AIDS Behav (2015) 19:2194–2203 DOI 10.1007/s10461-015-1096-9

ORIGINAL PAPER

Prevalence and Correlates of HIV Infection Among Sex Workers in Papua New Guinea: First Results from the Papua New Guinea and Australia Sexual Health Improvement Project (PASHIP) Handan Wand1 • Peter Siba2

Published online: 28 May 2015  Springer Science+Business Media New York 2015

Abstract The primary objective of this study was to estimate the individual and combined impacts of socio-demographic and sexual behaviours on HIV diagnosis among 523 female sex workers who participated in the Papua New Guinea and Australia Sexual Health Improvement Project. Logistic regression models were used to identify the factors associated with HIV positivity. We estimated their population level impacts in order to quantify the proportion of HIV seropositivity is attributed to these factors. Less than 40 % of women consented to get tested for HIV. HIV prevalence was 7 % (95 % CI 4–11 %); lack of education and knowledge/awareness of HIV accounted for *70 % of the HIV diagnoses. A major obstacle is lack of interest for testing. Our study underscored the major challenges in this culturally, linguistically heterogeneous country. The epidemic in Papua New Guinea requires targeted prevention interventions among those at highest risk of acquiring or transmitting infection. Keywords HIV prevalence  Sex workers  Papua New Guinea  Education

& Handan Wand [email protected] 1

Kirby Institute, University of New South Wales, Sydney, Australia

2

Papua New Guinea Institute of Medical Research, Goroka, Papua New Guinea

123

Introduction Papua New Guinea is one of the most geographically, culturally and linguistically diverse countries in the world. Unfortunately, Papua New Guinea is also known to be disproportionally affected by the HIV epidemic and is estimated to have the highest number of HIV infections in southwestern Pacific Ocean [1–6]. The primary mode of HIV transmission is unprotected heterosexual contact. Factors that increase the risk of HIV transmission are reported to be engaging in transactional sex, multiple concurrent sex partners and diagnosis with other sexually transmitted infections [6, 7]. With estimated prevalence of approximately 1–2 %, HIV has become a generalised epidemic in Papua New Guinea since 2003 [1, 2, 7–9]. In addition to this, mother-tochild transmission is also one of the common routes for HIV acquisition. Annually, 2 % of births in Papua New Guinea General Hospital Labour ward have been reported to be HIV seropositive [10]. Without effective education, awareness and prevention strategies, HIV epidemic is projected to affect 10 % of the general population by 2025 [5]. These figures are estimated to be much higher in high risk populations, such as individuals reporting transactional sex. Female sex workers have been consistently reported to have had the highest HIV infections, particularly in middle and low income countries. In a meta-analysis using data from 50 countries, HIV prevalence is estimated to be 14 times higher among female sex workers compared to the women in the general population [11]. Consistent with these results, in Papua New Guinea, HIV prevalences are estimated to be ten times higher (12–17 %) among groups who have transactional sex compared to the general population [1, 7]. Papua New Guinea is therefore considered to be the country most in need of effective HIV prevention methods.

AIDS Behav (2015) 19:2194–2203

However, developing, implementing and expanding these programs present unique challenges due to its large cultural and linguistic variations. In addition to this, Papua New Guinea is a country with sparsely populated large geographical areas where 90 % of the population is residing in rural areas. Many villages in such areas can only be reached by foot [12]. Lack of human resources, inadequate modern infrastructure and low-grade health services and facilities also bring significant challenges in integrating effective HIV prevention programs. Other factors such as poverty, violence, particularly against women, lack of education may contribute to the increasing number of HIV infections in Papua New Guinea. In addition to low level of information about HIV, a major barrier is lack of interest in testing for HIV, due to stigma and fear of discrimination [11, 12]. According to the World Health Organization estimates, approximately 80 % of HIV infected adults in sub-Saharan Africa are unaware of their status [13]. Also, more than 90 % of them do not know their partners’ HIV status. Whilst such estimates currently are not available in Papua New Guinea, it is estimated to be as high as 80 %. The primary aim of this study was to determine the factors associated with HIV positivity in a cohort of female sex workers who consented to participate in a bio-behavioural survey: The Papua New Guinea and Australia Sexual Health Improvement Project (PASHIP) took place in the Eastern Highlands of Papua New Guinea 2009–2011. This is the first study to report results from the sex workers study, within PASHIP. After more than 30 years of research, we have a good understanding of how HIV is transmitted. The most important challenge is to quantify the patterns of risk factors and their population level impacts, in order to prioritise prevention strategies. To address these issues, we also estimated population attributable risks, i.e. the proportion of HIV diagnoses in the study population could be attributed to the potentially modifiable factors such as socio-economic, sexual behaviours, HIV/AIDS knowledge and beliefs when they were considered separately and combined.

