High Prevalence of Sexually Transmitted Infections Among Female Sex Workers in the Eastern Highlands Province of Papua New Guinea: Correlates and Recommendations

Abstract

Background: 

More than 200 female sex workers (FSWs) participating in commercial sex along the Highlands Highway of Papua New Guinea were identified in a previous survey. This has implications for the spread of sexually transmitted infections (STIs) and human immunodeficiency virus (HIV) to areas and population groups serviced by the road.

Goal: 

The goal of this study was to estimate the prevalence of gonorrhea, chlamydia, syphilis, trichomoniasis, and HIV among FSWs in Goroka and Kainantu in the Eastern Highlands Province (EHP) and to identify correlates that could be considered in intervention and control.

Study: 

Self-identified FSWs recruited through the Goroka Sex Workers Peer-Mediated Programme were invited to participate. All consenting FSWs underwent pretest counseling and provided sociodemographic and behavioral data using a structured questionnaire. The women were also asked to self-collect vaginal specimens and to provide peripheral blood to detect the respective STIs and HIV.

Results: 

Results were available for 211 FSWs. None of the women were positive for HIV. The overall estimated rates for gonorrhea, chlamydia, syphilis, and trichomoniasis were 21%, 19%, 24%, and 51%, respectively. Seventy-four percent were positive for at least 1 STI and 43% had multiple STI infections. High-risk sexual behaviors were found to be common among the women, including low and inconsistent use of condoms, with most of them attributing this to unavailability, dislike by or familiarity with clients, and being drunk and/or high on marijuana.

Conclusions: 

STIs are prevalent among FSWs in Goroka and Kainantu in the EHP and are maintained by widespread high-risk sexual behaviors, including low use of condoms. Implications for their spread through the highway warrants increased efforts in intervention. Apart from a need to promote condom acceptance, distribution, and use, other high-risk sexual behavior and correlates identified in this study provide important considerations for intervention and control in this population.

In Brief

A study among female sex workers working near a major economic route in the Eastern Highlands Province of Papua New Guinea found a high prevalence of sexually transmitted infections and widespread high-risk sexual behavior.

SEXUALLY TRANSMITTED INFECTIONS (STIs) cause considerable morbidity and suffering among women of reproductive age and children in Papua New Guinea (PNG)^1–4 and, as such, constitute a significant challenge to public health. The rates of gonorrhea, syphilis, and chlamydia have remained high in the last 2 decades in PNG. Among selected, presumably low-risk groups, estimated rates in the range of 18% to 80% for gonorrhea, 4% to 30% for syphilis, and 17% to 44% for chlamydia have been reported, and similar rates have also been noted for trichomoniasis and bacterial vaginosis^3,5–12 (Passey, 1994–1998, unpublished data; Passey, 1999, unpublished data). The escalating epidemic of human immunodeficiency virus (HIV) is a particular worry in this context. By June 2004, 8200 HIV antibody-positive cases had been confirmed,¹³ although government estimates place the true figure in the community at anything up to 50,000 infected of the 5.5 million people in PNG. This emergent epidemic is likely to impose difficult challenges, coming on the background of increased movement, poverty, and underdeveloped preventive and therapeutic response.¹⁴ At the present rate of increase, it has been estimated that 10% of the adult population in PNG might be infected with HIV by 2014.¹⁵ How best to address these problems with the limited resources that are available is a key question for PNG.

Among the options that have been studied in other resource-poor settings, targeted interventions in high-risk groups have been found to be a cost-effective strategy.^16–20 As a result of their social status and lack of economic independence in many such settings, the female sex workers (FSWs) are particularly vulnerable and constitute a well-recognized high-risk group for STIs and HIV.²¹ Most studies done in Africa and Asia show that 80% to 90% of STIs originate from FSWs and their clients.^22–25 The high STIs and HIV disease burden observed in these countries occur in the background of sociodemographic and economic profiles that are similar to the PNG setting. Hence, such groups are likely to play a major role also in PNG in spreading and sustaining high rates of these infections.

Determination of the network, the burden, and dynamics of STI transmissions among FSWs and bridging groups like their male clients is crucial in formulating related policies and implementing strategies that are specific, affordable, and effective.²⁶ Unfortunately, unlike many other countries, commercial sex networking in PNG communities is not well defined. It is not commercially organized¹⁵ and therefore the data that is required for target-specific intervention is difficult to obtain.

