Correlates of HIV knowledge and Sexual risk behaviors among Female Military Personnel

Abstract

Objective

Uniformed services personnel are at an increased risk of HIV infection. We examined the HIV/AIDS knowledge and sexual risk behaviors among female military personnel to determine the correlates of HIV risk behaviors in this population.

Method

The study used a cross-sectional design to examine HIV/AIDS knowledge and sexual risk behaviors in a sample of 346 females drawn from two military cantonments in Southwestern Nigeria. Data was collected between 2006 and 2008. Using bivariate analysis and multivariate logistic regression, HIV/AIDS knowledge and sexual behaviors were described in relation to socio-demographic characteristics of the participants.

Results

Multivariate logistic regression analysis revealed that level of education and knowing someone with HIV/AIDS were significant (p<0.05) predictors of HIV knowledge in this sample. HIV prevention self-efficacy was significantly (P<0.05) predicted by annual income and race/ethnicity. Condom use attitudes were also significantly (P<0.05) associated with number of children, annual income, and number of sexual partners.

Conclusion

Data indicates the importance of incorporating these predictor variables into intervention designs.

Introduction

Human Immunodeficiency Virus infection is a major public health challenge in the Federal Republic of Nigeria. Nigeria is the second most affected country in sub-Saharan Africa (SSA) with HIV disease, representing 14% of HIV/AIDS cases in the region¹. While official reports have suggested that the HIV epidemic has been slower to impact Nigeria than other countries in Africa, research evidence in Nigeria suggests that HIV prevalence is high, as well as geographically and socially widely distributed. For example, Esu-Williams et al.² report that in a sample of 2,300 persons from five states in Nigeria, HIV-1 appears in over 60% of commercial sex workers (CSW), 8% of blood donors in some states, with male clients of CSWs, truck drivers, and STD patients having respectively 8%, 9%, and 21%. More recently, a United States Naval Health Research Center funded study found a 15% Seroprevalence rate among Nigerian military personnel.³ As the most populous country in Africa (Population > 130 million) and one of the most populous countries in the world, even a small increase in the HIV/AIDS prevalence rate in Nigeria would represent a significant share of the global HIV/AIDS burden⁴.

In 2007 women accounted for 58 percent of all adults aged 15 and above living with HIV. ⁵ Some of the factors responsible for the growing epidemic among women in Nigeria include various cultural practices like polygamy and a culture of silence over sexuality, the rising popularity of multi-partner mating among urban-based women in the 15-29 years' age group, women's inability to negotiate safe sex with their partners, lack of sexual education both at home and in schools, religious leaders' teaching which has led to stigma and which has not been empowering to women and an unwillingness of most men to use condoms. Research has also shown that because of cultural and economic reasons, many women feel unable to refuse the sexual advances of partners even when they know they risk infection. ^{6, 7} Poverty has pushed some young women between the ages of 15 and 25 into sex work ⁶ or to be involved in transactional sex with older men (sugar daddies) who give them monies, school fees or gifts in exchange for sex. In exploring the reasons for extramarital sexual relationships among married women in South-western Nigeria, Oruboloye and others ⁷ noted that 60% of married urban women and 33% of rural married women have extramarital sex for enjoyment, while 34% of married rural women and 14% of married urban women had sex as a means of securing economic benefits. This is particularly important since unprotected sex among married couples is the norm, yet many married men and women engage in unprotected extramarital sexual relations, and thus risk infecting their spouses. Thus, the level of high-risk sexual networking within or outside marriage in Nigeria tends to expose large sections of the population to the risk of HIV and other sexually transmissible diseases.

Epidemiologic evidence has consistently shown that military personnel are a high risk subpopulation with social norms that place them at an elevated risk of HIV infection.^{8, 9} Nwokoji and Ajuwon ¹⁰ explored the HIV related risk behaviors among military personnel in Nigeria by asking 480 enlisted men to complete a 70-item questionnaire that assessed HIV/AIDS knowledge, sexual behavior and risk-perception. The study revealed that 41% of the respondents did not use a condom during their last sexual encounter with a commercial sex worker and posting on international assignments was a positive predictor of lack of condom use. Similarly, Essien et al ^{11, 12} examined the determinants of HIV risk behaviors among Nigerian military personnel and found a direct correlation between alcohol and marijuana use and HIV risk perception. Their study also showed that knowledge of how to correctly wear a condom and male gender were positive predictors of intent to wear a condom. From a broader perspective, other investigators ^{10, 13} have shown a positive relationship between alcohol and marijuana use and inconsistent condom use among Nigerian military personnel.