Method Study Design and Population Study recruitment and behavioural data collection protocols have been previously described elsewhere [14]. Briefly, a cross-sectional bio-behavioural survey, PASHIP was conducted among men and women in seven provinces of Papua New Guinea during the period of 2009–2011. The present study included a sub-group of 523 women who identified themselves as sex workers in the province of

2195

Eastern Highlands. Respondent driven sampling was used to recruit women to the study as there was no available sampling frame for this hard-to-reach population. Only those participants who produced their identification card were provided tested for HIV and Syphilis. Behavioral Data Collection The questionnaires were developed based on the ‘‘Guidelines for Repeated Behavioral Surveys in Populations at Risk of HIV’’ [14]. The questionnaire collected information on socio-demographics, sexual behaviors, HIV/AIDS awareness as well as use of various health services [9]. Current analysis included selected socio-demographic characteristics: age (\25 vs. 25 or older), level of education (no formal education vs. some education), marital/cohabitation status (married/cohabiting vs. not married/cohabitating); selected risk behaviors including age at sexual debut (\16 years vs. 16? years), raped in past 12 months (yes/no), frequent alcohol intake (daily/[4 per week) (yes/ no), total number of clients (paying/non-paying) in past 7 days (\5 vs. 5 or more), anal/oral sexual practices in past 12 months (yes/no), condom used (paying/non-paying clients) in past 4 weeks (always vs. some/never), ever heard of HIV/AIDS (yes/no), know transmission route of HIV (yes/no), ever diagnosed/treated with sexually transmitted infections (yes/no). We also studied other risk factors for HIV testing and diagnosis, including age at first sex work, places to meet clients, age of paying/non-paying clients, another job/regular income, number of biological children, consistent use by type of clients (paying/non-paying), condom use in last sexual act, reasons not using condom consistently, places to buy condoms, injecting drugs, using other types of drugs (marijuana, ecstasy, heroin, cocaine). We did not present results in relation to these variables here, as no relationships were detected. Biological Data Collection All participants were given pre-test counselling for HIV and other sexually transmitted infections. 10 ml of venous blood was collected from participants consenting to HIV and syphilis testing. Serum was separated from the blood via centrifugation. All samples were stored at -20 C. Consistent with the National algorithm, syphilis screening was conducted using the Rapid Plasma Reagin (RPR) test (Biotech, Suffolk, UK). Reactive samples were titrated and a titre greater than 1 in 8 was considered to be potentially representative of a current infection. The syphilis status of these samples was confirmed with a treponemal particle haem agglutination assay (TPHA) (Human Diagnostics, Wiesbaden, Germany) [15]. ‘‘HIV testing was performed

123

2196

according to the laboratory based HIV confirmatory algorithm for PNG. Samples were screened with the Serodia HIV1/2 particle agglutination assay (Fujirebio, Tokyo, Japan). All reactive samples were run on the lateral flow Determine HIV1/2 strip (Allere, Australia) before reactive samples were confirmed on the Enzyme Immunoassay, Immunocomb HIV1/2 (Orgenics, Israel). Samples were considered HIV positive if they were reactive on all three tests and discrepant if reactive on either or both of the first two tests but non-reactive on the third. For research purpose, all discrepant samples were tested with the p24 antigen ELISA (BioRad, New Zealand)’’ [15]. All test results were independently interpreted and recorded by two trained technicians. All HIV and syphilis testing was conducted in the HIV and STI laboratory at PNGIMR. Statistical Analysis We presented descriptive data as frequencies and percentages. The Chi square test was used to formally compare the characteristics of women who consented to get tested for HIV with those who did not. In a sub-analysis, univariable and multivariable logistic regression models were used to determine the independent predictors for HIV seropositivity among women who consented to get tested; odds ratios (ORs) and 95 % confidence intervals (CIs) were calculated. All marginally significant variables with p value \ 0.10 in univariable analysis were included in multivariable analysis; final multivariable logistic regression model was built using a forward stepwise approach and included only factors. The fitness of the model was tested using Hosmer–Lemeshow criteria. Non-significant p-value (i.e. low Chi square test statistics) was considered as an indicator for an acceptable model. In an additional analysis, we also investigated the combined impact of the factors associated with HIV diagnosis and past STI diagnosis as well as current signs/symptoms of STI(s). For this analysis, we used the risk factors significantly associated with HIV diagnosis. For each of these factors, we assigned a score 0 (no risk) and 1 (high risk). We then added the scores for each factors to identify those at the lowest (minimum) to highest (maximum) score. Population Attributable Risk In an attempt to better understand the individual and combined impact of the risk factors on HIV testing and diagnosis, we calculated population attributable risk (PAR). Briefly, we calculated point and 95 % interval estimates of the proportion of HIV infections that could have been prevented if all women had been in the lowest risk category for the modifiable risk factors (at least theoretically). We assessed their relative contributions to

