaire (CDQ).⁷ To measure sexual risk behaviour, the RBA items included 30-day recall measures of penetrative vaginal sex, penetrative and receptive oral sex, and penetrative and receptive anal sex for men and receptive vaginal, receptive oral and receptive anal sex for women. Respondents who screened positive for a given behaviour during the past 30 days were then asked ‘of these times, how many times did you use a condom, latex or other barrier protection?’ Consistent condom use was defined as using condoms during every instance of reported vaginal or anal sex during the past 30 days. As part of the RBA, respondents were also asked how many days during the past 30 days they engaged in sexual activity as well as how many partners they had, and whether they engaged in transactional sexual activities. In addition to screening for transactional sex, respondents were asked whether they had ever had sex with a commercial sex worker (CSW), and with how many CSWs they had sex during the past 12 months. For respondents screening positive to sex with CSWs, condom use during last sexual contact was also measured. Respondents who had sex with more than one partner during the past 30 days were classified as having multiple concurrent partners (MCPs).

Self-reporting a previous sexually transmitted infection (STI) diagnosis is an important behaviour to examine in a cross-sectional study, as it represents the outcome of previous and/or ongoing engagement in sexual risk behaviour.^8,9 Subjects were asked a series of separate questions, by STI, to indicate if a doctor or a nurse had ever told them that they had hepatitis B and C, gonorrhoea, syphilis, genital warts, chlamydia, genital herpes or trichomoniasis. Respondents ever screening positive for any of the individual STIs listed were coded as having screened positive for an STI. This variable was combined with those whose tested HIV-positive to form a self-reported STI/HIV outcome variable.

The use of sexual coercion is relevant to sexual risk as it inhibits partner communication and places both the victim and perpetrator of abuse at risk for acquiring an STI. The Sexual Coercion Scale is a seven-item instrument derived from the revised Conflict Tactics Scale, which measures the frequency of three levels of coercion (insistence, threats of force and actual force) during sexual activities in the past year. Respondents were classified as having used sexual coercion in the past year if they screened positive for at least one of the behaviours measured.¹⁰

Post-traumatic stress disorder (PTSD) is an anxiety disorder, which arises from exposure to overwhelmingly stressful events, with symptoms of re-experiencing, avoidance and numbing and arousal.¹¹ We measured PTSD using Breslau's seven-item screen for PTSD.¹² Respondents were primed to recall a traumatic event using the PTSD inventory from the CDQ.⁷ We used a cut-off point of four to classify a respondent as having probable PTSD.¹²

Depression symptoms were assessed with the Patient Health Questionnaire-2 (PHQ-2).¹³ A 30-day recall period was used.

We measured alcohol abuse using the Rapid Alcohol Problems Screen 4-Quantity Frequency (RAPS4-QF), a short screening instrument which has shown high sensitivity to alcohol abuse.¹⁴ We classified respondents as having probable alcohol abuse for a score of 1 or greater on the 6-item RAPS4-QF, and probable alcohol dependency for a score of 1 or greater on the 4-item RAPS4.¹⁴

Data analysis

Quantitative data were analysed using STATA 10 software.¹⁵ Subject identifiers were scrambled after linking the serological and behavioural data-sets.

Unless specified, sample sizes remained floating by variable analysed. To characterize the prevalence of HIV and risk behaviours, means and frequencies were analysed to report distributions of items and behaviours. Confidence intervals (CI) for the HIV prevalence estimate were calculated using binomial normal theory. To test for associations with behavioural outcomes, two-tailed t-tests were used to test for mean differences by group for variables with continuous distributions, and the Pearson's chi-squared statistic was used to test for differences by group for variables with dichotomous distributions. The general linear model was used to control for potential confounders in analysing outcome variables with continuous distributions, and logistic regression was used to estimate odds ratios (ORs) and to control for potential confounders for outcome variables with dichotomous distributions. Analyses of consistent condom use during vaginal or anal sex were restricted to respondents who reported vaginal or anal sexual activity during the past 30 days.

