HIV prevalence and risk in long-distance truck drivers in South Africa: a national cross-sectional survey

Abstract

We estimated the prevalence of HIV and assessed correlates of HIV infection in long-distance truck drivers in South Africa. Between October 2003 and July 2004, 1900 long-distance truck drivers aged ≥18 years consented to interview and for testing for HIV. Participants were selected from a 10% stratified random sample of registered truck depots. A proximate-determinants framework was used to assess the hierarchical relationship between risk factors and HIV infection using logistic regression. HIV prevalence was 26% (95% confidence interval 24% to 28%). In multivariate analyses, HIV infection was associated with spending 2–4 weeks on the road (adjusted odds ratio 1.4; 95% confidence interval 1.1 to 1.9). There was modest evidence of a dose-response relationship between time on the road and HIV risk. Mobility increased risk by creating conditions for unsafe sex and reducing access to health services. Targeted HIV interventions for long-distance truck drivers are needed.

Keywords

HIV human immunodeficiency virus AIDS Africa location epidemiology high-risk behavior viral disease men truck drivers

Introduction

High rates of sexually transmitted infections (STIs), including HIV, have been documented in long-distance truck drivers in Africa,^1–3 Asia,^4–6 and the Americas,^7–9 and their sexual behaviour is thought to place them at high risk for infection. While on the road, truck drivers may be absent from their regular partners for long periods of time, less subject to social constraints, more likely to select commercial sex partners, and be exposed to high-risk sexual networks.^10–12 Truck drivers may have limited access to health services and commodities, including prompt and effective treatment of STI, condoms, and other prevention interventions, while on the road.^13–15

Truck drivers in South Africa are an important group for targeted HIV interventions. The spread of HIV in South Africa has been attributed in part to efficient transport networks which have linked areas and populations of high prevalence to those of lower prevalence, leading to subsequent dissemination of HIV within the general population.^16,17 Most of the 3000 registered road freight companies in South Africa are private and are estimated to operate over 200,000 vehicles.¹³ Despite the size of this industry, only one study involving 100 truck drivers has been published. In this study, low levels of condom use (29%), high levels of casual sex (37%), and high levels of self-reported STIs (66%) and HIV (56%) were observed.¹⁸ In 2003, the Department of Health and the National Bargaining Council and the Road Freight Association (NBCRFI), which represents the vast majority of small (1 to 10 drivers), medium (11 to 99 drivers), and large (100 or more drivers) companies within the road-freight industry in South Africa, piloted a programme to provide STI services at six clinics established on major truck routes around the country.¹⁹ Given the paucity of nationally representative data, we conducted a survey to determine the prevalence of HIV, STIs, and associated risk behaviour in long-distance truck drivers in South Africa to inform the development and expansion of HIV prevention and treatment interventions for truck drivers. Specifically, we examined the effects of time on the road on sexual risk behaviour and access to health services.

Methods

Study design and population

A nationally representative cross-sectional survey of 1900 truck drivers was conducted between October 2003 and July 2004. The study enrolled those who were 18 years or more, employed full-time as a long-distance trucker, co-driver or assistant in NBCRFI-affiliated organisations, and consented to be interviewed and provide urine and saliva samples for testing.

Sampling strategy

A two-stage stratified sampling design was employed. The sampling frame for the first stage included the 1988 depots of all companies affiliated with the NBCRFI. The NBCRFI provided the complete list of depots with 67% classified as small (<10 drivers), 31% as medium (10 to 99 drivers), and 2% as large (100 or more drivers). A 10% random sample of depots (n = 199) was drawn from the list, stratified by small, medium, and large depots.

In stage two, the study team updated information from depot representatives on the exact number of drivers who worked at the depot and the days on which drivers would be available for interview. Field workers arranged a private venue for interviews and confirmed interview days with depot staff. Interviewers returned to each depot until interviews with all eligible drivers were completed.

Data collection

Data were collected from all consenting adult truck drivers during a 45-minute face-to-face interview using a structured questionnaire that had been field-tested and translated into local languages. The questionnaire contained items on socio-demographics, sexual behaviours, medical history, and knowledge and perceptions of HIV. Saliva and urine samples were collected and frozen within 24 hours and sent from each depot to Johannesburg for testing in a central laboratory. HIV testing was performed on saliva using the Orasure Oral Specimen Collection Device (Orasure Technologies, Inc., Beaverton, Oregon, USA). Urine was tested by PCR technique, using Roche Cobas Amplicor CT/NG Test (Roche Diagnostics, Branchburg, NJ, USA).

Statistical analysis

Data were double-entered using Epi-Info and statistical analyses were performed using Stata version 10.0 (StataCorp, 2008, Texas, USA). Descriptive statistics summarise socio-demographic, risk behaviour, and infection data.