123

AIDS Behav (2015) 19:2194–2203

HIV seropositivity when they were considered separately and combined [16, 17]. Unadjusted and adjusted estimates of the PARs and their 95 % CIs for the factors associated with HIV diagnosis were calculated using unadjusted and adjusted odds ratios from univariable and multivariable logistic regression models, respectively. Prevalences for the combinations of the various risk factors were estimated using multinomial probabilities at each unique level combination of the factors. The PAR is formulated as a function of odds ratio (OR) (s) and the prevalence (p) (s) of the risk factor(s). When there is only one risk factor at two levels (1 vs. 0) PAR ¼

pðOR  1Þ 1 ¼ 1  P2 pðOR  1Þ þ 1 s¼1 ps ORs

ð1Þ

where OR is the odds ratios, p is the prevalence of the risk factor in the population and sindexes the two strata determined by the value of the risk factor. Equation 1 can be generalized to the multi-factorial setting when there are more than one risk factors at multiple levels, as PS ps ðORs  1Þ 1 PAR ¼ ¼ 1  P2 Ps¼1 S 1 þ s¼1 ps ðORs  1Þ þ 1 s¼1 ps ORs ð2Þ where ORs and ps, s = 1, …, S, are the odds ratios and the prevalences in the target population for the sth combination of the risk factors. Full PAR can be estimated by using Eq. 2 and interpreted as the percent reduction expected in the number of HIV diagnosis if all the known risk factors were eliminated from the target population. In a multifactorial disease setting, at least some key risk factors such as age and sex cannot be modifiable. This limits the practical utility of the full PAR which is based on modification of all variables of interests. In an evaluation of a preventive intervention in a multifactorial disease setting, the interest is in the percent of cases associated with the exposures to be modified, when other risk factors, particularly non-modifiable, exist but do not change as a result of the intervention. Therefore we derived and used partial PAR which kept unmodifiable variable(s) unchanged. Under the assumption of no interaction between the modifiable and non-modifiable risk factors of interest, the partial PAR is formulated as P PS PT P pst OR1s OR2t  Ss¼1 Tt¼1 pst OR2t PAR ¼ s¼1 t¼1PS PT p OR OR PT s¼1 t1 st 1s 2t p OR t 2t ¼ 1  PS Pt¼1 T s¼1 t¼1 pst OR1s OR2t ð3Þ where t denotes a stratum of unique combinations of levels of all background risk factors which are not modifiable

AIDS Behav (2015) 19:2194–2203

and/or not under study, t = 1,…, T and OR2t is the odds ratio in combination t relative to the lowest risk level, where OR2,1 = 1. As previously, s indicates a risk factor defined by each of the unique combinations of the levels of the modifiable risk factors, that is, those risk factors to which the PAR applies, s = 1, …, S, and OR1s is the relative risk corresponding to combinations relative to the lowest risk combination, OR1,1 = 1. The joint prevalence of exposure group s and stratum t is denoted by pst, and p.t = RSs=1pst. The PAR represents the difference between the number of cases expected in the original cohort and the number of cases expected if all subsets of the cohort who were originally exposed to the modifiable risk factor(s) had eliminated their exposure(s) so that their relative risk compared to the unexposed was 1, divided by the number of cases expected in the original cohort. The principle of the approach for PAR was to determine the joint impact of several theoretically modifiable risk factors on HIV diagnosis while keeping non-modifiable and/or background risk factors unchanged. All analyses were conducted using SAS statistical software, version 9 (SAS Inc., Cary, NC, USA) and Stata 12.0 (College Station, TX, USA).

Results Population Characteristics Over fifty percent (60.04 %) of the women were younger than 25 years of age, 30 % of women reported that they had no formal education; majority of women (78.97 %) were either married and/or cohabiting with a sexual partner; more than 60 % of women had their sexual debut when they were younger than 16 years of age and 14 % of them reported that they got raped in past 12 months. Almost 70 % of the women reported at least 5 or more clients (paying/non-paying), one quarter of them were classified as frequent alcohol consumer. In terms of sexual practices in past 12 months, just over fifty percent of them reported to have had anal sex and/or oral sex. Although vast majority of them (90 %) had heard about HIV/AIDS and more than 82 % knew the most common transmission route of HIV, only a third reported using condom consistently in past 4 weeks. In addition to this, less than 40 % of them volunteered to get tested for HIV infection. Characteristics of Women by Their HIV Testing Status A total 203 (38.81 %) of the women consented to be tested for HIV infection. Table 1 compares the characteristics of the women who tested for HIV with those who did not.