Nevertheless, a previous baseline survey (Transex Project, 1996, unpublished data) conducted by the Papua New Guinea Institute of Medical Research (PNGIMR) in Goroka, Eastern Highlands Province (EHP), found the commercial sex industry to be thriving along the Highlands Highway. This highway serves as the major economic route, linking coastal ports in Morobe and Madang Provinces with destinations in the PNG highlands and the hinterland and may be the main conduit for the transmission of STIs and HIV between provinces that are serviced by the road. In the baseline survey, more than 200 FSWs were identified through peer-mediated contact. The current study was performed specifically to document epidemiology of these infections in the FSWs and also to identify sociodemographic variables and sexual behavior that could be considered in intervention and control. The study was approved by Medical Research Advisory Committee of PNG and includes findings from January 2001 to August 2002.

Methods

Study Setting and Specimen Collection

Self-identified FSWs from Goroka and Kainantu Districts of the EHP were recruited. No current census data on population is available. However, the projected 2000 population from the last census estimate 440,000 people living in EHP of whom 72,000 and 92,000, respectively, reside in the 2 districts. The majority of people in Goroka and Kainantu are rural villagers, and apart from the 2 urban centers, the periurban vicinity of the 2 towns is increasingly being occupied by settlers who migrate to the province from remote parts of EHP and other highlands and coastal provinces. Both of these districts are serviced by the Highlands Highway, which acts as the major transport and economic route and links the coastal ports in Madang and Morobe Provinces on the east coast with those in the Highlands, including EHP. Most of the people are subsistence farmers or semiskilled and work on coffee plantations or other self-employed trades for cash income, which usually requires regular use of the highway.

The study focused on the urban settings of Goroka and Kainantu towns with their immediate periurban areas, which are composed of the surrounding villages and settlement areas. Several sex workers operating around nightclubs were initially engaged by our FSWs program team, who then contacted their peers and invited them to participate in the study. The FSWs were asked to attend a 1-week workshop held at the PNGIMR for those in Goroka and the Kainantu Women's Center for those in Kainantu. At this workshop, awareness on STIs and HIV/AIDS was conducted. This also included advocating good sexual health and practices through talks and demonstrations using teaching aids. Teaching materials and condoms were also distributed and diagnosis, treatment, and counseling for those infected were offered.

At the end of the workshop, the study objectives and procedures were explained to the women. Those who gave consent to participate in the study underwent pretest counseling by a trained social worker. They were then asked to complete a structured questionnaire used in a previous FSWs study,²⁷ in which data on sociodemographic and sexual history were obtained. This included information on age, marital status, frequency of sexual encounter, and behavior about type of commercial sexual act performed and where it took place. They were also asked to provide information on the type of clients that are involved in the sex trade.

The FSWs were taught how to self-collect vaginal swabs. This entailed inserting a Dacron-tipped swab into the vagina and rotating several times before placing it into a cryotube containing phosphate-buffered saline (PBS), and each of them was asked to provide a vaginal swab specimen for the detection of Neisseria gonorrhoeae, Chlamydia trachomatis, and Trichomonas vaginalis. Ten milliliters of peripheral blood was also obtained from each woman for syphilis and HIV serology. The swabs were immediately taken to the laboratory, vortexed in PBS, and then discarded. Samples from Kainantu, which is approximately 80 km from Goroka, were transported on ice the same day to the PNGIMR laboratory in Goroka for processing. The vaginal and blood samples were stored at −80°C until they were processed.

FSWs at both sites who reported symptoms of having STIs or complained of related complications, including vaginal discharge, abdominal pain, and cervicitis or urethritis, were begun immediately on presumptive treatment for gonorrhea and chlamydia, which included a 2-g stat of amoxicillin, 1-g stat of probenecid, 2 tablets of Augmentin, and 100 mg twice daily of doxycycline for 10 days. Except for suspected ulcerative case, which in this group of women was rarely reported, syphilis and trichomoniasis were not treated until after the laboratory diagnosis. Results were available within a week and the FSWs were treated from day 3 to 2 weeks after procurement of the specimens, depending on accessibility. Unfortunately, 20% of these women did not turn up for follow up, but those who were present and were confirmed positive for syphilis and trichomoniasis were treated intramuscularly with 2.4 million units of single-dose benzathine penicillin and a 2-g stat of tinidazole, respectively.

Serology

A Venereal Disease Research Laboratory test was done on sera obtained from the collected blood samples for the detection of Treponema pallidum according to the standard procedure (Murex Biotech, U.K.). A cutoff point of 1/8 was used as positive for acute infection and all repeat reactive samples were confirmed with T. pallidum hemagglutination assay (TPHA; Murex Biotech). In all cases, a positive and a negative sample were used as controls. Three different methods were used for HIV testing; an enzyme immunoassay-based test: ImmunoComb, HIV 1, and 2 BiSpot kit (PBS-Orgenics, France); and 2 particle-agglutination tests, Serodia HIV-1/2 (Fujerebio, Tokyo, Japan) and Capillus HIV-1/2 (Trinity Biotech, Wicklow, Ireland). All the samples from FSWs were screened twice using Serodia. Repeat reactive sera were then sent to the National Reference Laboratory in Port Moresby (Central Public Health Laboratory) for retesting and confirmation using both ImmunoComb and Capillus.