The proximate-determinant framework,²⁰ which categorises potential HIV risk factors hierarchically into underlying, proximate, and biological determinants of HIV infection, was used to explore the relationship between mobility and HIV infection, and to identify other independent risk factors for HIV infection and the pathways through which they act (Figure 1).

Figure 1.

The four models constructed and variables modelled as underlying and proximate determinants*, using the proximate determinant framework.²⁰ *Significant variables in bold.

Associations between HIV and risk factors categorised as potential underlying and proximate determinants were assessed using multiple logistic regression with Taylor linearized variance estimation to account for sampling strategy. Several models were built which explored the hierarchical relationship between risk factors and HIV infection (Figure 1). First, the associations between the proximate determinants and HIV infection were investigated (model 1). Then, a model for the association between the underlying determinants and HIV infection was developed (model 2). Thereafter, a full model containing significant underlying determinants, adjusted for significant proximate determinants, was evaluated (model 3). Like other investigators,²¹ we hypothesised that if the proximate determinants had been measured sufficiently well, then the significance of the underlying determinants would be removed if they were on the causal pathway to HIV infection. Finally, the associations between underlying determinants and proximate determinants were explored for two of the key proximate determinants (recent unsafe sex and history of STIs) (model 4). For this analysis, recent unsafe sex which was defined as either “no risk” (no multiple concurrent partnerships and no reported sex with sex workers), “low risk” (concurrency but condom use at last sex) or “high risk” (concurrency and no condom use at last sex) in the past 6 months was subsequently categorised as either no risk or risky, where having any concurrent relationship was categorised as risky. Associations between exposure and outcome in the adjusted analyses were assessed using multivariate logistic regression with results reported as adjusted odds ratios (aOR) and 95% confidence intervals (95%CI).

Ethical considerations

The study was reviewed and approved by the University of the Witwatersrand Human Research Ethics Committee. Participants received R50 reimbursement (∼USD 5) for their participation.

Results

Of the 199 NBCRFI-affiliated depots selected for interviews, 60 did not respond after several attempts to contact them, 17 did not have contact information, 11 refused participation, and 10 were no longer operating. Results are presented for 1896 participants who reported sexual experiences; four participants who reported no sexual debut were excluded from analysis.

Description of participants

All respondents were male and had a median age of 39 years (range 18–71), had mostly completed or partially completed secondary education (80%), were mostly South African (96%), and had worked as drivers for a median 8 years (interquartile range [IQR] 3–15). Over half had spent more than a week on the road in the previous month, with 81% away from home each week in the past year. A total of 44% reported travelling more than 12,000 km in the previous month. Median income per month was R3225 (∼460 USD; IQR ZAR 2645–4711).

At time of interview, 69% were married or living with a partner. Half reported having 10 or more sex partners in their life (IQR 5–20). In all, 46% reported concurrent partnerships in the previous 6 months, 30% had ever had sex with a sex worker, and 1% ever had sex with another man. Reported condom use at last sex was high with sex workers (92%) and non-steady partners (83%), but considerably lower with current steady partners (9%). Based on study definitions, 50% of the participants were considered to have had recent unsafe sex.

Prevalence of HIV, N. gonorrhoeae, and C. trachomatis

The prevalence of HIV, C. trachomatis (CT) and N. gonorrhoeae (NG) was 26% (95% CI 24% to 28%), 7% (95% CI 6% to 8%), and 2% (95% CI 1% to 3%), respectively. Of those who had CT/NG, only 11% gave a history of genital discharge in the past 12 months, suggesting that the majority of infections were asymptomatic.

Associations between proximate determinants of infection and HIV infection

In model 1, risk of HIV infection was highest in those who practised genital cleaning before or after sex, reported a history of STIs, or who reported recent unsafe sex, even after adjustment for covariates (Table 1). A dose-response relationship was observed between recent unsafe sex and HIV infection, with the rate of infection highest for men in the “high risk” group (concurrency and no condom use at last sex) and intermediate in the “low risk” group, compared to “no risk” men. Circumcised men were less likely to have HIV (aOR 0.7, 95% CI 0.6 to 0.9).

Table 1.

Association between proximate determinants and HIV infection (N = 1880).