2197

Women who had some level of education were more likely to get tested for HIV infection compared to those who had no formal education (75.86 vs. 65.63 %, p = 0.013). Women who reported using condom consistently in past 4 weeks (33.50 vs. 25.00 %, p = 0.036), those who had heard of HIV/AIDS (95.57 vs. 86.56 %, p = 0.001) and knew transmission route of HIV infection (89.16 vs. 79.06 %, p = 0.003) were all more likely to get tested for HIV. Women who were diagnosed and/or treated with a sexually transmitted infection previously were less likely to get tested compared to those who have never been diagnosed (23.15 vs. 40.63 %, p \ 0.001); while those who diagnosed with syphilis were more likely to get tested for HIV compared to those who did not (6.40 vs. 3.13 %, p = 0.075). Factors Associated with HIV Diagnosis and Their Population Level Impacts In a sub-analysis, we determined the factors associated with HIV diagnosis (Table 2). Among 203 women who tested for HIV infection, 14 (7 %) were determined to be HIV seropositive. In univariable analysis, women who had no formal education were approximately 10 times more likely to be diagnosed with HIV infection compared to those who had some level of education (OR 9.62, 95 % CI 2.86, 33.30, p \ 0.001); not being married and/or not cohabiting with a sexual partner (OR 3.55, 95 % 1.15, 10.90), frequent alcohol use (OR 3.73, 95 % CI 1.23,11.23, p = 0.020), being raped (past 12 months) (OR 3.48, 95 % CI 1.08, 11.21, p = 0.036) were associated with HIV seropositivity. Inconsistent condom use (sometime/never vs. always) was also determined to be significantly associated with HIV infection (OR 2.54, 95 % CI 1.25, 5.19, p = 0.010). Women who did not know transmission route for HIV infection were approximately 6 times more likely to be tested positive compared to those who did know; those diagnosed and/or treated with other sexually transmitted infections previously (OR 3.73, 95 % CI 1.23, 11.24, p = 0.020) and current diagnosis with syphilis were all significantly associated with HIV positivity (OR 4.88, 95 % CI 1.17, 20.33, p = 0.029). In multivariable analysis, same factors remained significant in the model; however, ORs were slightly attenuated (Table 2). Table 3 presents unadjusted and adjusted estimates of the population attributable risks and their 95 % CIs for the factors associated with HIV diagnosis in univariable and multivariable analysis, respectively. Overall, most influential factor was determined to be lack of education which was accounted for 61 % (95 % CI 44 %, 75 %) of the all HIV infections; while inconsistent condom use in past 4 weeks accounted for more than 50 % of the HIV cases. Third most influential factor was determined to be frequent

123

2198

AIDS Behav (2015) 19:2194–2203

Table 1 Demographics of sex workers who contented to get tested for HIV among the sex workers in highlands of PNG Tested for HIV infection

All participants

Overall N (%)

No N (%)

Yes N (%)

523 (100)

320 (61.19)

203 (38.81)

Age (years) 314 (60.04)

188 (58.75)

126 (62.07)

25?

209 (39.96)

132 (41.25)

77 (37.93)

364 (70.00)

210 (65.63)

154 (75.86)

159 (30.00)

110 (34.38)

49 (24.14)

Level of education

0.013 (6.15)

Some level of education No formal education Age at sexual debut

0.588 (0.29)

16?

196 (37.48)

117 (36.56)

79 (38.92)

\16

327 (62.52)

203 (63.44)

124 (61.08)

Married/cohabiting with a sexual partner/husband

413 (78.97)

249 (77.81)

164 (80.79)

Not married/not cohabiting

110 (21.03)

71 (22.19)

39 (19.21)

No

449 (85.85)

277 (86.56)

172 (84.73)

Yes

74 (14.15)

43 (13.44)

31 (15.27)

160 (30.59)

91 (28.44)

56 (33.99)

363 (69.41)

229 (71.56)

134 (66.01)

396 (75.72)

240 (75.00)

156 (76.85)

127 (24.28)

80 (25.00)

47 (23.15)

Cohabitation status

0.416 (0.66)

Rape in past 12 months

0.558 (0.34)

Total number of clientsa Less than 5 5 or more

0.179 (1.90)

b

None/rare Frequent (daily/[4 per week) Sexual practices in past 4 weeks

0.631 (0.23)

Anal/oral sex

0.127 (2.32)

No

246 (47.04)

87 (42.86)

159 (49.69)

Yes

277 (52.96)

116 (57.14)

161 (50.31)

Always

148 (28.30)

80 (25.00)

68 (33.50)

Sometime/never

375 (71.70)

240 (75.00)

135 (66.50)

No

52 (9.94)

43 (13.44)

9 (4.43)

Yes

471 (90.06)

277 (86.56)

194 (95.57)

434 (82.98)

253 (79.06)