RESULTS

Background characteristics

The most common ethnic background with which respondents identified was Mayan (non-Ketchi), followed by Creole and Mestizo (Table 1). The majority of respondents had completed high school (54%), and 65% were married or in common law unions. Respondents were an average age of 29 years with 1.5 children, and had served an average of eight years in the military and approximately four years at their current posts. Eighty-nine percent of respondents were enlisted (Table 1).

Table 1

Respondent characteristics by HIV/self-reported STI status and condom use, n = 334

			In sexually active subjects, n = 274
	Never reported an STI	Reported an STI^†	Inconsistent condom users	Consistent condom users
Characteristics	(n = 293)	(n = 41)	(n = 223)	(n = 51)
Male (%)	91.1	90.2	93.3	88.2
Age, (mean, (SD))	28.3 (6.4)**	31.1 (6.2)**	28.8 (6.5)	28.6 (6.9)
Less than high-school education (%)	34.5	63.4	36.8	41.2
Marital Status (%)
Married	37.5	41.5	42.6**	21.6**
Common law	27.0	26.8	30.8	31.4
Single	25.3	14.6	17.5**	35.3**
Enlisted	89.1	87.5	90.1	94.1
Number of years in the military, (mean, (SD))	7.5 (5.9)*	9.7 (5.9)*	8.1 (5.9)	7.3 (5.8)
Branch of the military (%)
Special Boating Unit	4.8	2.6	4.1	10.0
Air Wing Unit	3.4	2.6	3.6	6.0
Service and Support Battalion	29.2*	48.7*	29.3	30.0
Battalion 1 or 2	58.4*	43.6*	58.1	54.0
Force head quarters	4.1	2.6	5.0	0.0
Ever been on foreign deployment, %	13.6	19.5	16.1	11.8
Number of months deployed, n = 48 (mean, (SD))	5.6 (4.6)	5.8 (4.5)	5.1 (4.1)	3.8 (3.7)
Number of months at current post, (mean, (SD))	19.0 (35.1)	24.0 (48.6)	16.8 (30.7)	23.5 (45.9)
Number of biological children, (mean, (SD))	1.5 (1.6)	1.7 (1.4)	1.6 (1.6)	1.2 (1.2)
Christian/Catholic religious identification, %	81.2**	61.0**	26.9	27.5
Ethnic background (%)
Creole	17.1	29.3	17.0	21.6
Maya	27.3	24.4	28.3	25.5
Ketchi	16.4**	0.0**	12.6	21.6
Garifuna	16.7	26.8	18.8	13.7
Mestizo	18.8	17.1	18.8	17.7
Latin	2.1	2.4	2.7	0.0
Other	1.7	0.0	1.8	0.0

STI = sexually transmitted infection

^†Includes respondents who tested positive for HIV infection at the time of the study

*P < 0.05; **P < 0.01; ***P < 0.001

HIV, self-reported STIs and risk behaviour

The prevalence of HIV infection was 1.14% (95% CI 0.029–2.25%) (Table 2). Among the four respondents testing positive, two reported having been aware that they were infected with HIV. Seventy-six percent of respondents had ever had an HIV test prior to this study. Among those who had been previously tested, 50% reported receiving pretest counselling, and 47% reported receiving post-test counselling. Problems in overall knowledge included 28% thinking that water-based lubricants made a condom less effective, and 27% not knowing the answer to the question. Twenty-three percent thought that an HIV-infected woman could not infect other people with HIV while she is menstruating, and 21% did not know the answer to the question. Ninety-two percent of respondents reported that they would use condoms if they were available for free at work. Seventy-two percent reported that they had access to free condoms. Forty-three percent of respondents reported that they always carry condoms. Of respondents reporting vaginal or anal sex during the past 30 days (n = 274), 19% consistently used condoms.