Determinant (n)	% HIV seropositive	Unadjusted OR (95% CI)	Adjusted OR (95% CI) (Model 1)^a [n = 1660]
Age of sexual debut
Less than 17 (694)	26.4	1
17–19 (795)	27.6	1.1 (0.8–1.5)
20 and above (365)	24.4	0.9 (0.6–1.3)
No. of lifetime sexual partners
1–5 (356)	19.1	1
6–10 (392)	26.3	1.5 (1.0–2.2)*
11–20 (331)	32.9	2.1 (1.4–3.2)**
21 and above (342)	36.3	2.4 (1.4–4.1)**
Partner age group (years)
18–24 (102)	24.5	1
25–34 (499)	31.5	1.4 (0.8–2.6)
35–49 (761)	23.9	1.0 (0.5–1.7)
50 and above (123)	12.2	0.4 (0.17–1.1)
Male sexual partner^b
No (1473)	28.9	–	–
Yes (14)	0
Recent unsafe sex
No risk (937)	18.7	1	1
Low risk (561)	31.3	2.0 (1.4–2.7)**	1.4 (1.0–2.0)*
High risk (373)	37.3	2.6 (1.9–3.4)**	1.8 (1.4–2.3)**
Consumed alcohol in the past 4 weeks
No (1084)	25.6	1
Yes (734)	27.5	1.1 (0.8–1.6)
Circumcised
No (1071)	28.6	1	1
Yes (807)	23.4	0.8 (0.6–1.0)*	0.7 (0.6–0.9)*
Genital cleaning before or after sex
No (1753)	12.7	1	1
Yes (126)	27.4	2.6 (1.1–6.2)*	2.4 (1.1–5.4)*
Ever had an STI
No (754)	17.6	1	1
Yes (917)	35.9	2.6 (1.8–3.7)**	2.1 (1.5–2.9)**
History of genital ulcers in past year
No (1746)	25.3	1
Yes (134)	41	2.1 (1.2–3.4)**
History of genital discharge in past year
No (1753)	25.2	1
Yes (127)	44. 1	2.3 (1.5–3.6)**
Current CT/GC
No (1725)	26	1
Yes (157)	31.2	1.3 (0.8–2.1)

OR: odds ratio; p values are indicated for each OR; *≤0.05; **≤0.01.

Adjusted for recent unsafe sex, genital cleaning, history of STIs, history of genital ulcers in past 12 months, and circumcision status.

OR could not be calculated as none of the 14 respondents who reported sex with men were HIV-seropositive.

CT, Chlamydia trachomatis; GC, Neisseria gonorrhoeae

Associations between underlying determinants and HIV infection

In model 2 (Table 2), men older than 24 had a higher risk of HIV infection compared to men aged 18–24 years. Being unmarried and living with a sexual partner was associated with the highest odds of HIV infection (aOR 2.2, 95% CI 1.4 to 3.4), relative to married men living with their spouses. Married and unmarried men who did not live with a sexual partner were also at increased risk of HIV infection. Being resident in one of three provinces in South Africa increased the odds of HIV infection, with men resident in KwaZulu-Natal having the highest risk of infection (aOR 9.8, 95% CI 2.9 to 33.2), followed by men living in Gauteng (aOR 7.5, 95% CI 2.2 to 25.7) and the Free State (aOR 7.1, 95% CI 1.9 to 26.6). Men currently living outside of South Africa also had an increased risk of HIV (aOR 6.7, 95% CI 8.17 to 25.9). Conversely, men whose primary language was Setswana, Sepedi, English, or Afrikaans had lower odds of infection relative to Zulu speakers (Table 2).

Table 2.

Association between underlying determinants and HIV infection (N = 1880).