Polymerase Chain Reaction Detection of Neisseria gonorrhoeae, Chlamydia trachomatis, and Trichomonas vaginalis

Cell suspensions were centrifuged and DNA extracted from sedimented pellets using a cell lysis buffer solution containing 1% Triton X-100, 0.5% of Tween 20, and Nonidet P-40; and 100 μg/mL proteinase K. N. gonorrhoeae and C. trachomatis were detected by multiplex polymerase chain reaction (PCR), whereas detection for T. vaginalis was done in a separate PCR. Both PCRs were done at PNGIMR as previously described.^11,27 Briefly, PCR reactions were performed in 50 μL volume using 5 μL of the crude lysate, 1× PCR buffer (Gibco BRL), 1.5 μmol/L MgCl₂, 0.2 μmol/L dNTP, 10 μmol/L of the respective primer pairs, and 0.025 U of DNA Taq polymerase (Promega). Multiplex PCR was performed by denaturation at 95°C for 1 minute, annealing at 55°C for 1 minute, and extension at 72°C for 1 minute for 35 cycles with a final extension at 72°C for 10 minutes. Samples for the detection of T. vaginalis were initially heated at 94°C for 5 minutes, followed by 35 cycles of denaturation at 94°C for 1 minute, annealing at 47°C for 1 minute, and extension at 67°C for 1 minute with a final extension at 67°C for 10 minutes. Amplified products were resolved by electrophoresis on 2% agarose gels, stained using ethidium bromide, and then visualized. A positive sample and a negative control were always used. Occasionally, because we had primers available, we performed a nested PCR for the detection of C. trachomatis using primers FLS/FLA and N2/N4 as previously described²⁸ to assess consistency.

Data Management and Analysis

Data were double-entered using FoxPro 2.6 (Microsoft Corp.). These data were then converted to Stata 7.0 (Stata Corp.) using Stat/Transfer version 5. Associations between concurrent STIs as well as between STIs and other demographic and behavioral characteristics were analyzed using contingency tables for categorical variables and analysis of variance for numerical variables. Multivariate logistic regression analysis was performed with backward elimination. P values were calculated using the Pearson chi-squared test.

Results

Sociodemography

A total of 211 FSWs between the ages of 16 and 59 (mean age, 27 years) were recruited. Local EHP women comprised 71% of the total and 19% were originally from the neighboring Simbu Province, which shares the western border with EHP. Between 1% and 2% of the women came from other provinces served by the same highway, in particular Western Highlands and Enga further west of EHP and Madang and Morobe on the east coast. Similar number of women also came from other provinces not linked by the highway, which included East Sepik, MilneBay, East New Britain, and the National Capital District.

Most of the women were either living in settlements (45%) or rural hamlets (39%). The remaining 16% were residents of the 2 towns, Goroka and Kainantu. More than half (54%) of the women interviewed at the time of the study were divorced, 8% were separated with their husbands living away from home for studies or work, 5% were widows, 15% were married, and 19% claimed they had never married. Forty-eight percent of the women reported having no children. The number of children ranged between 1 and 5 among FSWs who had children, with more (21%) of them reporting having only 1 child. Almost two third (66%) of the women had a primary school education of whom 23% reported having completed grade 6. Only 6% of the FSWs completed secondary education and had been to college or a tertiary institution. Twenty-eight percent of them had never been to school.

Prevalence of Sexually Transmitted Infections

Prevalence of the respective STI was determined in all participating FSWs, except in 35 women for whom results of syphilis were not available. Eleven of the latter women (11 of 35) were negative for the other STIs, hence of the 200 FSWs that were analyzed, 74% were positive for at least 1 of the 4 pathogens that were detected. As summarized in Table 1, T. vaginalis predominated at 51% (108 of 211). Prevalence rates of N. gonorrhoeae, C. trachomatis, and T. pallidum were 21% (44 of 211), 19% (41 of 211), and 24% (42 of 176), respectively. None of the FSWs was positive for HIV, although samples from 7 women that appeared reactive for HIV-1 during screening were confirmed negative independently using both ImmunoComb and Capillus.

TABLE 1: Prevalence of Sexually Transmitted Infections Among Female Sex Workers in Eastern Highlands Province (EHP)

Higher frequencies of these STIs were detected in Goroka than in Kainantu (Table 1). Although this difference was seen to be statistically significant for T. vaginalis (P <0.01), results were pooled in subsequent analysis as a result of a comparatively few FSWs recruited in Kainantu than Goroka and the fact that sociodemographic and sexual behavior among women in the 2 districts were observably very similar.