Table 2

HIV prevalence and risk behaviours

	Frequency (%)
HIV positive, n = 351	1.14
Ever told by a doctor/nurse they had an STI*, n = 334	11.7
Sex with a CSW in past 12 months, n = 334	31.1
Did not use a condom last time they had sex with a CSW^†	26.3
Ever had sex with a CSW, n = 327	55.1
Did not use a condom last time they had sex with a CSW^‡	18.8
Did not know sexual partner's HIV status at last sexual encounter, n = 311	50.8
More than one sexual partner during the past 12 months, n = 314	52.2
More than one sexual partner during the past 30 days, n = 293	25.9
Sex partner in last 30 days has multiple sexual partners, n = 305	35.1
Sexual partner(s) likely to have been injecting drug users or shooters, n = 290	4.5
Have had sex while drunk or high in the last 30 days, n = 321	37.7
Inconsistent condom use^§	54.2
Injected drug or popped skin with a needle in the past 30 days, n = 330	0.0
Ever injected drug or popped skin with a needle, n = 329	0.0
Have given sex to get money, drugs or other goods in the past 30 days, n = 315	4.1
Engaged in sexual coercion during the past 12 months, n = 330	44.2
Have gotten a permanent tattoo, n = 329	69.9

CSW= commercial sex worker

*Derived from individual questions for hepatitis B and C, gonorrhoea, syphilis, genital warts, chlamydia, genital herpes, trichomoniasis

^†Among those who reported sex with CSW during the previous 12 months, n = 104

^‡Among those who ever reported sex with a CSW, n = 180

^§Among those who reported having sex while drunk or high, n = 121

Twelve percent of respondents reported ever having an STI or screened positive for HIV infection (Table 1). Seven percent of respondents had ever been told by a doctor/nurse that they had gonorrhoea, 2.5% had been told that they had hepatitis B, 2.4% had been told they had genital herpes, 2.1% had been told that they had genital warts and 1.8% had been told that they had syphilis. Consistent condom use was not associated with a previous STI diagnosis (OR = 1.3, 95% CI 0.51–3.1, P = 0.62).

The prevalence of sex with CSWs was high, with 26% not using condoms during last sexual contact with a CSW (Table 2). Inconsistent condom use while having sex drunk or high was a notable problem (Table 2). Fifty-four percent of those reporting that a sexual partner was likely to have multiple partners also had sex with a CSW during the past 12 months (χ ² = 36.8, P < 0.001). There was no report of injection drug use; however, there was evidence of sexual partners likely being injecting drug users (Table 2).

For use of sexual coercion among men during the past 12 months, 46.3% reported using insistence to have sex, 1.7% reporting using threats and 1.7% reported using physical force to make a partner have sex.

Three percent of women reported having ever had bacterial vaginosis. Seventy percent of women reported ever having douched, and of those, 70% reported having douched before or after having vaginal sex. While only four women were classified with HIV/ever having an STI, all four reported having ever douched before or after having vaginal sex.

A high-risk subsample of respondents who reported MCPs was isolated for further analysis (76 respondents). Within men, the prevalence of inconsistent condom use during penetrative vaginal sex was high (81%), and a higher rate of penetrative anal sex was observed in comparison to those without MCPs (19% versus 2%, χ ² = 23.6, P < 0.001). Among women in the high-risk group, the prevalence of receptive vaginal sex was 86% with 100% showing inconsistent condom use.

Correlates of a self-reported STI

Demographic correlates of ever reporting an STI included greater age, a greater number of years in the military, branch of service, non-Christian religious identification and non-Ketchi ethnic identification (Table 1). Multivariable models were thus adjusted for the effects of age, number of years in the military, branch of military service (service and support battalion or battalion 1/2), religious identification (Christian/Catholic versus non-Christian/Catholic) and Ketchi ethnic identification.