Determinant (n)	HIV (%)	Unadjusted OR (95% CI)	OR (95% CI), adjusted for underlying determinants [n = 1797]	OR (95% CI), adjusted for underlying and proximate determinants [n = 1607]
Age
18–24 (78)	12.8	1	1	1
25–34 (526)	28.8	2.7 (1.5–5.2)**	3.0 (1.5–5.8)**	3.0 (1.5–6.2)**
35–49 (960)	27.2	2.5 (1.2–5.1)*	3.9 (1.9–8.1)**	3.9 (1.7–8.9)**
50 and above (320)	23.8	2.1 (1.0–4.2)*	3.5 (1.6–7.5)**	2.5 (1.5–8.2)**
Marital status
Married living with spouse or sexual partner (1135)	21.7	1	1	1
Married lives alone (206)	37.4	2.2 (1.4–3.3)**	1.7 (1.0–2.7)*	1.6 (1.0–2.6)*
Unmarried living with sexual partner (162)	37	2.1 (1.4–3.3)**	2.2 (1.4–3.4)**	2.4 (1.5–3.7)**
Not married not living with sexual partner (375)	30.7	1.6 (1.2–2.1)**	1.8 (1.3–2.4)**	1.6 (1.2–2.3)**
Current province of residence
Eastern Cape (81)	3.7	1	1	1
Free State (91)	28.6	10.4 (3.0–35.6)**	7.1 (2.1–23.7)*	10.1 (1.0–105)
Gauteng (770)	30.3	11.3 (3.3–38.8)**	7.4 6 (2.2–25.7)*	12.1 (1.1–127)*
Kwazulu Natal (445)	40	17.3 (4.9–61.0)**	9.8 (2.9–33.2)**	17.0 (1.6–180)*
Limpopo (90)	21.1	7.0 (2.0–24.2)**	5.6 (1.4–20.8)	8.3 (0.7–93)†
Mpumalanga (55)	23.6	8.0 (2.0–31.9)**	5.7 (1.3–24.5)	11.2 (0.9–133)
North West (57)	10.5	3.1 (0.6–17.0)	2.8 (0.6–13.2)	5.0 (0.4–66)
Northern Cape (61)	4.9	1.3 (0.2–8.2)	***	***
Western Cape (203)	5.4	1.5 (0.4–5.4)	3.8 (1.0–14.3)	5.4 (0.5–58)
Other Southern African country (13)	23.1	7.8 (1.4–44.0)*	6.7 (1.7–25.9)**	8.9 (0.8–104)
Primary language
Zulu (556)	43.4	1	1	1
Sesotho (219)	31.1	0.6 (0.4–0.9)*	0.8 (0.5–1.2)	0.8 (0.5–1.3)
Xhosa (168)	26.2	0.5 (0.3–0.8)**	0.8 (0.5–1.3)	1.0 (0.5–1.8)
Setswana (116)	22.4	0.4 (0.3–0.5)**	0.5 (0.4–0.8)**	0.6 (0.3–0.9)*
sePedi (173)	22	0.4 (0.3–0.5)**	0.5 (0.3–0.7)**	0.5 (0.3–0.8)**
Ndebele (31)	25.8	0.5 (0.2–0.9)†*	0.7 (0.3–1.3)	0.7 (0.3–1.5)
Siswati (43)	25.6	0.5 (0.2–0.9)*	0.6 (0.3–1.3)	0.6 (0.3–1.3)
Venda (22)	31.8	0.6 (0.3–1.2)	0.8 (0.3–1.7	1.0 (0.4–2.4)
Tsonga (44)	29.6	0.6 (0.3–1.0)†	0.8 (0.4–1.3)	0.8 (0.5–1.5)
English (198)	12.6	0.2 (0.1–0.4)**	0.3 (0.1–0.5)**	0.3 (0.1–0.7)**
Afrikaans (294)	3.4	0.05 (0.02–0.09)**	0.1 (0.04–0.2)**	0.1 (0.03–0.2)**
Other (16)	31.3	0.6 (0.2–1.6)	1.8 (0.4–8.8)	1.8 (0.4–7.4)
Time on the road in past 30 days
Less than 1 week (818)	22.6	1	1	1
1–2 weeks (427)	23.9	1.1 (0.8–1.5)	1.1 (0.8–1.6)	1.2 (0.9–1.6)
2–4 weeks (375)	28.8	1.4 (0.9–2.2)	1.4 (1.1–1.9)*	1.2 (0.9–1.7)
4 weeks (227)	41.0	2.4 (1.4–4.0)**	1.5 (0.9–2.3)	1.5 (0.9–2.5)
Not on the road (32)	25	1.1 (0.3–4.2)	0.9 (0.2–3.3)	0.9 (0.3–3.1)
Another home
No (1134)	19.7	1
Yes (747)	36.7	2.4 (1.6–3.4)**
Ever heard of sexually transmitted infections
No (184)	13.6	1
Yes (1697)	27.8	2.5 (1.4–4.15)**
Perceived risk of HIV
No (1100)	23.1	1
Yes (715)	31.3	1.5 (1.2–1.9)**
Don’t know (61)	29.5	1.4 (0.8–2.4)
History of HIV testing
No (1162)	30	1	1	1
Yes (719)	20.7	0.6 (0.5–0.8)**	0.7 (0.5–1.0)*	0.7 (0.5–1.0)*

OR: odds ratio; p values are indicated for each OR: †≤0.1; *≤0.05; **≤0.01, ***adjusted model could not be fitted.

Indicators of mobility including years as a truck driver, time on the road in the past 30 days, distance travelled in the past 30 days, having another home, and frequency of time away from home were assessed for their relationship with HIV infection. Spending 2–4 weeks on the road in the past 30 days was significantly associated with HIV infection, even after adjustment. There was evidence of a dose-response relationship between time on the road and risk of HIV infection, with those who spent 4 weeks on the road in the past 30 days having 1.5 times the odds of HIV infection compared to those that spent less than a week on the road (95% CI 0.9 to 2.3).

Knowledge and health-seeking behaviour were also assessed. Almost 60% of men did not consider themselves to be at risk for HIV infection. Of these, 23% were HIV-positive. A prior history of HIV testing was the only variable that remained independently associated with HIV infection when compared with those that reported no previous HIV testing (aOR 0.7, 95% CI 0.5 to 1.0).