181 (89.16)

89 (17.02)

67 (20.94)

22 (10.84)

346 (66.16)

190 (59.38)

156 (76.85)

177 (33.84)

130 (40.63)

47 (23.15)

No

176 (33.65)

101 (31.56)

75 (36.95)

Yes

347 (66.35)

219 (68.44)

128 (63.05)

No

500 (95.60)

310 (36.88)

190 (93.60)

Yes

23 (4.40)

10 (3.13)

13 (6.40)

Condom used

0.036 (4.42)

Ever heard of HIV/AIDS

0.001 (11.25)

Know transmission route of HIV Yes—know No—don’t know

0.003 (8.97)

\0.001 (16.93)

Ever diagnosed/treated with STIs No Yes Current STI signs/symptoms

0.204 (0.16)

Diagnosed with syphilis

a

Past 7 days

b

Past 4 weeks

123

– 0.450 (0.57)

\25

Alcohol consumption

p-value (test statistics)

0.075 (3.18)

AIDS Behav (2015) 19:2194–2203

2199

Table 2 Correlates of HIV diagnosis among women who consented to get tested for HIV Univariable analysis Odds ratio (95 % CI)

Multivariable analysis p-value (test statistics)

Odds ratio (95 % CI)

p-value (test statistics)

Age (years) (median) \25

1

25?

1.74 (0.58, 5.21)

0.322 (0.99)



Level of education Some level of education

1

No formal education

9.62 (2.86, 32.30)

Age at sexual debut 16?

1

\16

1.64 (0.50, 5.43)

1 \0.001 (3.66)

0.415 (0.82)

9.70 (2.87, 32.83)

\0.001 (3.65)



Cohabitation status Married/cohabiting with a partner/husband

1

Not married/not cohabiting

3.55 (1.15, 10.90)

1 0.027 (2.21)

3.82 (1.02, 14.36)

0.048 (2.18)

Rape in past 12 months No

1

Yes

3.48 (1.08, 11.21)

1 0.036 (2.10)

3.42 (1.06, 11.06)

0.126 (1.53)



0.040 (2.21)

Total number of clientsa Less than 5

1

5 or more

3.29 (0.72, 15.16)

Alcohol consumption None/rare

1

Frequent (daily/[4 per week)

3.73 (1.23, 11.23)

1 0.020 (2.35)

3.81 (1.26, 11.58)

0.577 (0.56)



0.018 (2.33)

Sexual behaviours in past 4 weeks Anal/oral sex No

1

Yes

1.37 (0.45, 4.27)

Condom used Always

1

Sometime/never

2.54 (1.25, 5.19)

1 0.010 (2.65)

2.42 (1.17, 5.00)

0.614 (0.50)



0.017 (2.31)

Knowledge of HIV/AIDS Ever heard of HIV/AIDS Yes

1

No

1.74 (0.20, 15.00)

Know transmission route of HIV Yes—know

1

No—don’t know

5.62 (1.69, 18.69)

1 0.005 (2.82)

4.65 (1.35, 16.05)

0.015 (2.29)

Ever been diagnosed/treated for an STI(s)b No Yes Current STI signs/symptoms

1 3.73 (1.23, 11.24)

No

1

Yes

1.06 (0.34, 3.29)

1 0.020 (2.33)

3.19 (1.03, 10.0)

0.921 (0.10)



0.044 (2.10)

Current diagnoses with syphilis No

1

Yes

4.88 (1.17, 20.33)

a

Past 7 days

b

Chlamydia, gonorrhoea or syphilis

1 0.029 (2.26)

4.69 (1.20, 20.0)

0.035 (2.18)

123

2200

AIDS Behav (2015) 19:2194–2203

Table 3 Predictors of HIV seroprevalence and their population level impacts among those who consented to test for HIV Modifiable factors

Unadjusted PAR (95 % CI)

Adjusted PAR (95 % CI)a

Full PARb



97 % (95 %, 99 %)

61 % (44 %, 75 %)

61 % (43 %, 77 %)

Individual impact No formal education Inconsistent condom use

52 % (43 %, 64 %)

54 % (45 %, 64 %)

Frequent alcohol consumption

35 % (21 %, 53 %)

39 % (22 %, 57 %)

Not married/Not cohabiting

29 % (16 %, 46 %)

20 % (8 %, 40 %)

Raped in past 12 months

27 % (15 %, 42 %)

26 % (13 %, 45 %)

Not knowing transmission route of HIV

25 % (13 %, 46 %)

23 % (12 %, 39 %)

Ever been diagnosed/treated for an STI(s)

24 % (12 %, 46 %)

22 % (10 %, 43 %)

Current diagnosis with syphilis

20 % (12 %, 32 %)

25 % (14 %, 41 %)