Mixed infections with these STIs were observed to be common. Among the 176 FSWs for whom complete results of all STIs were available, 72% (126) were infected, of whom 43% (54 of 126) had more than 1 STI. Double infections accounted for 63% (34 of 54) in this group, 33% (18 of 54) had triple infections, whereas 4% (2 of 54) had all of the 4 detected STIs. Coinfection with N. gonorrhoeae and C. trachomatis was found to be statistically significant (odds ratio [OR] = 11.32; 95% confidence interval [CI] = 4.88–26.60; P <0.001), and both these STIs and T. vaginalis were the most common (13 of 18) concurrent triple infections detected. Although not significant, some association was also shown for double infections involving T. vaginalis with either C. trachomatis (OR = 1.81; 95% CI = 0.86–3.85; P = 0.09) or N. gonorrhoeae (OR = 1.86; 95% CI = 0.90–3.87; P = 0.07). The risk of being infected with 1 STI when a FSW was already exposed to the other 1 or 2 STIs was also determined for all concomitant STI combinations. This is summarized in Table 2. Risk of being infected with C. trachomatis when a FSW was concomitantly exposed to N. gonorrhoeae and T. vaginalis (OR = 8.5; 95% CI = 3.35–21.86; P <0.001) and N. gonorrhoeae and T. pallidum (OR = 8.51; 95% CI = 1.26–70.76; P <0.01), as well as the risk of being infected with N. gonorrhoeae when a FSW had concomitant C. trachomatis and T. vaginalis (OR = 9.09; 95% CI = 3.47–24.19; P <0.001) and C. trachomatis and T. pallidum (OR = 5.14; 95% CI = 0.91–30.73; P = 0.02) infections, were highly significant.

TABLE 2: Associations Among Neisseria gonorrhoeae, Chlamydia trachomatis, Treponema pallidum, and Trichomonas vaginalis

Sexually Transmitted Infections and Sociodemographic Variables

Except for age, education, and number of live children, none of the other sociodemographic variables studied was found to be associated with the detected STIs. Formal education was found to be associated only with C. trachomatis infection in both the univariate and multivariate logistic regression analysis (OR = 2.92; 95% CI = 1.16–7.34; P = 0.02 and OR = 3.56; 95% CI = 1.18–10.70; P <0.02, respectively). Having 1 or more children was predictive for C. trachomatis (OR = 2.14; 95% CI = 1.08–4.27; P = 0.03), N. gonorrhoeae (OR = 1.96; 95% CI = 1.00–3.81; P = 0.05), and T. pallidum (OR = 2.18; 95% CI = 1.05–4.52; P = 0.04). This continued to be a significant risk factor for T. pallidum in the multivariate analysis, (OR = 8.92; 95% CI = 2.44–32.64; P <0.01).

Prevalence of each STI by age is shown in Figure 1. Prevalence rates of C. trachomatis and N. gonorrhoeae were highest in the 16- to 19-year age group. Although a tendency to increase was observed for N. gonorrhoeae among the 25- to 29-year-old age group, there was a general decline with age for both these STIs. T. vaginalis showed 2 peak prevalence rates; among 16- to 19-year and 25- to 29-year age groups, which subsequently declined with age. Rates of T. pallidum peaked among the 25- to 29-year age groups and also declined thereafter. C. trachomatis and N. gonorrhoeae were found to be significantly associated with FSWs with mean ages of 23 (23.3 ± 6.7; P <0.01) and 22 (22.2 ± 5.9; P <0.001), respectively. In the multivariate logistic regression analysis, however, only age between 25 to 29 years was found to be a significant risk factor for T. pallidum infection (OR = 8.23; 95% CI = 1.33–50.80; P = 0.02). Although not significant, being a widow and unmarried or divorced showed some association for T. vaginalis (P = 0.06, 0.07) and T. pallidum (P = 0.06), respectively, in this analysis.

Fig. 1: Age-specific rates of Chlamydia trachomatis, Neisseria gonorrhoeae, Trichomonas vaginalis and Treponema pallidum among female sex workers in Goroka and Kainantu, Papua New Guinea. Pearson's chi-squared P values of mean age groups for the respective sexually transmitted infections were <0.01, <0.001, 0.66, and 0.23, respectively. In multivariate logistic regression analysis, only ages between 25 to 29 years was significantly associated with syphilis (P = 0.02).