Associations between underlying determinants and HIV infection, after adjustment for proximate determinants

Model 3 assessed the hierarchical relationship between proximate and underlying determinants and HIV. By adjusting for proximate determinants in this model it was possible to assess whether the observed relationship between underlying factors and HIV was mediated by the proximate determinants of infection. After adjusting for recent unsafe sex, genital cleaning, a history of STIs, and circumcision status, five of the six previously significant underlying factors remained independently associated with HIV infection; specifically being older than 24 years, marital status, current province of residence, primary language, and prior history of HIV testing (Table 2). The relationship between residence in Gauteng or Kwazulu-Natal and HIV infection strengthened in this model. As hypothesised, time on the road in the past 30 days was no longer significantly associated with HIV infection in this model.

Associations between underlying determinants and proximate determinants

Finally, previously identified significant underlying determinants of HIV infection, namely recent unsafe sex, and a history of STIs were modelled as separate outcome variables (Table 3; Model 4). There was strong evidence that marital status, place of residence, and speaking Ndebele were positively correlated with recent unsafe sex, while prior history of HIV testing was associated with lower odds of recent unsafe sex. There was some evidence that time on the road was predictive of recent unsafe sex, with a suggestion that risk increased with increasing time away from home, although the evidence for this was less strong. Being older than 24 years, marital status, and place of residence were associated with a history of STIs. There was a suggestion that time on the road was also associated with a history of STIs, but the evidence for this was weak.

Table 3.

Association between selected proximate and underlying determinants of HIV infection.

Determinant	Unsafe sex		Ever had STI
Determinant	Unadjusted OR (95% CI)	Adjusted OR (95% CI)^a	Unadjusted OR (95% CI)	Adjusted OR (95% CI)^a
Age group
18–24 (78)	1	1	1	1
25–34 (526)	0.7 (0.4–1.3)	0.8 (0.4–1.5)	2.2 (1.4–3.6)**	2.4 (1.4–4.2)**
35–49 (n = 960)	0.4 (0.2–0.9)*	0.7 (0.4–1.4)	2.9 (1.7–5.0)**	4.1 (2.1–7.9)**
50 and above (320)	0.3 (0.2–0.6)**	0.4 (0.2–0.8)*	2.5 (15–4.2)**	3.2 (1.7–5.8)**
Marital status
Married living with spouse or sexual partner (1135)	1	1	1	1
Married not living with spouse or sexual partner (206)	2. (1.5–3.1)**	1.7 (1.2–2.4)**	2.4 (1.7–3.5)**	1.9 (1.3–2.7)**
Not married living with sexual partner (162)	2.4 (1.6–3.4)**	2.0 (1.3–2.9)**	1.5 (1.0–2.2)*	1.8 (1.1–2.8)*
Not married not living with sexual partner (375)	3.2 (2.4–4.3)**	2.7 (2.0–3.8)**	1.2 (0.9–1.6)	1.4 (1.1–1.9)*
Current province of residence
Eastern Cape (81)	1	1	1	1
Free State (91)	5.1 (2.2–11.9)**	3.7 (1.8–8.1)**	18.1 (2.8–118)**	10.1 (1.8–57.4)**
Gauteng (770)	4.5 (2.1–9.4)**	4.0 (2.2–7.4)**	23.4 (3.4–161.)**	11.4 (1.9–68.6)**
Kwa-Zulu Natal (445)	4.4 (2.0–9.7)**	4.1 (2.3–7.4)**	26.4 (4.0–173)**	14.8 (2.5–87.6)**
Limpopo (90)	5.7 (2.4–13.2)**	5.5 (2.4–12.5)**	37.8 (4.8–300**	14.6 (2.0–105)**
Mpumalanga (55)	3.9 (1.8–8.3)**	3.2 (1.5–6.7)**	32.8 (4.5–241)**	15.6 (2.4–101)**
North West (57)	1.4 (0.6–3.1)	1.6 (0.6–3.8)	13.9 (1.6–118)**	10.4 (1.5–73.1)**
Northern Cape (61)	1.6 (0.7–3.6)	2.8. (1.4–5.7)	3 (0.4–22.6)	4.5 (0.6–36.8)
Western Cape (203)	1.2 (0.6–2.8)	2.6 (1.2–5.6)	3 (0.4–20.4)	5.9 (1.0–36.5)
Southern Africa (13)	11.7 (2.3–60)**	15.7 (3.6–67.9)**	40 (5.1–314)**	32.4 (4.6–231)**
Primary language
Zulu (556)	1	1	1	1
Sesotho (219)	1.0 (0.7–1.4)	1.3 (0.9–1.7)	0.5 (0.4–0.7)	0.6 (0.4–0.8)**
Xhosa (168)	0.7 (0.5–1.0)*	1.0 (0.7–1.4)	0.4 (0.2–0.7)	0.6 (0.4–1.0)
Setswana (116)	0.6 (0.4–1.0)	1.0 (0.5–1.7)	0.6 (0.4–0.9)	0.7 (0.4–1.2)
sePedi (173)	0.9 (0.6–1.2)	0.9 (0.5–1.3)	0.9 (0.5–1.4	0.9 (0.5–1.4)
Ndebele (31)	1.9 (0.8–4.6)	3.1 (1.2–8.3)*	1.3 (0.6–2.9)	1.2 (0.5–3.2)
Siswati (43)	1.1 (0.5–2.2)	1.3 (0.5–3.6)	0.6 (0.3–1.3)	0.6 (0.3–1.3)
Venda (22)	1.0 (0.3–3.3)	1.0 (0.3–3.3)	0.8 (0.3–2.0)	0.7 (0.3–2.0)
Tsonga (44)	1.1 (0.5–2.1)	1.2 (0.–2.3)	0.7 (0.4–1.3)	0.8 (0.4–1.7)
English (198)	0.3 (0.2–0.5)**	0.5 (0.3–0.7)**	0.1 (0.08–0.2)	0.1 (0.1–0.2)**
Afrikaans (294)	0.2 (0.1–0.3)**	0.4 (0.3–0.6)**	0.05 (0.03–0.08)	0.1 (0.06–0.2)**
Other (16)	1.1 (0.5–2.5)	1.6 (0.3–8.8)	0.8 (0.3–2.2)	0.7 (0.2–2.8)
Time on the road in past 30 days
Less than 1 week (818)	1	1	1	1
1–2 weeks (427)	0.8 (0.6–1.1)	0.9 (0.6–1.2)	0.8 (0.6–1.2)	0.9 (0.7–1.2)
2–4 weeks (375)	0.9 (0.6–1.4)	1.2 (0.9–1.6)	0.9 (0.6–1.6)	1.0 (0.8–1.4)
4 weeks (227)	1.8 (1.2–2.6)**	1.8 (1.1–2.5)*	1.6 (1.0–2.8)	1.2 (0.7–2.1)
Not on the road (32)	0.9 (0.3–3.0)	0.8 (0.2–2.6)	1.9 (0.6–5.7)	1.8 (0.6–5.1)
Previous HIV test
No (1162)	1	1	1
Yes (719)	0.6 (0.5–0.7)**	0.7 (0.6–0.9)**	0.7 (0.6–0.9)**