A few reports in the literature have addressed the issue of HIV transmission among Nigerian military personnel. Most soldiers are young and sexually active with a sense of invulnerability that may lead to risky sexual behaviors and reduced condom use.¹³ Soldiers are often deployed from home for extended periods of time, have a regular income and the opportunity for casual sex.¹⁴ For instance, it has been reported that almost half of the military personnel that participated in the various peacekeeping operations admitted having sexual partners during their time away from home and with these sexual partners, only half of the respondents used condoms.⁸ In addition, societal norms that do not support condom use have been known to also contribute to the efficiency of HIV transmission among Nigerian military personnel.¹⁵

Like their male counterparts, the high mobility of women in the armed forces also places them at risk of HIV infection for the same reasons they share. Also, female military personnel sexual interactions with local partners while on peacekeeping missions, and with officers returning from peacekeeping missions to the barracks may also put them at risk of HIV infection since the lifestyles of militaries on such missions are often characterized by high levels of multiple sexual partners, including sex with commercial sex workers; low condom use, and exposure to blood transfusions in the line of duty. In addition, they are subject to sex under duress, transactional sex for favors from superior officers; and sometimes are at risk of outright rape. Therefore, the complexity of sexual networking within or outside the militaries suggest that some female military personnel may serve as a significant vector in a concentrated HIV epidemic in the barracks, as well as being a potential bridge to the general population through sexual relationship with civilians. While there is a growing body of literature examining HIV risk behaviors among Nigerian military personnel, these studies have been based predominantly on male samples and provide only limited information about risk behaviors among female military personnel. The current research sought to address this void in the literature by examining the correlates of HIV knowledge and risk behaviors among Nigerian female military personnel.

Methods

Participants

Study participants were female military personnel drawn from two cantonments in Southwestern Nigeria. The study population was comprised of 346 females, ages 18 and above, who were recruited to participate in an HIV intervention study. Female military personnel were eligible to participate in the study if they were: (1) sexually active and had a history of unprotected vaginal intercourse in the past twelve months; (2) had a history of multiple sexual partners; (3) did not plan to retire from the military in the next twelve months; (4) had a history of alcohol and drug use; (5) and had the ability to communicate in English. Participants were excluded from the study if they were unable to sign an informed consent form or had an emotional disorder that could interfere with the study.

Study Design

The study used the baseline data that was collected for a videotape-based HIV prevention intervention to examine the correlates of HIV knowledge and risk behaviors in the study population. Baseline data were collected in 2006 and 2008 using a cross-sectional design. The study used an adapted version of a previously validated instrument, ¹⁶ which was designed to assess HIV/AIDS knowledge, HIV risk behaviors, alcohol and drug use, condom use practices, HIV prevention self-efficacy and peer norms. Also, captured were the socio-demographic characteristics of the participants. The instrument was cross-validated with the present study population.

Measures

The investigative team administered the assessments using a group format. The facilitator used overhead projection transparencies of the instrument to walk the participants through the measures. This procedure has been found to be particularly effective in eliciting accurate responses to HIV risk assessments among populations with a low level of literacy.¹⁷ Measures included demographic characteristics, HIV/AIDS knowledge, HIV prevention self-efficacy, condom use attitudes and barriers, substance use and sexual behaviors and peer norms.

Socio-demographic characteristics

The participants reported their age, marital status, race/ethnicity, religion, level of education, and employment status. Information was also obtained on socio-economic characteristics such as annual income, sexual relationships, and personally knowing someone with HIV/AIDS.

AIDS-related knowledge

We used a 10-item test to assess HIV/AIDS risk prevention knowledge. The items elicited information on HIV transmission knowledge, condom use knowledge, and AIDS-related knowledge. A categorical scale with three levels of responses (Yes, No, Don't know) was used. Example items included “Can a woman give the AIDS virus to a man?” (yes), “Can you get AIDS by touching a person with AIDS?” (no). Each correct answer was scored 1 point, and a total score of 10 points was attainable based on the number of questions. Participants that scored 5 points or less were classified as having ‘poor’ knowledge and those that scored above 5 points were classified as having ‘good’ knowledge. The internal consistency of the HIV/AIDS knowledge scale in our sample was, alpha= 0.74.

HIV Prevention Self-efficacy

We assessed HIV prevention self-efficacy using a 6-item scale that examined the participants' self-efficacy for condom use, HIV testing and substance use prior to sexual intercourse. Examples of items are: “Talked with sex partner about using male condoms or safer sex in the past three months”, “Did not have sex because you did not have a condom”, “drank less or used drugs less before having sex”. The questions were anchored on a three-month timeframe. The responses were categorized, with the highest score indicating a more favorable HIV prevention self-efficacy. Participant taking a positive action on any of the items for 5 times or less in the past 3 months was classified as having ‘low’ HIV prevention self-efficacy and those taking positive action for more than 5 times as having ‘High’ HIV prevention self-efficacy. The internal consistency of the HIV prevention self-efficacy scale in our sample was Cronbach alpha = 0.78

Condom use attitudes and barriers

Attitudes toward male and female condoms were assessed using items that measured intent and utilization of condoms. A binary scale (yes or no) was used to score the item for actual condom use: I have used latex condoms. Condom attitude and barriers were assessed using a four-item test. The participants responded on a four-point scale ranging from 1 (strongly disagree) to 4 (strongly agree). The items included: “Female condoms take away pleasure” (reverse score), “male condoms reduce the fun of sex (reverse score)”. “I would be embarrassed to buy condoms” (reverse score), and male “condoms are a hassle to use”. Overall, a positive attitude towards condom use was determined by the lowest score. Participants with a total score less than or equal to 55 were regarded as having positive attitudes towards condom use, and those scoring above 55, as having negative attitudes. The internal consistency of the condom use scale in our sample was Cronbach alpha = 0.66.