Sexual History and Behavior

The time of involvement in commercial sex among the FSWs ranged from 1 month to 10 years with more than 90% having been engaged in the sex industry for 1 to 5 years. FSWs who were engaged in the sex trade for 2 years or more were found to be at high risk for T. vaginalis (OR = 2.25; 95% CI = 1.15–4.39; P = 0.02) and protective for C. trachomatis (OR = 0.46; 95% CI = 0.21–1.00; P = 0.05) in the univariate analysis. These associations were, however, not significant in the multivariate logistic regression analysis. In the multivariate analysis, only 24% (50 of 211) of the FSWs who also had steady noncommercial sex partners were remotely associated with T. vaginalis (P = 0.09). Sixty-four percent (32 of 50) of these women claimed to be married.

The number of clients seen by the FSWs ranged between 1 and 8 per week, with a single FSW having on average 3 clients per week. The amount of cash paid per sexual act ranged from 2 Kina to 30 Kina (US $0.50–8.00). This was not associated with frequency and types of sexual acts performed or whether this occurred in the night or during the day, although charging different rates was found to be protective for T. pallidum (OR = 0.06; 95% CI = 0.01–0.44; P <0.01). The number of clients seen and amount of cash paid were also found not to be associated with age, marital status, or whether the women had been to school.

Apart from vaginal intercourse, 63% and 67% of the FSWs admitted practicing penile-anal and penile-oral sex, respectively. No association of extravaginal sex was established with any STIs. Twenty-one percent (45 of 210) of the women gave histories of being gang-raped, commonly known as “line-up” in PNG. This usually occurred when the women were either drunk or high on drugs. From the responses that were received, 67% of the women were raped while under the influence of alcohol, 2% while under the influence of marijuana, and 27% while under the influence of both alcohol and marijuana. No association with drug use for any STIs was established, and FSWs who reported having been gang-raped were determined to be at increased risk of being infected with C. trachomatis only (OR = 3.59; 95% CI = 1.18–8.68; P <0.01).

Condom use among the FSWs was found to be both low and inconsistent. None of the women reported using a condom for every sexual act they were involved in. Forty-two percent (17 of 40) of FSWs who had steady noncommercial partners admitted never using condoms with their partners, which, although not significant, showed some association with the risk of acquiring T. vaginalis (OR = 2.36; 95% CI = 0.98–5.71; P = 0.06). Seventy-four percent (156 of 210) of them admitted that they did not use condoms in the last sexual act performed. FSWs who had T. pallidum infection were more likely to use condoms (OR = 4.44; 95% CI = 1.89–10.49; P <0.001) in this group. The FSWs gave different reasons why condoms were not used. Of 454 responses that were received, 43% (194) of them cited unavailability of condoms as the main drawback. Thirty-nine percent (179) said their partners, some of whom were frequent clients, did not like it, 16% (73) attributed it to being under the influence of alcohol and marijuana, whereas 2% (8) said they did not like using condoms.

The client profile and place of initial contact and where sex took place are summarized in Table 3. FSWs were required to indicate “never,” “sometimes,” “most of the time,” and/or “all the time” for the type of clients that were seen and places of pick up and where sex took place. A total of 1630 positive responses were received for the type of clients and 1628 and 1149, respectively, for the places where the FSWs were picked up and where sex took place. Those who reported having sex at all with village men, public motor vehicle (PMV) drivers, and businessmen irrespective of the reported frequency comprised 12.5%, 12.2%, and 10.6%, respectively, followed by office workers (9.9%), teachers (9.6%), tradesmen (8.8%), security (8.8%), and policemen (7.7%). Similarly, irrespective of the reported frequency, at least 12.8%, 12.4%, 12%, 11.9%, and 11.6% reported having used markets, street, bars, card games, and PMV stops as more common places of initial meeting and pickup, followed by hotels (9.5%), snooker halls (9.2%), offices (8%), vacant lots/bushes (6.3%), and guest houses (5.1%). The more common places reported by the FSWs where sex took place were the woman's house (16.2%), client's house (15.8%), PMVs (15.1%), hotels (12.2%), bars, (7.3%) and guest houses (7.3%). Dormitories, offices, and vacant lots were used less often for sex (6.9%, 6.6%, and 6.4%, respectively). Hotels, offices, and bars followed by vacant lots and guest houses were the only exceptions where both initial meeting and sex commonly occurred. So-called “Pamuk Haus” (brothels) were also cited, although they were used only occasionally (4.7%). The inclusion of wharfs for initial contact, pickup, and sex indicated that the women were mobile and travel between the Highlands and the coastal provinces.