Sexual behavioural correlates of ever reporting an STI included engaging in penetrative oral sex and penetrative anal sex (Table 3). After controlling for potential confounders, male respondents who engaged in penetrative oral sex were over twice as likely to have self-reported STI/HIV (adjusted odds ratio [AOR] = 2.2, 95% CI 1.02–4.9, P < 0.05), and male respondents who engaged in penetrative anal sex were over four times more likely to have a self-reported STI/HIV (AOR = 4.1, 95% CI 1.2–14.4, P < 0.05). While not statistically significant, a trend level association was detected for respondents who reported having sex with a CSW during the past 12 months (AOR = 1.9, 95% CI 0.9–3.9, P = 0.1).

Table 3

Respondent behaviour by HIV/self-reported STI status, n = 334

Characteristics	Never reported an STI (n = 293)	Reported an STI (n = 41)	OR (95% CI)	Adjusted OR^† (95% CI)
Sex with a CSW during the past 12 months	30.0%	39.0%	1.49 (0.76–2.9)	1.86 (0.90–3.9)
Sexual partner has multiple concurrent partners	33.2%	48.7%	1.8 (0.94–3.5)	1.8 (0.87–3.6)
Sex with multiple concurrent partners during the past 30 days	26.3%	23.5%	0.86 (0.37–2.0)	0.77 (0.31–1.9)
Knew sexual partner's HIV status	46.1%	43.9%	0.92 (0.47–1.8)	0.73 (0.36–1.5)
Had sex while drunk or high during the past 30 days	36.9%	43.6%	1.3 (0.67–2.6)	1.4 (0.67–2.9)
Used sexual coercion	46.9%	47.5%	1.02 (0.52–2.0)	1.1 (0.56–2.3)
Inconsistent condom use during the past 30 days
Penetrative vaginal sex	66.7%	73.5%	1.39 (0.62–3.1)	1.01 (0.43–2.4)
Receptive vaginal sex	–	–	–	–
Penetrative oral sex	26.8%	47.1%	2.4 (1.2–5.0)*	2.2 (1.02–4.9)*
Receptive oral sex	1.9%	0.0%
Penetrative anal sex	3.1%	12.2%	4.4 (1.4–13.8)*	4.1 (1.2–14.4)*
Receptive anal sex	0.4%	2.8%	7.7 (0.47–126.1)	8.1 (0.38–173.3)
Psychosocial characteristics
Worry about contracting HIV	69.2%	79.5%	1.7 (0.76–3.9)	1.7 (0.72–4.0)
Talks with partner about HIV	91.3%	85.4%	0.55 (0.21–1.4)	0.47 (0.16–1.4)
Mental health characteristics
MDE	10.5%	23.7%	2.7 (1.2–6.1)*	2.9 (1.2–7.3)*
PTSD	18.0%	32.5%	2.2 (1.1–4.6)*	2.5 (1.1–5.7)*
Alcohol abuse	86.3%	95.0%	3.0 (0.7–13.0)	2.4 (0.5–10.8)
Alcohol dependency	62.1%	72.5%	1.6 (0.77–3.3)	1.8 (0.8–3.8)

STI = sexually transmitted infection; OR = odds ratio; CI = confidence interval; CSW = commercial sex worker; MDE = major depressive episode; PTSD = post-traumatic stress disorder

^†Estimates are adjusted for the effects of age, number of years in the military, branch of military service (service and support battalion or battalion 1/2), religious identification (Christian/Catholic versus non-Christian/Catholic) and Ketchi ethnic identification

*P < 0.05; **P < 0.01; ***P < 0.001

Mental health correlates included probable major depressive episode (MDE) and PTSD (Table 3). After controlling for potential confounders, respondents with a probable MDE were nearly three times more likely to have a self-reported STI/HIV (AOR = 2.9, 95% CI 1.2–7.3, P < 0.05), and respondents with probable PTSD were 2.5 times more likely to have a self-reported STI/HIV (aOR = 2.5, 95% CI 1.1–5.7, P < 0.05).