Substance use and sexual behaviors

We classified drug use into major substances with which ingestion could result in behavioral impairment and altered mentation. These agents included alcohol, marijuana, cocaine, amphetamines, and ecstasy. The agents were assessed in relation to sexual encounters. First, we asked the participants to provide a Yes or No response to questions that elicited information on the use of substances prior to sexual encounters during the past three months. Secondly, the participants were also asked to indicate the frequency of substance use during the past three months. These responses were later categorized into four groups and used in the scale development. The internal consistency of the substance use and sexual risk behavior scale in our sample was Cronbach alpha = 0.62.

Human Subjects and Ethical Considerations

The research protocol was reviewed and approved by the relevant Institutional Review Boards at University of Houston in the United States and Institute for Health Research and Development in Nigeria.

Statistical analysis

First, descriptive analysis using frequency runs was carried out to determine the distribution of socio-demographic characteristics of the study participants. Chi-square statistic, (with Fisher's exact test applied where applicable to correct for small cells count), was used to assess the association of the Socio-demographic characteristics of study participants with HIV/AIDS knowledge, condom use attitudes, HIV prevention self-efficacy, and substance use and sexual behaviors, respectively. Based on the outcome of this analysis, the predictor variables for use in multivariable logistic regressions were selected a priori, if they were epidemiologically important or if they were significant at the 0.10 level in a bivariate analysis. Multivariable logistic regression was used to model HIV/AIDS knowledge, condom use attitudes and HIV prevention self-efficacy, and substance use. Adjusted Prevalence Odds Ratio (APOR) with 95% confidence interval was computed for each association. All tests were two-tailed, with probability value of 0.05 used as the statistical significance level. Data management and statistical analyses were performed using SPSS software version 14.0 (SPSS Inc, Chicago, IL).

Results

Sociodemographic Characteristics

The Sociodemographic characteristics of the study population are presented in Table 1. There were a total of 346 participants from two military cantonments with majority of them of age range 30-39 years (56.4%). Most of these participants had high school diploma and above (303, 87.8%), with 42 (12.2%) having less than high school education. The marital status of the study population indicates that 64.4% of participants were singles, with only 4.1% living as partners. Only 9.9% of them were married, with 40.8% of the participants claiming to have one or more children. Among the female military personnel, there were more Christians (77.7%) than Muslims (21.1%). Also, there were more participants of Yoruba tribe (33.8%), compared to Hausa and Ibo tribes. However, other minor tribes together represented about 33.2% of the study population (Table1). Approximately 30% of the respondents reported knowing one or more people that have been infected with HIV/AIDS. Ninety six percent of the participants were still in active service, with majority of them (51%) in the lower income brackets of ₦241, 000- ₦360,000 per annum (Table 1).

Table 1.

Socio-demographic characteristics of participants

Variable	N	%a
Total	346	100
Age (yrs)
18-29	82	23.7
30-39	195	56.4
40+	69	19.9
Education Level
Less than High School	42	12.2
High School and above	303	87.8
Marital Status
Single	221	64.4
Married	34	9.9
Separated/divorced/widowed	74	21.6
Living as partner	14	4.1
Number of Children
No child	205	59.2
One or more child(ren)	141	40.8
Race/Ethnicity
Hausa	64	18.5
Ibo	50	14.5
Yoruba	117	33.8
Other	115	33.2
Religion
Christian	269	77.7
Muslim	73	21.1
Other	4	1.2
Employment Status
Active Service	331	96.2
Trainee b	13	3.8
Annual Income (× 1000)
₦120-240	63	18.3
₦241-360	176	51.0
₦361-500	93	27.0
₦500+	13	3.8
Sexual Relationship
Sex with one partner	218	63.0
Sex with multiple partners	128	37.0
Personally known someone with HIV
Yes	105	30.3
No	241	69.7
Number of people known with HIV
None	241	69.7
1-2	85	24.6
> 2	20	5.8

^aWeighted % of those who responded; Some percentages may not add up exactly to 100% due to rounding.

^bRefers to military personnel undergoing some form of training.

Knowledge of HIV/AIDS

Table 2 shows the bivariate and multivariate correlates of HIV/AIDS knowledge among female military personnel in Nigeria. The knowledge scores ranged from 2 to 10 with a mean of 7.6 (SD=2.18). The results indicate that overall, 76.1% of the participants have good knowledge of HIV/AIDS compared to 23.9% with poor knowledge of the disease. Knowledge of HIV/AIDS among the group was independently associated with educational attainment (P=0.005), religion (P=0.021), number of sexual partners (P=0.050) and knowing someone with the HIV/AIDS disease (P=0.001). Of the participants with good knowledge of HIV/AIDS, 34.9% of them compared to 13.8% with poor knowledge of HIV/AIDS agreed that they knew people infected with the disease (table 2). However, multivariate logistic regression analysis showed that HIV/AIDS knowledge among the female military personnel was significantly predicted (R²=0.06; P<0.01) by level of education (P=0.050) and knowing someone with the HIV/AIDS disease (P=0.003). Those with high school diploma and above were twice more likely than those below high school, to have good knowledge of HIV/AIDS (APOR=2.08; 95%CI: 0.97-4.46).

Table 2.