OR, Odds Ratio; p values for each OR; *≤0.05; **≤0.01.

Adjusted for significant determinants in the respective models.

Discussion

To our knowledge, this is the largest representative cross-sectional study of long-distance truck drivers in South Africa, and to our knowledge, in Africa. The HIV prevalence of 26% amongst truck drivers is almost three times higher than the prevalence observed in the same age group of men in the general population in South Africa in 2005.²² We observed a dose-response relationship between length of time on the road and HIV infection, underlining the fact that longer periods away from home lead to riskier behaviours in truck drivers. Adjusting for significant proximate determinants in the underlying determinants model removed the relationship between time spent on the road and HIV infection, demonstrating that the association between time spent on the road and HIV is mediated through one or more of the significant proximate determinants that were identified. Time on the road thus does not lead to increased HIV infection by itself, but rather works through factors that either increase the risk of exposure to infection e.g. recent unsafe sex or that increase the probability of HIV transmission e.g. history of STIs, lack of male circumcision. While several studies have demonstrated the relationship between mobility and HIV,^17,23,24 to our knowledge, this is the first to use a conceptual framework to formally structure the analysis and thus confirm the hierarchical relationship that exists between them.

Truck drivers who were married and lived with their spouses had a lower risk of HIV infection, and were less likely to report recent unsafe sex or have a history of STIs, suggesting that being in a stable relationship reduces the likelihood that men will engage in risky sexual behaviour while on the road. Other studies of truck drivers and mobile men have found that married men were more likely to use condoms with non-spousal partners²⁵ and less likely to be at risk of new HIV infection than unmarried men.²⁶ Structural interventions such as workplace-based policies that reduce time on the road and periods of absence from home may have an important impact on HIV risk in this population.

Prevention interventions targeted at most at risk populations are cost-effective.^27,28 In South Africa, long distance truck drivers are an easily identifiable occupational cohort with a higher risk for HIV than the general population. Interventions to reduce unsafe sex and to promote access to health services are a priority for this population. At the time of the survey, several roadside clinics had recently been established on major highways but less than 50% of respondents were aware of these centres and an even smaller proportion reported using them, suggesting the need for greater promotion of these services within the industry. As a result of this survey, roadside services were expanded from 6 to 22 clinics currently to cover the national road network.²⁹ Expansion of services provides the opportunity to promote a range of effective HIV-prevention interventions. In this study, male circumcision was associated with 30% lower odds of infection, but 57% of men were not circumcised. Previous HIV testing was also associated with a lower HIV risk but only 38% of drivers had tested for HIV before. Promotion of HIV testing and linkage to care is a priority in this population where almost a quarter of those who were surveyed and did not consider themselves to be at risk for HIV infection were HIV-positive. Early initiation of HIV treatment may be an important HIV prevention intervention in this population and should be evaluated further. ³⁰ Efforts by the NBCRFI to expand access to HIV treatment for truck drivers subsequent to this survey should be acknowledged. Men who reported living outside of South Africa had amongst the highest rates of HIV infection and were also at highest risk for recent unsafe sex and a history of STIs arguing for interventions to be extended transnationally.