Correlates of consistent condom use

After adjusting for relationship status (single versus not single), the odds of consistent condom use were lower for those using sexual coercion (AOR = 0.37, 95% CI 0.19–0.72, P = 0.004) and decreased with number of days engaged in sex during the past 30 days (AOR = 0.90, 95% CI 0.84–0.97, P = 0.001) (Table 4). The odds of consistent condom use were higher for those worrying about contracting HIV in comparison to those who were not worried (AOR = 2.4, 95% CI 1.1–5.5, P = 0.033), higher for those who always carried condoms in comparison with those who did not always carry condoms (AOR = 2.4, 95% CI 1.2–4.5, P = 0.010), and higher for those who received post-test HIV counselling in comparison with those who had not received post-test counselling (AOR = 1.9, 95% CI 1.01–3.5, P = 0.045). When each of the significant factors described above was entered into a model for consistent condom use, McKelvey and Zavoina's R ² statistic was 0.36.

Table 4

Respondent behaviour by consistent condom use among currently sexually active subjects, n = 274

Characteristics	Inconsistent condom users (n = 223)	Consistent condom users (n = 51)	OR (95% CI)	Adjusted OR^† (95% CI)
Knew sexual partner's HIV status	48.0%	45.1%	0.89 (0.48–1.6)	1.01 (0.54–1.9)
Used sexual coercion	50.9%	27.5%	0.36 (0.19–0.71)**	0.37 (0.19–0.72)**
Worried about contracting HIV	68.5%	84.0%	2.4 (1.1–5.4)*	2.4 (1.1–5.5)*
Talks with partner about HIV	92.3%	92.2%	0.98 (0.32–3.1)	1.02 (0.32–3.2)
Always carry condoms	38.0%	62.8%	2.7 (1.5–5.2)**	2.4 (1.2–4.5)*
Has access to free condoms	72.5%	66.7%	0.76 (0.39–1.5)	0.85 (0.44–1.7)
Ever tested for HIV	76.4%	78.4%	1.1 (0.54–2.4)	1.1 (0.54–2.4)
Received post test HIV counselling	34.1%	49.0%	1.9 (1.01–3.4)*	1.9 (1.01–3.5)*
Number of sexual partners (during last 30 days)	1.4 (2.7)	1.2 (1.6)	0.97 (0.83–1.1)	0.97 (0.82–1.1)
Number of days engaged in sex (during last 30 days)	7.1 (7.5)	3.2 (4.7)	0.89 (0.83–0.96)**	0.90 (0.84–0.97)**
Mental health characteristics:
MDE	11.3%	17.7%	1.7 (0.74–3.9)	1.8 (0.77–4.2)
PTSD	21.0%	23.5%	1.2 (0.56–2.4)	0.96 (0.45–2.0)
Alcohol abuse	88.0%	88.2%	1.0 (0.4–2.6)	1.2 (0.45–3.1)
Alcohol dependency	65.8%	62.8%	0.88 (0.47–1.7)	0.89 (0.47–1.7)

OR = odds ratio; CI = confidence interval; MDE = major depressive episode; PTSD = post-traumatic stress disorder

^†Estimates are adjusted for the effects of relationship status (single versus not single)

*P < 0.05; **P < 0.01; ***P < 0.001

DISCUSSION

This study is the first to document the prevalence of HIV infection in the BDF, and to document key correlates of sexual risk behaviour and condom use in the BDF population. This study has identified several gaps to be addressed in upcoming prevention efforts.

While previous HIV testing was relatively common among personnel, there was still a notable portion that had never reported a test. Our findings indicate that receiving post-test counselling after an HIV test was a correlate of consistent condom use during the past 30 days. As part of this study endeavour, BDF medical personnel were trained in pre- and post-test counselling. By promoting voluntary counselling and testing (VCT) and by ensuring the confidentiality of the test results through professional practice, VCT centres within the BDF should be reinforced as a common and welcoming component of military health. Post-test counselling, a crucial component of regular HIV testing and prevention, is one key area in which to capitalize upon prevention with HIV-negative soldiers.