Bivariate and multivariate association between socio-demographic characteristics and HIV/AIDS knowledge among female military personnel

	HIV/AIDS Knowledge
Correlates	Good		Poor				Multivariate d
	n	% e	n	% e	χ2C	P-value	Wald	APOR (95%CI)a	P-value
Age (yrs)
< 30 yrs	61	23.9	17	21.3
>= 30 yrs	194	76.1	63	78.7	0.24	0.622	N/A	--
Education Level
Less than High School (Ref)b	24	9.4	17	21.3			3.53	2.08 (0.969-4.461)	0.050*
High School and above	230	90.6	63	78.7	7.87	0.005**
Marital Status
Single	198	78.3	60	75.9
Married	55	21.7	19	24.1	0.19	0.667	N/A	--	--
Number of Children
No child	155	60.8	49	61.3
One or more child(ren)	100	39.2	31	38.7	0.01	0.941	N/A	--
Race/Ethnicity
Hausa	41	16.1	21	26.3
Ibo	38	14.9	10	12.5
Yoruba	86	33.7	28	35.0	5.18	0.159	N/A	--	--
Other	90	35.3	21	26.3
Religion
Christian (Ref)b	206	82.1	56	70.0			1.15	0.698 (0.361-1.348)	0.284
Muslim	45	17.9	24	30.0	5.36	0.021*
Employment Status
Trainee (Ref)b	4	1.6	4	5.1		0.094 (f)	2.54	3.34 (0.757-14.732)	0.111
Active Service	250	98.4	75	94.9
Annual Income
Low	175	68.6	60	75.9
High	80	31.4	19	24.1	1.55	0.213	N/A	--
Sexual Relationship
Sex with one partner (Ref)b	165	64.7	42	52.5			0.64	1.25 (0.725-2.147)	0.425
Sex with multiple partners	90	35.3	38	47.5	3.84	0.050*
Personally known someone with HIV
No (Ref)b	166	65.1	69	86.3	13.01	0.001***	8.84	0.34 (0.166-.691)	0.003**
Yes	89	34.9	11	13.8

^aAPOR (95%CI): Adjusted Prevalence Odds Ratio; 95% Confidence Interval.

^bRef: Reference category.

^cExcept for race/ethnicity with df=3, all other correlates have df =1.

^dOnly correlates that met the entry criteria of P ≤ 0.10 in the bivariate analysis were included in the multivariate logistic regression model.

^eSome percentages may not add up exactly to 100% due to rounding.

f: Fisher's Exact Test.

N/A: Not Applicable

^*Significance level: P<0.05;

^**Significance level: P<0.01;

^***Significance level: P<0.001.

HIV Prevention Self-Efficacy

The confidence among the participants in performing a specific behavior was determined in relation to their socio-demographic characteristics (Table 3). On the average, more than 70% of the participants reported low HIV prevention self-efficacy, having taken a positive action for only 2.64 times (SD: 1.73) on the average in the last 3 months (result not presented). Univariate analysis indicated that significant independent associations were only noted between HIV prevention self-efficacy and the following variables: race/ethnicity (χ² = 7.56; P=0.053), religion (χ² = 5.71; P=0.017), and annual income (χ² = 3.99; P=0.046), respectively. However, multivariate analysis indicated that HIV prevention self-efficacy among the female military personnel was significantly predicted (P≤0.05) by their annual income and race/ethnicity (Table 3). The Yoruba race recorded a significant average protective APOR of 0.44, 95% CI 0.22–0.92. Female military personnel with high income earnings were 3.48 times (APOR=3.48; 95%CI: 1.09-11.08), more likely to have high HIV prevention self-efficacy and to engage in safe sex behaviors than those with low income earnings.

Table 3.

Bivariate and Multivariate association between socio-demographic characteristics and HIV Prevention Self-Efficacy among female military personnel

Correlates	HIV Prevention Self-Efficacy
	High		Low					Multivariate d
	n	% e	n	% e	χ2c	P-value	Wald	APOR (95%CI)a	P-value
Age (yrs)
< 30 yrs	26	27.1	52	21.8
>= 30 yrs	70	72.9	187	78.2	1.09	0.297	N/A	--	--
Education Level
Less than High School (Ref)b	7	7.3	34	14.3				0.61 (0.246-1.510)	0.285
High School and above	89	92.7	204	85.7	3.11	0.078	1.14
Marital Status
Single	78	82.1	180	75.9
Married	17	17.9	57	24.1	1.48	0.223	N/A	--	--
Number of Children
No child	63	65.6	141	59.0
One or more child(ren)	33	24.4	98	41.0	1.26	0.261	N/A	--	--
Race/Ethnicity
Hausa (Ref)b	14	14.6	48	20.1
Ibo	21	21.9	27	11.3	7.56	0.053*	3.26	0.33 (0.099-1.099)	0.071
Yoruba	34	35.4	80	33.5			4.82	0.44 (0.215-0.917)	0.028*
Other	27	28.1	84	35.1			1.36	0.70 (0.381-1.278)	0.244
Religion
Christian (Ref)b	84	87.5	178	75.7			1.00	4.85 (0.220-10.654)	0.317
Muslim	12	12.5	57	24.3	5.71	0.017*
Employment Status
Active Service	96	100.0	229	96.6
Trainee	0	0.0	8	3.4		0.111 (f)	N/A	--	--
Annual Income
Low (Ref)b	60	62.5	175	73.5			4.44	3.48 (1.091-11.075)	0.035*
High	36	37.5	63	26.5	3.99	0.046*
Sexual Relationship
Sex with one partner	62	64.6	145	60.7
Sex with multiple partners	34	35.4	94	39.3	0.44	0.505	N/A	--	--
Personally known someone with HIV
No	70	72.9	165	69.0	0.49	0.483	N/A	--	--
Yes	26	27.1	74	31.0

^aAPOR (95%CI): Adjusted Prevalence Odds Ratio; 95% Confidence Interval.