Place of residence was strongly associated with HIV infection, even after adjustment for other underlying determinants, and for proximate determinants of infection. This finding reflects the importance of epidemiological context in determining risk of HIV infection, and suggests that sexual risk-taking may not only occur in transit. In provinces with higher HIV prevalence such as KwaZulu-Natal or Gauteng, there is a greater chance that a sexual partner chosen from these communities will be HIV-infected than in areas of lower HIV prevalence. A community-based study in Zimbabwe, which used a similar analytical framework, found that community prevalence of HIV in the opposite sex was predictive of HIV infection.²¹ Similar findings for primary language spoken probably also reflect the combination of epidemiological context and other unmeasured socio-cultural factors that may be associated with increased risk.

This study has some limitations. A few underlying determinants remained significantly associated with HIV in the fully adjusted model, suggesting that these underlying determinants work through proximate determinants other than those measured in this survey, such as partner HIV status, for example. Other studies using this approach for analysis have shown that partnership characteristics are important predictors of individual risk.^21,31 The survey was an interviewer-administered survey and may have suffered from social desirability bias or poor recall in responses to sensitive questions. The low response rate of depots may have biased the sample in favour of drivers working for companies that more strongly encourage health-seeking behaviour amongst their workforce; given this, the survey is likely to underestimate HIV risk. Finally, the cross-sectional nature of this survey limits our ability to determine the temporal sequence of risks in relation to HIV infection.

In conclusion, this population-based survey of long-distance truckers to our knowledge is the first of its kind in South Africa and provides a precise estimate of HIV infection in this population. The data confirm the critical role that occupational mobility plays in promoting ongoing HIV risk behaviour as well as reducing access to health services, highlighting the need for enhanced industry-based targeted interventions for this population.

Authors’ contributions

SD, HR designed the study and MO supervised the implementation. IK provided guidance on the sampling frame and sampling strategies. BB analysed the data with assistance from PK and MC. SD, BB and PK interpreted the data and prepared the first draft of the manuscript which was approved by all authors.

Footnotes

Acknowledgements

We would like to acknowledge the participants and depot managers who participated in this survey; the South African National Department of Health and the National Bargaining Council of the Road Freight Industry (NBCRFI) who supported the project from its inception; the staff at Development Research Africa (DRA) who conducted the field work for the survey and entered the data; Contract Laboratory Services (CLS) who conducted all the laboratory work for the survey; Nomampondomise Koetle and Kgopotso Mokgope who assisted with project administration; and Harry Moultrie, Heena Brahmbhatt and Landon Myer for comments on the manuscript.

Conflict of interest

The authors declare no conflict of interest.

Funding

This project was funded through the USAID Bilateral Grant Agreement Number 674-0320-G-00-5053–10 Equity in Integrated Primary Health Care (EQUITY). The opinions expressed herein are those of the authors and do not necessarily reflect the views of USAID. The NBCRFI had no part in the analysis of the data or the interpretation of the results of this study.

References

Bwayo

Plummer

Omari

. Human immunodeficiency virus infection in long-distance truck drivers in east Africa. Arch Intern Med 1994; 154: 1391–136.

Mosoko

Macauley

Zoungkanyi

. Human immunodeficiency virus infection and associated factors among specific population subgroups in Cameroon. AIDS Behav 2009; 13: 277–287.

Morris

Ferguson

. Sexual and treatment-seeking behaviour for sexually transmitted infection in long-distance transport workers of East Africa. Sex Transm Infect 2007; 83: 242–245.

Pandey

Benara

Roy

. Risk behaviour, sexually transmitted infections and HIV among long-distance truck drivers: a cross-sectional survey along national highways in India. AIDS 2008; 22: S81–S90.

Chen

Yin

Gong

. Prevalence of sexually transmitted infections among long-distance truck drivers in Tongling, China. Int J STD AIDS 2006; 17: 304–308.

Mustikawati

Morineau

Nurhayati . Sexual risk taking, sexually transmitted infections and HIV prevalence among four “high-risk” occupational groups of Indonesian men. Sex Transm Infect 2009; 85: 391–396.