68 % (53 %, 80 %)

65 % (50 %, 77 %)

87 % (74 %, 93 %)

85 % (70 %, 90 %)

45 % (27 %, 63 %)

40 % (22 %, 59 %)

89 % (79 %, 94 %)

87 % (78 %, 91 %)

Combined impact of education/knowledge No formal education ? not knowing transmission route of HIV Combined impact of high risk factors Inconsistent condom use ? frequent alcohol consumption ? raped in past 12 months Combined impact of biological factors Ever been diagnosed/treated for an STI(s) ? current diagnosis with syphilis Combined impact of 3 most influential factorsa No formal education ? inconsistent condom use ? frequent alcohol use a

Calculated using the odds ratios from the multivariable logistic regression model presented in Table 2, assumed that only relevant variable(s) of interests was (were) modifiable while others are not (i.e. remain unchanged)

b

Calculated using the odds ratios from the multivariable logistic regression model presented in Table 2, assumed that all the variables in the model are modifiable

alcohol consumption associated with approximately onethird of the cases. High prevalences and high ORs of these factors were responsible for this impact among women. Other factors including not being married and/or not cohabitating with a sexual partner, being raped, knowledge about HIV (heard of HIV and knew transmission route) and past STI(s) and current syphilis diagnoses were accounted for 20–29 % of the all infections. When they were considered combined, lack of education and knowledge of HIV transmission route were accounted for 68 % (95 % CI 53 %, 80 %) of the cases; while combined impact of the high risk factors—namely, inconsistent condom use, frequent alcohol consumption and being raped in past 12 months collectively were responsible 87 % (95 % CI 74 %, 93 %) of the HIV infections. We also present estimates of the proportion of HIV diagnosis attributable to the three most influential factors, i.e. inconsistent condom use in past 4 weeks, no formal education and frequent condom use. In the study population, these three factors collectively accounted for 89 % (95 % CI 79 %, 94 %). Our estimates indicated that almost all HIV infections (97 %) could be attributed to at least one of the eight potentially modifiable factors presented in Table 3. We also assessed combinations of risk factors in relation to past STI(s) and current STI signs/symptoms. As

123

presented in the Fig. 1a, b, prevalent of past STI diagnosis and current STI signs/symptoms increased with an increasing number of risk factors. For example odds ratios of past STI diagnosis and current STI signs/symptoms were estimated to be 8.0 (95 % CI 3.0, 24.0, p \ 0.001) and 5.0 (95 % CI 2.0, 9.0, p \ 0.001), respectively for having at least four of the six risk factors i.e. lack of education, being single/not cohabiting, reporting frequent alcohol consumption, lack of knowledge for route of HIV transmission, being raped (past 12 months) and not using condom consistently (past 12 months) compared to none or one factor only. These risk factors were also accounted for more than 50 % of the past STI diagnosis and current STI signs/symptoms (both).

Discussion Our findings suggest that female sex workers in Eastern Highlands of Papua New Guinea are at high risk for HIV transmission. Particularly, lack of education and knowledge of the basic HIV/AIDS concepts, frequent alcohol consumption and being raped as well as low level of condom use have significant impact on HIV positivity. Biological factors such as diagnosis and/or treated with

AIDS Behav (2015) 19:2194–2203

2201

Fig. 1 a Odds ratios and population level impacts of combined risk factors on ‘‘Past STI diagnosis’’. b Odds ratios and population level impacts of combined risk factors on ‘‘current STI signs/symptoms’’

other sexually transmitted infections were all significantly associated with high prevalent of HIV diagnosis. Our results indicated that almost 100 % of the HIV infections could be ‘‘theoretically’’ prevented by modifying the risk factors selected in this study. Lack of education and low level of condom use among sex workers have been frequently reported in the previous studies [7, 18]. However, this is one of the first studies to show ‘‘very high’’ impact of lack of education on HIV testing and diagnoses not just at individual level but population level as well. One of the most striking results from our study is the lack of interest to get tested for HIV infection and low level

of condom use. Although vast majority of women (90 %) had heard about HIV/AIDS and more than 80 % knew the transmission route of HIV, less than 40 % agreed to get tested for HIV. Despite intensive condom counselling, many women in Papua New Guinea are unable to negotiate safe sex [7]; less than 30 % of women indicated using condom consistently in past 4 weeks. Approximately 70 % of the women reported to have had 5 or more clients (paying/non-paying) in past 7 days. These figures are alarming and have significant implications for the new infections and trajectory of the epidemic in Papua New Guinea.