^bRef: Reference category

^cExcept for race/ethnicity with df=3, all other correlates have df =1

^dOnly correlates that met the entry criteria of P ≤ 0.10 in the bivariate analysis were included in the multivariate logistic regression model.

^ePercentages may not add up exactly to 100% due to rounding.

f: Fisher's Exact Test.

N/A: Not Applicable

^*Significance level: P<0.05.

Condom use attitudes and barriers

Table 4 presents the bivariate and multivariate association of condom use attitude and barrier scale in the study population. An overall ratio of 1:1 was noted for participants with positive and negative condom use attitude and barrier with a mean scores of 59.2 (SD: 2.99) and 51.2 (SD: 3.98), respectively. Among the covariates considered, only participants' annual income (P= 0.054) and number of sexual partners (P=0.025) were significantly associated with condom use attitude and behavior. Majority of the participants (61.8%) indicated that they had sex with multiple partners in the past three months, with 34.3% and 27.5% of them having positive and negative condom use attitudes. Multivariate analysis of the association, however, indicates that positive condom use attitudes and behavior among female military personnel was significantly predicted by the number of children (P=0.003), annual income (P=0.036) and the number of sexual partners (P=0.001). Participants with high annual income were 71% more likely (APOR=1.71; 95%CI: 1.04-2.83) to have positive attitudes towards condom use and to engage in safe sexual behaviors than the low annual income earners. With respect to sexual activity, during the previous three months period, study participants on the average had 3 sexual encounters without condoms and 4 sexual encounters with condoms (Table 5). This implies that condoms were used about 57% of the time. There was a significant difference between the number of times they had ‘sex without a condom’ with a single partner (Mean = 4.0; 95%CI: 3.62-4.39) compared to ‘sex with multiple partners’ (Mean = 2.79; 95%CI: 2.27-3.32). However, no significant differences (P>0.05) were observed by type of partners (casual vs. non-casual).

Table 4.

Bivariate and Multivariate association between socio-demographic characteristics and Condom use Attitude and Behavior among female military personnel

Correlates	Condom use Attitude and Behavior
	Positive		Negative				Multivariate d
	n	% e	n	% e	χ2c	P-value	Wald	APOR (95%CI)a	P-value
Age (yrs)
< 30 yrs	46	27.1	32	19.4			2.61	0.63 (0.37-1.10)	0.106
>= 30 yrs	124	72.9	133	80.6	2.75	0.097
Education Level
Less than High School (Ref)b	21	12.4	20	12.1
High School and above	148	87.6	145	87.9	0.01	0.932	N/A	--	--
Marital Status
Single	133	79.2	125	76.2
Married	35	20.8	39	23.8	0.42	0.519	N/A	--	--
Number of Children
No child (Ref)b	111	65.3	93	56.4			9.06	0.44 (0.25-0.75)	0.003**
One or more child(ren)	59	34.7	72	43.6	2.80	0.094
Race/Ethnicity
Hausa (Ref)b	33	19.4	29	17.6
Ibo	23	13.5	25	15.2	1.92	0.589	N/A	--	--
Yoruba	53	31.2	61	37.0
Other	61	35.9	50	30.3
Religion
Christian (Ref)b	135	80.4	127	77.9
Muslim	33	19.6	36	22.1	0.30	0.584	N/A	--	--
Employment Status
Active Service	163	97.0	162	98.2
Trainee	5	3.0	3	1.8		0.723 (f)	N/A	--	--
Annual Income
Low (Ref)b	111	65.7	124	75.2			4.39	1.71 (1.04-2.83)	0.036*
High	58	34.3	41	24.8	3.59	0.054*
Sexual Relationship
Sex with one partner (Ref)b	55	32.4	73	44.2			11.64	0.40 (0.23-0.67)	0.001***
Sex with multiple partners	115	67.7	92	55.8	5.01	0.025*
Personally known someone with HIV
No	118	69.4	117	70.9	0.90	0.765	N/A	--	--
Yes	52	30.6	48	29.1

^aAPOR (95%CI): Adjusted Prevalence Odds Ratio; 95% Confidence Interval.

^bRef: Reference category

^cExcept for race/ethnicity with df=3, all other correlates have df =1

^dOnly correlates that met the entry criteria of P ≤ 0.10 in the bivariate analysis were included in the multivariate logistic regression model.

^ePercentages may not add up exactly to 100% due to rounding.

f: Fisher's Exact Test.

N/A: Not Applicable

^*Significance level: P<0.05;

^**Significance level: P<0.01;

^***Significance level: P<0.001.

Table 5. Condom Use for Vaginal Sex in the past three months among Sexually-Active Female Military Personnel in Nigeria⁺.