HIV risk behaviours which were notably problematic included sex with CSWs and inconsistent condom use with CSWs, sex with MCPs, having sex while drunk or high, engaging in sexual coercion and vaginal sex with inconsistent condom use among individuals with MCPs. Further, we found that the number of days engaged in sex during the past 30 days was associated with a decreased likelihood of consistent condom use. While these findings are inconclusive, it is probable that hypersexual behaviour and sexual risk are correlated in this population. While reducing sexual risk behaviour will take sustained prevention efforts, it is immediately feasible to promote consistent condom use in the BDF environment.

Gaps in access to and availability of condoms were identified. Access to free condoms is particularly important for high-risk individuals, as we found (in a post hoc analysis) that access to free condoms was associated with a decreased likelihood of engaging in inconsistent condom use during penetrative anal sex (AOR = 0.31, 95% CI 0.09–1.1, P = 0.06). Further, we found that personnel who reported always carrying condoms were more likely to consistently use condoms (Table 4). In order to ensure maximum accessibility and use, condoms should be prominently displayed, easily available and even promoted as a compulsory accessory at no cost to BDF personnel. The BDF has attempted to make free condoms available to all military service personnel; however, there have been challenges regarding inconsistent supply chains. It is critical that a procurement and supply management plan for condoms be developed for the BDF to ensure the availability of condoms and to avoid stock-outs.

We found that alcohol abuse, depression and PTSD were prevalent and associated with HIV-risk behaviours. These mental health factors have previously been associated with risk behaviours in uniformed personnel,² and may compromise condom use and sexual decision-making. In this sample, 87% of respondents screened positive for probable alcohol abuse on the RAPS4-QF, and 63% screened positive for probable alcohol dependency. Twelve percent of respondents screened positive for experiencing a probable MDE (17.9% in women, 11.5% in men) and 20% of respondents screened positive for probable PTSD (34.5% in women, 18.3% in men). Among respondents screening positive for probable PTSD, 25.4% reported experiencing a natural disaster such as a hurricane, flood or other similar disaster as their worst experience, followed by seeing someone seriously injured or violently killed (14.3%), and a serious accident or fire at home or at your job (12.7%). Among those with probable PTSD, 22% reported that PTSD affected their ability to do their jobs and 35% reported engaging in sex to feel better. The influence of these mental health factors on sexual risk behaviour in this population warrants that they are addressed in the context of HIV prevention efforts.

Focusing on reducing binge drinking during periods where alcohol use is culturally prevalent, such as the time period immediately following field operations, may be a key moment in which to address the role that alcohol abuse plays in HIV risk for BDF personnel. Further analysis revealed that respondents who had sex while drunk or high during the past 30 days exhibited 1.6 more symptoms of alcohol abuse (t = 9.1, df = 307, P < 0.001) than respondents who did not have sex while drunk or high. Reinforcing condom use intention and condom self-efficacy while under the influence of alcohol may be a beneficial component of training and messaging curricula. While treating PTSD necessitates psychiatric services which may not be available, prevention efforts can be tailored to decrease the associated sexual risk behaviours of those with symptoms of PTSD. Teaching basic symptom management, particularly aimed at alternative coping and anxiety reduction strategies (other than drinking or sexual activity) for those with symptoms of PTSD, may be a beneficial component of prevention.

Finally, sexual coercion was prevalent, and personnel who did not use sexual coercion were nearly three times more likely to consistently use condoms. Men who screened positive for using sexual coercion exhibited a greater number of symptoms of alcohol abuse (mean difference = +0.85, t = 4.4, df = 286, P < 0.001), and were more likely to be unaware of their partner's HIV status at last sexual contact (OR = 1.7, 95% CI 1.1–2.7, P = 0.026). Indeed, sexual coercion and sexual negotiation should be a core focus of HIV prevention efforts with this population.