123

2202

Studies reported complex interaction between genital infections, vaginal health and HIV acquisition [19–21]. A woman’s risk of HIV acquisition has been reported to be increased with presence of other curable sexually transmitted infections [22]. Our study also confirmed these reports. The biological factors including being diagnosed and/or treated with a sexually transmitted infection(s) as well as current diagnosis with syphilis were all significantly associated with increased prevalent of HIV diagnosis. Women are considered to be marginalised population in Papua New Guinea [23]. Apart from lack of human resources and functioning health care facilities, hidden and neglected issues such as sexual violence against women are very common and play serious challenges in spreading the epidemic. In our study population, women who reported being raped in past 12 months were more than three times likely to be diagnosed with HIV infection compared to those who did not. At population level, being raped was attributed to 30 % of the all HIV diagnoses. In addition to this, factors associated with HIV diagnosis have also been shown to have a significant impact on the past STI diagnosis and current STI signs/symptoms. Elimination of 4 or more risk factors from the target population would (theoretically) prevent more than 50 % of the past STI diagnosis and current STI signs/symptoms. Producing and communicating numbers that quantify the impact of risk factors for a disease at a population level has potentially crucial implications for prevention policy and practice [17]. PARs allow us to estimate the reduction in disease that may occur if particular risk factors are removed in target population. We assessed potentially modifiable risk factors known to be associated with HIV positivity to determine the proportion of HIV positive cases would have been prevented by evidence-based innovative health interventions such as reducing inconsistent condom use, increasing level of education and improving knowledge on HIV/AIDS and reducing high risk behaviours. Potential Limitations of Study The current study has several limitations. Female sex workers were recruited using non-random sampling frame, respondent driven survey, results which may impact the generalizability of the results. However, the demographic profile of the women included in this survey is similar to that of female sex workers population studied previously in Papua New Guinea [7]. We also cannot rule out the possibility that our findings may be due in part to unmeasured characteristics such as social and cultural beliefs of the sex workers and their paying/non-paying clients. Despite these limitations, we present reasonably robust analyses for calculating quantitative epidemiological measures of HIV

123

AIDS Behav (2015) 19:2194–2203

infection burden that provides policy-makers and health service providers with an important means to prioritise health services and prevention strategies. We believe, in the absence of representative nationwide HIV testing surveys, studies such as the current one likely to play crucial a role to provide guidance to the policy makers. Overall, impact of risk factors considered in this study was associated with significant reductions in HIV diagnosis. If these results were applied to the target population, almost all (97 %) of the HIV cases could potentially have been prevented by innovative public health interventions; particularly developing and expanding socially and culturally appropriate strategies aimed at increasing level of education and HIV/AIDS knowledge, reducing high risk behaviors including unprotected sexual intercourse and frequent alcohol consumption.

Conclusion Our study underscored the major challenges in this culturally, linguistically and geographically heterogeneous country. Apart from the lack of education, a major obstacle is lack of interest for HIV testing. In fact, no formal education has substantial impact on both not being tested for HIV testing and diagnosis. Future HIV/AIDS prevention programs should include socially and culturally appropriate effective education programs. We believe that results from this study will provide an effective way of identifying and alerting those who are in need of additional, non-routine HIV screening. Acknowledgments We gratefully acknowledge the contribution of the women who participated in The Papua New Guinea (PNG) and Australia Sexual Health Improvement Project (PASHIP) study. We also acknowledge the contributions of the all PASHIP team from Medical Research Institute, Goroka, Papua New Guinea. The PASHIP was established through funding from Australian international development agency (AusAID). Ethical Review Prior to the initiation of the survey, the protocol and data collection tools were approved by the PNGIMR Institutional Review Board (approval number 07.25), the Medical Research Advisory Council (approval number 07.33) and the National AIDS Council Research Advisory Council (approval number RES07-0013) Additional approvals were given by the government authorities in each project province by the Provincial Administration Offices by way of a Memorandum of Agreement.

References 1. Vallely A, Page A, Dias S, Siba P, Lupiwa T, Law G, Millan J, Wilson D, Murray JM, Toole M, et al. The prevalence of sexually transmitted infections in Papua New Guinea: a systematic review and meta-analysis. PLoS One. 2010;5(12):e15586. doi:10.15510. 11371/journal.pone.0015586.