Variable	n (%)	Without Condom Mean (95% CI) +	P-value	n (%)	with condom Mean (95% CI) +	P-value
Overall	278 (100)	3.40 (3.07-3.72)	---	318 (100)	4.18 (3.91-4.47)	---
Sexual Relationship
Single partner	178 (64)	4.00 (3.62-4.39)		196 (62)	4.15 (3.81-4.50)
Multiple partners	100 (36)	2.79 (2.27-3.32)	0.000***	122 (38)	4.22 (3.78-4.66)	0.825ns
Type of partner
Casual	128 (46)	3.23 (2.80-3.66)		148 (47)	4.21 (3.84-4.59)
Non-casual	150 (54)	3.56 (3.07-4.05)	0.317ns	170 (53)	4.16 (3.75-4.57)	0.853ns

⁺Mean (95% CI) number of times participants had vaginal sex with or without condom in the past three months

^***Significance level: P<0.001;

^nsnot significant (P>0.05).

Substance use and sexual behaviors

The overall proportional distribution indicates that about 31% of the study population reported being involved in sexual behaviors associated with substance use. However, of the correlates considered, the following were significantly associated with substance use and sexual behaviors of participants: educational attainment (P=0.012), race/ethnicity (P=0.039), employment status (P=0.05), number of sexual partners (P=0.000) and knowing someone with HIV/AIDS disease (P=0.000) (Table 6). Multivariable logistic regression analysis showed that substance use and risky sexual behavior among the female military personnel was significantly predicted by their educational level, number of sexual partners and knowing someone with HIV/AIDS disease. Participants with high school certificate and above, having sex with multiple partners, and knowing someone impacted by the HIV/AIDS disease had a significantly higher likelihood of substance use with adjusted prevalence odds ratios of 3.92, 3.08 and 2.58, respectively (Table 6).

Table 6.

Bivariate and Multivariate association between socio-demographic characteristics and Substance use and sexual behaviors among female military personnel

Correlates	Substance use and sexual behaviors
	Yes		No				Multivariate d
	n	% e	n	%	χ2c	P-value	Wald	APOR (95%CI)a	P-value
Age (yrs)
< 30 yrs	25	23.6	53	23.1
>= 30 yrs	81	76.4	176	76.9	0.08	0.929	N/A	--	--
Education Level
Less than High School (Ref)b	6	5.7	35	15.3				3.92 (1.47-10.48)	0.006**
High School and above	99	94.3	194	84.7		0.012**(f)	7.41
Marital Status
Single	83	78.3	175	77.4
Married	23	21.7	51	22.6	0.03	0.859	N/A	--	--
Number of Children
No child	69	63.9	137	59.8
One or more child(ren)	39	36.1	92	40.2	0.35	0.555	N/A	--	--
Race/Ethnicity
Hausa (Ref)b	19	17.9	43	18.8			2.15	0.51 (0.21-1.25)	0.143
Ibo	21	19.8	27	11.8	8.39	0.039*	2.18	1.79 (0.83-3.85)	0.140
Yoruba	26	24.5	88	38.4			0.05	1.09 (0.52-2.30)	0.824
Other	40	37.7	71	31.0
Religion
Christian	88	83.8	174	77.0
Muslim	17	16.2	52	23.0	2.02	0.155	N/A	--	--
Employment Status
Trainee (Ref)b	0	0.0	8	3.5
Active Service	106	100	219	96.5		0.05* (f)	0.00	0.00	0.99
Annual Income
Low	78	73.6	157	68.9
High	28	26.4	71	31.1	0.78	0.379	N/A	--	--
Sexual Relationship
Sex with one partner (Ref)b	46	43.4	161	70.3
Sex with multiple partners	60	56.6	68	29.7	22.22	0.000***	17.50	3.08 (1.82-5.22)	0.000***
Personally known someone with HIV
No (Ref)b	88	83.0	147	64.2	12.27	0.000***	8.39	2.58 (1.36-4.89)	0.004**
Yes	18	17.0	82	35.8

^aAPOR (95%CI): Adjusted Prevalence Odds Ratio; 95% Confidence Interval.

^bRef: Reference category

^cExcept for race/ethnicity with df=3, all other correlates have df =1

^dOnly correlates that met the entry criteria of P ≤ 0.10 in the bivariate analysis were included in the multivariate logistic regression model.

^ePercentages may not add up exactly to 100% due to rounding.

f: Fisher's Exact Test.

N/A: Not Applicable.

^*Significance level: P<0.05;

^**Significance level: P<0.01;

^***Significance level: P<0.001.

Discussion

Military populations, in general, have increased vulnerability to HIV infection, compared to the general civilian populations. Risk factors among the military include high rates of sexual partner change, elevated rates of STIs, low rates of condom use with commercial sex workers and other casual partners, and significant mixing between groups having high and low risk behavior patterns.²⁵ The study attempted to ascertain participants' level of knowledge with respect to HIV/AIDS through a number of questions. Although more than three quarters of the participants had good knowledge of HIV/AIDS including mode of transmission, few of them lacked adequate knowledge to make responsible decisions about HIV/AIDS risk behavior and possessed beliefs about HIV exposure that may increase their risk. Similar findings of armed forces personnel having a high degree of knowledge and engaging in risk-taking behaviors have been reported in other studies.^{19, 20, 21} However, an appreciable level of knowledge of the modes of transmission of HIV and how to prevent it are important prerequisites for behavior change.²¹ The significant association noted in the bivariate analysis between religious affiliation and HIV/AIDS knowledge suggests that religious organizations may constitute a strong cultural force for preventive education in the military. More than three quarter of the personnel in our study were Christians, even though Muslims represent nearly 45 percent of the general population in Nigeria.¹⁸