This study has several limitations. First, it is likely that subjects who opted not to participate in the study may have known that they were HIV-positive, which could affect the prevalence estimate. However, the random selection procedure, consent process, and high response rate mitigated this possibility. Second, results can only be generalized to active duty military personnel in the BDF. While the sample was representative of the population of BDF personnel, conclusions about women in the armed forces are limited since the sample was predominantly male. Third, social desirability bias is always a concern in measuring HIV risk behaviours. However, the use of ACASI may reduce the likelihood of such biases. Finally, HIV results were derived from a test and the STI outcomes were derived from self-reported recall of what a doctor said. This variable does not capture serological prevalence of STIs, but it does represent an outcome of a previous and/or ongoing sexual risk behavioural process. Separate questions were asked for each specific STI, thus reducing the probability of recall biases.

In this study, we found that BDF personnel take regular sexual risks, which can be mitigated through prevention efforts. These data will be used to inform efforts to reduce the risk of HIV infection and to improve the health of military personnel stationed throughout the country, as well as their sexual partners.

References

Brodine

, Starkey

, Shaffer

, Diverse HIV-1 subtypes and clinical, laboratory and behavioral factors in a recently infected US military cohort. AIDS 2003;17:2521–7

Tavarez

, Chun

, Anastario

. Correlates of sexual risk behavior in sexually active male military personnel stationed along border-crossing zones in the Dominican Republic. Am J Mens Health 2011;5:65–77

UNAIDS. Fact sheet: Key facts by region – 2007 AIDS Epidemic Update. 19 November 2007. See http://www.unaids.org/en/HIV_data/2007EpiUpdate/default.asp (last checked 12 December 2007)

Organization

TWH

. Summary Country Profile for HIV/AIDS Treatment Scale-up. 2005. See http://www.who.int/hiv/HIVCP_BLZ.pdf (last checked May 2010)

Ghanem

, Hutton

, Zenilman

, Zimba

, Erbelding

. Audio computer assisted self interview and face to face interview modes in assessing response bias among STD clinic patients. Sex Transm Infect 2005;81:421–5

Risk Behavior Assessment Questionnaire. National Institute on Drug Abuse: Community Research Branch, Rockville, MD. 1993:16–44

Aidala

, Havens

, Mellins

, Development and validation of the Client Diagnostic Questionnaire (CDQ): a mental health screening tool for use in HIV/AIDS service settings. Psychol Health Med. 2004;9:362–79

Jakopanec

, Schimmer

, Grjibovski

, Klouman

, Aavitsland

. Self-reported sexually transmitted infections and their correlates among men who have sex with men in Norway: an Internet-based cross-sectional survey. BMC Infect Dis 2010;10:261

Kissinger

, Liddon

, Schmidt

, Curtin

, Salinas

, Narvaez

. HIV/STI Risk behaviors among Latino migrant workers in New Orleans post-Hurricane Katrina disaster. Sex Transm Dis 2008;35:924–9

10.

Straus

, Hamby

, Boney-McCoy

, Sugarman

. The Revised Conflict Tactics Scales (CTS2): Development and Preliminary Psychometric Data. J Fam Issues 1996;17:283–316

11.

Diagnostic and Statistical Manual of Mental Disorders: Fourth Edition. Text Revision. Arlington, VA: American Psychiatric Association, 2000

12.

Breslau

, Peterson

, Kessler

, Schultz

. Short screening scale for DSM-IV posttraumatic stress disorder. Am J Psychiatry 1999;156:908–11

13.

Lowe

, Kroenke

, Grafe

. Detecting and monitoring depression with a two-item questionnaire (PHQ-2). J Psychosom Res 2005;58:163–71

14.

Cherpitel

. Screening for alcohol problems in the US general population: comparison of the CAGE, RAPS4, and RAPS4-QF by gender, ethnicity, and service utilization. Rapid Alcohol Problems Screen. Alcohol Clin Exp Res 2002;26:1686–91

15.

Stata 10 [computer program]. College Station, Texas: Stata Corporation, 2007