TABLE 3: Clients of the Female Sex Workers

None of the STIs that were detected was found to be associated with any 1 client group or places of initial meeting and sex in the univariate analysis. In the multivariate logistic regression analysis, having sex with loggers and soldiers were the only consistent client groups determined as significant risk factors for acquiring N. gonorrhoeae (OR = 3.0; 95% CI = 1.28–7.04; P = 0.01) and T. pallidum (OR = 8.93; 95% CI = 2.63–30.28; P <0.001), respectively. Similarly, guest house was the only consistent venue where having sex or initial contact and pickup were associated with a significant risk of acquiring C. trachomatis (OR = 2.97; 95% CI = 1.33–6.64; P <0.01), and N. gonorrhoeae (OR = 2.68; 95% CI = 1.20–6.03; P = 0.02) and T. vaginalis (OR = 2.56; 95% CI = 1.06–6.14; P = 0.04), respectively.

At the time of the study, 96% of the FSWs reported having STIs and/or associated signs and symptoms. Lower abdominal pain and vaginal discharge were reported by 81% and 68%, respectively. Thirty-seven percent of the FSWs also reported experiencing pain during urination and 3% said they had genital sores. Although N. gonorrhoeae and T. pallidum showed some association with genital ulcer (P = 0.09) and lower abdominal pain (P = 0.09), respectively, none of the signs and symptoms reported by the FSWs was found overall to be significant for a single STI.

The women were also asked whether they were infected in the past. Seventy-one percent (149 of 211) answered “yes,” 25% answered “no,” and 5% were either not sure or did not know. Seventy percent of them had ever been treated for a sexually transmitted disease. This ranged between 1 and 9 times. From data that was available for 94 FSWs, 88% (83 of 94) had sought treatment in the last 3 months. Almost 90% of those who reported having sought treatment in the past had presented to either the STD clinic (58%) or urban clinics (30%). Nine percent sought “off-the-shelf” treatment at the chemist and the remaining 3% at other avenues, including self-remedies and traditional healers.

Discussion

It is now well recognized that FSWs and their clients are high-risk groups for acquiring and transmitting STIs. Thus, identifying these groups and targeting them for intervention is important, because this has been found to be a cost-effective strategy for intervention and control of STIs in resource-poor countries.^16–20 Although to a large extent the practice and dynamics of commercial sex trade remains unknown in PNG, the few studies that have been done so far show it to be present and common.^4,29,30 In both Port Moresby and Lae, 2 major cities of PNG, Mgone et al²⁷ recently showed extremely high prevalence rates of STIs among 407 FSWs where N. gonorrhoeae, C. trachomatis, T. pallidum, and T. vaginalis accounted for 36%, 31%, 32%, and 33%, respectively. Ten percent of these women were also determined to be positive for HIV-1.

In this study, we confirm previous anecdotal and preliminary observations that commercial sex is present and thriving in Goroka and Kanaintu in the EHP. We only report here the findings among 211 FSWs who were recruited within the study period. There were, however, others who presented subsequently at peer education workshops for treatment and counseling, and recent observations indicate that there could be many more women currently engaged in commercial sex in the province. Previously, in 1998, Passey et al,^2,3 and subsequently Mgone et al¹¹ in a follow-up analysis, reported prevalence rates of 18%, 27%, 4%, and 47% for gonorrhea, chlamydia, syphilis, and trichomoniasis, respectively, among apparently low-risk rural women in a similar study in this province. Although their estimated rates in the current study (Table 1) are comparatively lower than those reported for FSWs in Port Moresby and Lae, they are still very high and suggest that this group of women is at a higher risk for STIs in comparison to apparently low-risk women in the province.

Despite the fact that we were able to confirm STIs in only 76% of FSWs, our results suggest that these infections could have been even more prevalent. Nearly all the women (96%) in the study reported current infection with the majority of them attributing this to lower abdominal pain and/or vaginal discharge, as well as pain on urination. Given the fact that vaginal swab specimens were self-collected, it is likely that we were not able to detect some infections as a result of inadequate specimens. In addition, a number of these women may have been infected by other STIs, including bacterial vaginosis, herpes simplex, donovanosis, and human papillomavirus, which, although not investigated in this study, have been found to be common also in the EHP in other studies.^2,3,8,31,32 The potentially confounding factor of female genital mutilation is not present in the study area.

In view of the current HIV epidemic reported in PNG, its absence in this high-risk group is rather surprising, especially when it is now known that HIV infection in EHP accounts for 4% of the 8200 recorded cases in the country.¹³ So far, 95% of HIV positivity that have been documented in the EHP has been seen in adults ranging from 15 years to just over 50 years old, with children below the age of 14 years accounting for the rest. The age distribution of FSWs recruited in this study is roughly representative for women in the 15- to 45-year age bracket as estimated by the extrapolated 2000 population data for EHP. Hence, on this basis, the observed result could be genuine for women in this age group. Nevertheless, the proportion that are infected of the total population of 433,000 in EHP is still very small (<1%) and may explain its nondetection in our study population group. Whatever the case may be, its apparent absence in the study group is consistent with, and may therefore be a true reflection of, the trend of distribution as also supported by comparatively lower prevalence rates of STIs and HIV reported in EHP compared with Port Moresby (National Capital District), Lae (Morobe Province), and the Western Highlands Province.¹³ It is, however, important to note that HIV is present in the province and that given the low sensitivity of the methods used to detect HIV in the so-called window period, the observed result should be interpreted with caution.