In our study population we found that knowledge of HIV/AIDS was significantly associated with level of education and the personnel's knowing of someone infected with HIV/AIDS disease. Majority of the respondents were literate, with only 12.2 percent having less than high school education. Unlike in the past, new entrants into the Nigerian armed forces are now required to have some level of education either from within the military system through the Nigerian Defense Academy or from outside the system. Thus, the Nigerian military population is quite well educated compared with the general population.¹⁸ Educational level might protect against HIV infection through information and knowledge that may affect long-term behavioral change, particularly for women by “reducing the social and economic vulnerability that exposes [them] to a higher risk of HIV/AIDS than men”, including prostitution and other forms of economic dependence on men.²³ We noted that knowing someone impacted by the HIV disease tends to encourage the females' quest for HIV/AIDS knowledge. Proximity to the disease (e.g., knowing someone with HIV/AIDS) has similarly been shown to be useful in educating others about the disease.²⁴ This finding confirms the wisdom of many community programs that utilize persons with HIV to reach and motivate the community. However, the low level of variance explained by the correlates in our study is an indication that several other factors not captured here may be responsible for determining the female military personnel's knowledge of HIV/AIDS.

In our study, we found that more than 70% of the respondents had lower HIV prevention self-efficacy; with race, religious affiliations and annual income being significant correlates. Although this finding differs from previous reports where women were said to have greater condom use self-efficacy, ^{26, 27} it does support the gender difference in perceived control during a sexual encounter.^{28, 29} Specifically, the lack of perceived control over a sexual encounter by women may explain the lower HIV prevention self-efficacy and the higher perceived condom use barriers reported by the females in our sample. Also, the series of misconceptions about the nonexistence of AIDS and/or the myths about the availability of cure, ¹⁰ may encourage some females to practice risky behaviors. Culturally, women carrying condoms in Nigeria are stigmatized as being sexually permissive. Therefore, as reported by many researchers, condom use self-efficacy in women may generally reflect their ability to apply condoms, negotiate condom use, exert self-control during sexually arousing encounters, and develop acceptance of sexuality. ^{28, 30} In our study, the low HIV prevention self-efficacy may have contributed to the less than optimal use of condoms. The variable odds of lower HIV prevention self-efficacy by ethnicity in the study population may help account for the role of ethnicity in predicting sexual risk behaviors. It is also possible that the ability to tease apart cultural and social bias from personal choice may dramatically impact HIV prevention self-efficacy. In Nigeria there are three major ethnic groups namely Yoruba, Ibo and Hausa. The Hausas who occupy the northern part of the country are mainly Muslims, while the Yorubas found in the southwestern region are a blend of Christians and Muslims. The Ibos inhabit the southeastern part of the country and are mainly Christians of Catholic faith. It is a generally held view among these major ethnic groups that decisions on safe sex are left with men. Women are rarely in a position to insist on the use of a condom if their partners do not want it. Nor can they protect themselves by using a female condom without their partners' permission or they may be accused of infidelity. With females constituting approximately 6-10% of the military in Nigeria, ¹³ and being exposed to the same - and sometimes even greater - pressure as men to enter into casual sexual relationships, there is need to promote ingenious ways of female self-protection from the highly dominant males. Therefore being safe in the circumstance would require the female military personnel to change their individual behaviors and to develop strategies to change the social context of their lives in the barracks.

On a fundamental level, Christians and Muslims in Nigeria have similar views on why HIV continues to spread: both groups see promiscuous behavior as the root cause of the HIV crisis and promiscuity is frowned upon heavily because of religious teachings and because of underlying cultural traditions within the Nigerian society. Although Christians do not believe people should engage in sexual behavior before marriage, a social stigma is the harshest punishment a person would receive from society if their extra-marital sex is discovered. Muslims, on the other hand, could be punished for their decisions about extra marital sex through the system of Sharia law. While morals and ethics in Islamic laws generally tend to shape women lives through antagonism toward sex and sexual relations and explain the very low level of access to the women, there is however, no strong evidence that this may have a direct impact on risky sexual behavior among Muslims. Ethnic comparison associates the Yoruba with the greatest incidence of extramarital sexual activity when compared to the Ibos and Hausas. ³¹ This is because the Yorubas tend to be more permissive of both male and female infidelity. ^{32, 33} The protective odds of HIV prevention self-efficacy noted among participants of Yoruba ethnicity may be related to high exposure to western education, which tend to empower them with control over their sexuality including ability to negotiate condom use.