Pulerwitz

Michaelis

Lippman

. HIV-related stigma, service utilization, and status disclosure among truck drivers crossing the Southern borders in Brazil. AIDS Care 2008; 20: 764–770.

Lichtenstein

Hook

3rd Grimley

. HIV risk among long-haul truckers in the USA. Cult Health Sex 2008; 10: 43–56.

Valway

Jenison

Keller

. Risk assessment and screening for sexually transmitted infections, HIV, and hepatitis virus among long-distance truck drivers in New Mexico, 2004–2006. Am J Public Health 2009; 99: 2063–2068.

10.

Lagarde

Schim van der Loeff

Enel

. Mobility and the spread of human immunodeficiency virus into rural areas of West Africa. Int J Epidemiol 2003; 32: 744–752.

11.

Msisha

Kapiga

Earls

. Place matters: multilevel investigation of HIV distribution in Tanzania. AIDS 2008; 22: 741–748.

12.

Lippman

Pulerwitz

Chinaglia

. Mobility and its liminal context: exploring sexual partnering among truck drivers crossing the Southern Brazilian border. Soc Sci Med 2007; 65: 2464–2473.

13.

IOM. Mobile Populations and HIV/AIDS in the Southern African Region: Recommendations for Action. Pretoria: International Organisation for Migration, 2003 May 2003. Report No.

14.

Parker

Easton

Klein

. Structural barriers and facilitators in HIV prevention: a review of international research. AIDS 2000; 14: S22–S32.

15.

Lau

Tsui

Cheng

. A randomized controlled trial to evaluate the relative efficacy of adding voluntary counseling and testing (VCT) to information dissemination in reducing HIV-related risk behaviors among Hong Kong male cross-border truck drivers. AIDS Care 2010; 22: 17–28.

16.

Lurie

Williams

Zuma

. Who infects whom? HIV-1 concordance and discordance among migrant and non-migrant couples in South Africa. AIDS 2003; 17: 2245–2252.

17.

Welz

Hosegood

Jaffar

. Continued very high prevalence of HIV infection in rural KwaZulu-Natal, South Africa: a population-based longitudinal study. AIDS 2007; 21: 1467–1472.

18.

Ramjee

Gouws

. Prevalence of HIV among truck drivers visiting sex workers in KwaZulu-Natal, South Africa. Sex Transm Dis 2002; 29: 44–49.

19.

SouthAfrica.info. Trucking against AIDS 2003, http://www.southafrica.info/ess_info/sa_glance/health/truckingaids.htm (accessed November 2011).

20.

Boerma

Weir

. Integrating demographic and epidemiological approaches to research on HIV/AIDS: the proximate-determinants framework. J Infect Dis 2005; 191: S61–S67.

21.

Lewis

Donnelly

Mare

. Evaluating the proximate determinants framework for HIV infection in rural Zimbabwe. Sex Transm Infect 2007; 83: i61–i69.

22.

Shishana

Rehle

Simbayi

. South African national HIV prevalence, HIV incidence, behaviour and communication survey, Cape Town: HSRC Press, 2005.

23.

Coffee

Garnett

Mlilo

. Patterns of movement and risk of HIV infection in rural Zimbabwe. J Infect Dis 2005; 191: S159–S167.

24.

Hargreaves

Bonell

Morison

. Explaining continued high HIV prevalence in South Africa: socioeconomic factors, HIV incidence and sexual behaviour change among a rural cohort, 2001–2004. AIDS 2007; 21: S39–S48.

25.

Gibney

Saquib

Metzger

. Behavioral risk factors for STD/HIV transmission in Bangladesh's trucking industry. Soc Sci Med 2003; 56: 1411–1424.

26.

Barnighausen

Tanser

Gqwede

. High HIV incidence in a community with high HIV prevalence in rural South Africa: findings from a prospective population-based study. AIDS 2008; 22: 139–144.

27.

Vickerman

Terris-Prestholt

Delany

. Are targeted HIV prevention activities cost-effective in high prevalence settings? Results from a sexually transmitted infection treatment project for sex workers in Johannesburg, South Africa. Sex Transm Dis 2006; 33: S122–S132.

28.

Lafort

Geelhoed

Cumba

. Reproductive health services for populations at high risk of HIV: Performance of a night clinic in Tete province, Mozambique. BMC Health Services Res 2011; 10: 144–144.

29.

NBCRFI. http://www.nbcrfli.org.za/ 2013 (accessed 16 August 2013).

30.

Cohen

Chen

McCauley

. Prevention of HIV-1 Infection with Early Antiretroviral Therapy. New Eng J Med 2011; 365: 493–505.

31.

Lopman

Nyamukapa

Mushati

. HIV incidence in 3 years of follow-up of a Zimbabwe cohort–1998-2000 to 2001-03: contributions of proximate and underlying determinants to transmission. Int J Epidemiol 2008; 37: 88–105.