This observation provides a unique window of opportunity to control and contain the HIV epidemic from spreading in the province. Strategies for control, which at a minimum includes diagnosis and treatment of curable STIs, behavioral change, and consistent use of condoms, have resulted in the decline of STIs and HIV among FSWs in Cotonou (Benin), Kinshasa (Zaire), Nairobi (Kenya), Lima (Peru) and Thailand,^{16,18,33–35} and holds promise for the PNG setting. In addition to the high rates of gonorrhea, chlamydia, syphilis, and trichomoniasis estimated in the study, almost half of the women had multiple infections. Hence, this may warrant presumptive treatments to be incorporated at the sexually transmitted disease clinics and related healthcare centers in PNG for this strategy to be successful both in the province and the country. This is being done for syphilis, gonorrhea, and chlamydia. However, it has been difficult to assess their impact, which has been compromised further by a very low treatment seeking behavior.³⁶ High prevalence of T. vaginalis noted in this study and others before,^{2,3,5,9,10,12,27} and results of our most recent study,³⁷ show that this infection is equally prevalent in men and therefore also warrants presumptive treatment. Although a 2-g stat dose of tinidazole has been suggested for syndromic management among male patients in PNG only recently, presumptive treatment should be encouraged where mixed infections are suspected following the observations made in the study. The high rates of infection and widespread high-risk sexual behavior observed in these women, which includes low and irregular use of condoms and the fact that they were actively engaged in unprotected sex despite reported signs and symptoms show, however, that efforts for control and prevention can prove to be very difficult in this setting.

The results of this study highlight the need for increased and more comprehensive effort in awareness, treatment, and ensuring supply and availability of condoms with equal emphasis on advocating their acceptance and use. The observed protection afforded by condom use by those with syphilis in the study may relate to the fact that this infection is readily manifested clinically compared with others and therefore detectable visually to prompt caution. This is also evident in the fact that FSWs in the study with this infection were likely to charge different rates for sex, perhaps inferring an attempt to secure cheaper sex or resorting to extravaginal sex with the implied negotiation for alternative rates. This may, however, be an isolated event because this represents only those who reported condom use (26%) in their last sexual act performed before our interview. Because most of these infections can either be asymptomatic in the majority of cases or be clinically indistinguishable, it provides a false sense of security that can be easily compromised by the high rate of drug use observed in the study.

We were not able to determine the use of alcohol and marijuana in all the women. However, this appeared to be a widespread practice and it was very common among those women who reported having been gang-raped. It may therefore be associated with a high risk of acquiring C. trachomatis observed in this subgroup of FSWs with implications for other STIs, although this was not shown to be the case in the study. However, with evidence from studies elsewhere and PNG^30–41 suggesting it is responsible for high-risk sexual behavior and exposure, a need to promote abstinence from drug use, complimented by skill training in negotiation and consensual sex, is further highlighted for this high-risk population group in EHP.

A diverse client group (Table 3) and a series of varied locations reported in the study where initial contact and sex occurred (Table 3) make intervention efforts even harder. Although our results may suggest 1 or 2 client groups and some sites of initial contact and sex to be significantly associated with these STIs, we think that this is transient and may only be reflecting a current source of infection and transmission that may change in time. Although this information is crucial to advise intervention of the implicated STIs in the short term, the entire profile of clients and places identified in the study provides vital information for comprehensive intervention in the longer term for all STIs. In particular, given the strategic location of FSWs in the 2 towns on the major Highlands transport route, we expect that truckers will play an effective role in introducing HIV into the province. This is especially so with higher prevalence rates of this infection already reported in 2 of the flanking provinces served by the same highway, Morobe (Lae) and Western Highlands, and the indications in the study that few FSWs in EHP appear to be mobile along this highway also. The results of this study also support making condoms freely available in hotels, guest houses, and bars as some of the most likely places suggested in the study for initial contact and sex (Table 3). Street vending machines might also promote accessibility, given the stigma and perceived restriction associated with obtaining them from public facilities like the hospitals or purchasing them from shops.