Consistent condom use is the most effective way to reduce exposure to HIV and other sexually transmitted diseases among sexually-active individuals. Approximately, 50% of the respondents in our sample had positive condom use attitudes and behaviors, possibly because this measure composed of beliefs with opposing valences that condoms are both effective and also reduce pleasure.³⁴ Negative attitudes toward condoms were generally associated with irregular or non-use of condoms by some participants. Despite their negative attitudes, those participants claimed to use condoms when they perceived their partner to be of high-risk. Some participants reported that they dislike condoms because its use reduces sexual pleasure, a reason that has commonly been cited by other researchers for none use of condoms during sexual encounters. ^{28, 35, 36}

In line with prior qualitative and quantitative research, ^{36, 37} respondents were unwilling to use condoms with their steady partners, because they believed condom use connotes distrust and a lack of intimacy, because they did not feel that their partner was at risk, or because they felt that condom use with long-term partners was unnecessary. In contrast, participants with multiple partners had more positive attitudes, and were more likely to protect themselves during sexual encounters. However, consistent users tended to use condoms with both steady and casual partners, while less frequent condom users preferred to use them with people whom they considered to pose comparatively higher risk such as new partners, and casual partners. But despite the relative good knowledge of this group about sexual transmission of HIV, our analysis indicates that 46% of the participants had sex with casual partners without using condoms. Unprotected sex with a casual partner is a risk factor for HIV infection among mobile populations like the military.¹⁹

A third of the participants reported substance use during sexual encounters, with 18% of those having multiple sexual partners. Our current study also showed that female military personnel who use substances were three times more likely to have multiple partners. Previous studies similarly associated substance use with multiple sex partners.³⁸ The assumptions that alcohol and/or drug use will enhance a person's sexual attraction, behavior, or performance can also have an impact. For example, it has been noted that adolescents who expect alcohol to lead them to be less inhibited sexually are more likely to participate in risky sexual behavior when they drink.³⁹ It is very common to have restaurants and bars around military barracks, where alcohol, and sometime illicit drugs are sold. Such social environments also support the meeting of new sexual partners and may help to explain the relationship between substance use and the likelihood of having multiple partners. We noted in our study, that participants with multiple partners were three times more likely to use alcohol and drugs during sexual encounters than those who had a single partner. Similarly, widespread alcohol use and sexual relationships in the context of alcohol were noted among Nigerian soldiers with no significant difference between the genders. ¹³ If substance use leads to unsafe sexual activity, understanding the dynamics of this relationship can contribute to research and preventive and educational efforts to contain the spread of HIV.

If sexual risk taking is caused by lessened inhibitions due to substance use, then education might warn about the impact of alcohol and drugs on one's judgment and the potential consequences of such situations, such as the increased risk of STD and HIV transmission. Unfortunately, our study found that although participants with high school education and above had higher levels of HIV knowledge, they were approximately four times more likely to use alcohol and drugs than those who had less than high school education. It's been also reported that not all types of risky sexual behavior were avoided with increased levels of education. ²³ Therefore, it is important to distinguish the differential impact that education has on different sexual behaviors, including condom use and multiple sexual partnership.

Implications for Interventions

Our findings from this investigation have implications for the design of HIV/AIDS prevention interventions for Nigerian women in the military service. The Nigerian military authority should conduct regular and sustained STD and HIV prevention education programs among military personnel and their families in the barracks and schools to reinforce health promoting behaviors. Such programs should include HIV prevention self-efficacy, and also the addition of an HIV positive peer educator should be considered. Female military personnel have already been identified as desirable peer educators than their male counterparts based on their educational attainments. ¹³ Efforts should be focused on those segments of the military population shown to have lower knowledge levels, lower levels of preventive actions, low self-efficacy for condom use, and higher exposure to risk.

Self-efficacy has been identified as an important component of condom use ^{40, 41} and thus, the low levels of HIV prevention self-efficacy reported in this group is a concern. Since the demands of sexual communication may vary across sexual encounters based on the partner, the intimacy level, and a host of other situational factors, there is need to gear risk reduction self-efficacy interventions (sexual communication training) toward specific contexts. Specifically, the differing cultural and religious backgrounds of the female military personnel could be examined within the context of an intervention to articulate and discuss cross-cultural sex role norms, reduction in number of sexual partners and consistent use of condoms. This may inform individuals and their partners of underlying social influences that impact how each person approaches sexuality. Understanding these factors is critical for the design of a culturally and contextually tailored intervention for the reduction of sex-related health risks among Nigerian military personnel and their families.

Limitations

The data have the usual limitations of sexual behavior research, being self-reported information it is subject to reporting errors and biases, which could not be practically or ethically validated. They are also cross-sectional in nature, with current measures of psychosocial variables being used to predict reports of past behavior. Thus, while the study provides useful information about the associations between variables, conclusions should not be drawn about causation or prediction. Also, since the study used military cantonments from one of the six regions in Nigeria, the findings may not be generalizable or representative of the female military personnel in Nigeria. Future studies are needed to assess the effectiveness of HIV prevention interventions among this venerable subpopulation.

Conclusion

Despite the relatively good knowledge about sexual transmission of HIV, there was a low level of confidence in HIV prevention self-efficacy and many in this subpopulation engaged in behaviors that elevate their risk of exposure to HIV infection. Since these female service personnel live and interact freely with civilian population they represent a potential bridging group for disseminating HIV into the larger population. Although there were no consistent associations of the correlates to a number of the outcome variables considered in our study, targeted intervention programs for this group, taking the variables associated with the outcome measures into account could help to minimize the consequences of the epidemic. Since female military personnel are more vulnerable to HIV transmission through sex with infected partners, efforts are needed to ensure that their needs are met through gender-sensitive HIV intervention programs. The Nigeria military service, like many others, has an organized structure that could provide an excellent opportunity for the implementation of such intervention programs, which could help reduce military personnel chances of exposure to HIV. It is recommended that further studies of sexual behaviors in this group be carried out using a larger sample size and considering other correlates in addition to those identified in the current study.