AIDS Behav (2015) 19:2194–2203 2. National_AIDS_Council: The 2007 Estimation Report on the HIV Epidemic in Papua New Guinea. Port Moresby; 2007. 3. UNAIDS/WHO: Oceania: AIDS epidemic update. Regional summary. Geneva, Switzerland: Joint United Nations Programme on HIV/AIDS (UNAIDS) and World Health Organization (WHO); 2008. 4. Worth H, Henderson K. AIDS is a tear in the social fabric of PNG: HIV and its impact 2005–2025. Health Sociol Rev. 2006;15(13):293–304. 5. AusAID: impacts of HIV/AIDS 2005–2025 in Papua New Guinea, Indonesia and East Timor. Final report of HIV epidemiological modelling and impact study. Canberra, Australia: Australian Agency for International Development, Commonwealth of Australia; 2006. 6. WPRO: Second Generation Surveillance Surveys of HIV, other STIs and Risk Behaviours in 6 Pacific Island Countries (2004–2005). Manila: World Health Organization Western Pacific Regional Office; 2006. 7. Kelly A, Kupul M, Man WYN, Nosi S, Lote N, Rawstorne P, Halim G, Ryan C, Worth H: Askim na save (ask and understand): people who sell and/or exchange sex in Port Moresby. Key quantitative findings. Sydney, Australia: Papua New Guinea Institute of Medical Research and University of New South Wales; 2011. 8. Daoni E: Prevalence of HIV in Papua New Guinea. In: Paper presented at the Australasian Sexual Health/HIV conference (ASHM), Brisbane; 2009. 9. Papua New Guinea, Operational Plan. Report. http://www.pepfar. gov/documents/organization/222178.pdf (2013). Accessed 3 Oct 2014. 10. Wand H, Kumbu J: Missed opportunities to prevent mother-tochild transmission in Papua New Guinea: results from a retrospective cohort of HIV exposed infants (unpublished article, under the review, AIDS Care—Vulnerable Children); 2014. 11. Baral S, et al. Burden of HIV among female sex workers in lowincome and middle-income countries: a systematic review and meta-analysis. Lancet Infect Dis. 2012;12:538–49. 12. UNAIDS/WHO: AIDS epidemic update: December 2009. Geneva: UNAIDS/WHO (2009). Accessed 3 Oct 2014. 13. Bunnell R, Opio A, Musinguzi J, Kirungi W, Ekwaru P, Mishra V, Hladik W, Kafuko J, Madraa E, Mermin J. HIV transmission

2203

14. 15.

16.

17.

18.

19.

20.

21.

22.

23.

risk behavior among HIV-infected adults in Uganda: results of a nationally representative survey. AIDS. 2008;22(5):617–24. doi:10.1097/QAD.0b013e3282f56b53. Behavioural Surveillance Survey. http://www.who.int/hiv/strate gic/en/bss_fhi2000.pdf (2000). Accessed 3 Oct 2014. Ryan CE, Maibani G, Kurumop T, Gare J, John B, Kaldor J, Wand H, Siba PM: Large scale population based survey validates recent HIV and syphilis prevalence estimates in Papua New Guinea (Unpublished abstract). Wand H, Ramjee G. Combined impact of sexual risk behaviors for HIV seroconversion among women in Durban, South Africa: implications for prevention policy and planning. AIDS Behav. 2011;15(2):479–86. doi:10.1007/s10461-010-9845-2. Wand H, Spiegelman D, Law M, Jalaludin B, Kaldor J, Maher L. Estimating population attributable risk for hepatitis C seroconversion in injecting drug users in Australia: implications for prevention policy and planning. Addiction. 2009;104(12):2049–56. doi:10.1111/j.1360-0443.2009.02704.x. Hosain GMM, Chatterjee N. Beliefs, sexual behaviours and preventive practices with respect to HIV/AIDS among commercial sex workers in Daulatdia, Ban Bangladesh. Public Health. 2005;119:371–81. doi:10.1016/j.puhe.2004.08.015. Ba¨rnighausen T, Hosegood V, Timaeus IM, Newell ML. The socioeconomic determinants of HIV incidence: evidence from a longitudinal, population based study in rural South Africa. AIDS. 2007;21:S29–38. Laga M, Manoka A, Kivuvu M, Malele B, Tuliza M, Nzila N. Non-ulcerative sexually transmitted diseases as risk factors for HIV-1 transmission in women: results from a cohort study. AIDS. 1993;7:95–102. Leon N, Naidoo P, Mathews C, Lewin S, Lombard C. The impact of provider-initiated (opt-out) HIV testing and counseling of patients with sexually transmitted infection in Cape Town, South Africa: a controlled trial. Implement Sci. 2010;5:8. Martin H, Richardson B, Nyange P, Lavreys L, Hillier S, Chohan B, et al. Vaginal lactobacilli, microbial flora, and risk of human immunodeficiency virus type 1 and sexually transmitted disease acquisition. J Infect Dis. 1999;180:1863–8. Lewis I, Maruia B, Walker S. Violence against women in Papua New Guinea. J Fam Stud. 2008;14:183–97.

123

Prevalence and Correlates of HIV Infection Among Sex Workers in Papua New Guinea: First Results from the Papua New Guinea and Australia Sexual Health Improvement Project (PASHIP).

The primary objective of this study was to estimate the individual and combined impacts of socio-demographic and sexual behaviours on HIV diagnosis am...
557KB Sizes 3 Downloads 11 Views