The enormity and complexity of the challenges associated with STI/HIV intervention is undeniable in PNG as anywhere else, in particular in the resource-poor countries. Hence, it is crucial to be strategic in addressing some of these challenges with the resources that are available. A successful and cost-effective approach, beginning with treating high-risk groups like FSWs reported in other developing countries,^16–20 prompts unreserved adoption of this strategy in PNG. However, improper and inadequate diagnoses resulting from lack of facilities in many developing countries, including PNG, can prevent this strategy from achieving its full potential. Moreover, although this and similar studies in PNG have identified a few brothels, the network of commercial sex trade is not clearly enough defined in many communities to facilitate targeted intervention. The results of this study set the stage for such an approach to be adopted in the ongoing efforts to successfully implement intervention of STIs and HIV in the EHP of Papua New Guinea.

As a result of confounding by multiple factors in a high-risk group like the FSWs, interactions of STIs with their related variables like sociodemographic and sexual history and behavior can be variable for different settings. Hence, these associations or their absence need to be clearly defined for each setting, with the subsequent development of target-specific algorithms to aid the ongoing efforts. In this study, these interactions were found to be inconsistent and varied in most instances, depending on whether univariate or multivariate analysis was performed. The only consistent association observed that could assist in formulating and implementing intervention are exposure to formal education and having 1 or more children as significant risk for acquiring C. trachomatis and T. pallidum, respectively. Having been gang-raped was also predictive for chlamydia. Some associations were also shown with being a widow and unmarried or divorce as possible independent risks for T. vaginalis (P = 0.06 and 0.07) and T. pallidum (P = 0.06), respectively. FSWs who have sex with noncommercial sex partners were determined to be at a possible risk for T. vaginalis (P = 0.09) also. Whereas sexually active age groups appear to be a significant factor in aiding treatment in nonsex worker population groups, this becomes almost irrelevant, except for syphilis, as shown by the results of multivariate analysis in this study for sex workers. Reported signs and symptoms were found not to be associated with any single STI, although genital ulcers and lower abdominal pain were, respectively and tentatively, associated with N. gonorrhoeae (P = 0.09) and T. pallidum (P = 0.09).

The observations of this study also prompt the approval and administration of broad-based presumptive treatment to manage suspected mixed infections. Our results show that in nearly half of the FSWs who had mixed infections, those infected with chlamydia are more likely to be infected also with gonorrhea and vice versa. We also show that those who are found to be positive for gonorrhea and trichomoniasis or gonorrhea and syphilis are more likely to be infected by chlamydia also. Similar associations have been shown for infection with gonorrhea when a FSW is doubly infected with either chlamydia and trichomoniasis or chlamydia and syphilis. Hence, these observations can be used to advise treatment where technical capacity to diagnose is not available or lacking.

Given the current limitations in logistic support and finances, an initial and cost-effective approach would be to target for awareness, screening, and treatment most likely places (high-effort areas) and subgroups (high-risk groups) where maximum impact can be achieved. Fortunately, for this population group, our results show a predominance of the FSWs in the periurban areas. These are the surrounding villages and settlements where concerted effort in intervention and control could be initiated. More often than not, members of the major male clients identified in the study (Table 3) reside in these settings. This is true for village men, truckers, and security guards, and sometimes policemen and other government and nongovernment workers; hence, this approach has the advantage of targeting both the FSWs and the majority of their clients at the same time.

Despite low treatment-seeking behavior previously reported in the country,³⁸ our results for FSWs in the study show improvement in that over 90% of the women who reported past infections were seeking treatment where it could be appropriately treated (sexually transmitted disease clinics, 58%; urban clinics, 30%; and chemists, 9%). Unfortunately, the widespread clinical manifestations reported by the women indicate that many of them were doing so after the infections had already progressed, thus emphasizing a need for routine screening and early detection. Our observation among 94 women shows this change in behavior in the previous 3 months, before the actual study. This coincides with our preliminary awareness campaigns on STIs and safe sex in the province. Hence, although we see a major challenge in this province, positive indications such as this and associations and priority areas identified in this study are reasons for optimism in the continued effort.

The findings of this study also provide opportunities for applications in regions with similar disease and socioeconomic background. However, apart from being broad and varied, a slightly different client profile shown in the current study to that previously determined in Port Moresby and Lae²⁷ cautions against generalization and emphasizes a need for population-specific study every time. This is especially true for the PNG setting where these differences can be marked.

In all, this study demonstrates that the sex industry is thriving in the Goroka and Kainantu areas of EHP in PNG. The high STI rates and widespread high-risk sexual behavior, including low and irregular use of condoms, reported by the FSWs warrant increased effort to advocate behavioral change and treatment. Related correlates identified in this study can help facilitate this. Although the findings and recommendations of this study have implications for applications in settings with similar sociodemographic and economic profiles, a need for intervention using population-specific data are also emphasized.