Population Characteristics
A total of 340 SW took part in the study, of whom 100 (29.4%) were heterosexual MSW. The median age was 27 years (interquartile range [IQR] 25–30) for MSW and 26 years [IQR], (23–29) for FSW (Table 1). The median duration of sex work was 36 and 30 months for MSW and FSW, respectively. Relative to FSW, MSW were more likely than FSW to have obtained secondary or high education (89% vs 45%; p = 0.001), and less likely to be biological parents (64% vs. 82%; p = 0.03). Most MSW (70%) had never married and only 11% engaged in full time sex work. They preferred female clientele aged 35 years or greater, whom they solicited through pimps, dating sites, social media and recreational venues. By contrast, FSW solicited clients on the street and in lodges, clubs, bars and brothels. Most MSW (61%) serviced 1 to 2 clients per week compared with 92% of FSW who had 5 or more clients per week. Condom use at last sex was less likely among MSW than FSW (16% vs. 66%; p = 0.001) (Table 2). Similarly, consistent condom use was reported by only 10% of MSW compared with 63% of FSW (p = 0.001).
Table 1
Socio-demographic characteristics
Variable | Male (N = 100) N (%) or median (IQR) | Female (N = 240) N (%) or median (IQR) | p-value |
Variable | | | |
Age (years) | 27 (25–30) | 26 (23–29) | 0.22 |
Duration of sex work (months) | 36 (24–60) | 30 (13–48) | 0.52 |
Biological children | 1 (0–2) | 1 (1–2) | 0.01** |
Other dependants | 1 (0–2) | 1 (1–2) | 0.76 |
Education level | | | 0.001** |
None | 0 (0) | 25 (10.4) | |
Primary | 11 (11.0) | 108 (45) | |
Secondary | 54 (54.0) | 100 (41.7) | |
Higher education | 35 (35.0) | 7 (2.9) | |
MarItal status | | | 0.001** |
Married | 7 (7.0) | 12 (5) | |
Separated | 23 (23.0) | 83 (34.6) | |
Widow | 0 (0) | 9 (3.8) | |
Never married | 70 (70.0) | 136 (56.4) | |
Solicitation of clients | | | N/A |
Street | 0 (0) | 154 (31.6) | |
Home | 0 (0) | 9 (1.9) | |
Lodge | 0 (0) | 171 (35.2) | |
Bar/Club | 40 (10.2) | 123 (25.3) | |
Brothel | 0 (0) | 29 (6.0) | |
Dating site | 55 (14.0) | 0 (0) | |
Pimp | 60 (15.3) | 0 (0) | |
Dating site | 48 (12.2) | 0 (0) | |
Social media (Facebook, histogram & whatsapp) | 50 (12.8) | 0 (0) | |
Recreation venues (swimming pool, sausan, beaches & hotels) | 49 (12.5) | 0 (0) | |
Escort service | 30 (7.7) | 0 (0) | |
Client referrals | 40 (10.2) | 0 (0) | |
Work place (Saloon, car washing bay) | 20 (5.1) | 0 (0) | |
Description of sex worker | | | 0.001** |
Full time, no any other source of income | 16 (16.0) | 147 (61.2) | |
Full time, supplement my income | 11 (11.0) | 24 (10.0) | |
Part time, have other sources | 69 (69.0) | 66 (27.5) | |
Part time, am student | 4 (4.0) | 3 (1.3) | |
Average number of clients per week | | | 0.001** |
1–2 | 61 (61.0) | 0 (0) | |
3–4 | 29 (29.0) | 20 (8.3) | |
≥ 5 | 10 (10.0) | 220 (91.7) | |
Estimated age in years for female clients | | | |
20–29 | 12 (6.1) | N/A | |
30–34 | 42 (21.2) | N/A | |
35–39 | 65 (32.8) | N/A | |
≥ 40 | 79 (39.9) | N/A | |
Mobility | | | 0.02** |
Work only in this town | 39 (39.0) | 185 (77.1) | |
Move regularly in many towns in Uganda | 52 (52.0) | 50 (20.8) | |
Even move outside Uganda | 9 (9.0) | 5 (2.1) | |
Table 2
Condom use and STI/HIV testing behaviors
Variable | Male (N = 100) N (%) | Female (N = 240) N (%) | P-value |
Variable | | | |
Condom use at last sexual intercourse | | | 0.001** |
Yes | 34 (34.0) | 201 (83.8) | |
No | 66 (66.0) | 39 (16.2) | |
Condom use practice | | | 0.001** |
All the time will all clients | 10 (10.0) | 152 (63.0) | |
Sometimes | 90 (90.0) | 88 (37.0) | |
Ever had a serological test for Syphilis/STI | | | 0.001** |
Yes | 32 (32.0) | 150 (62.5) | |
No | 68 (68.0) | 90 (37.5) | |
Syphilis/STI serological testing frequency in the last 12 months | | | 0.001** |
00 | 81 (81.0) | 114 (47.5) | |
1 | 10 (10.0) | 49 (20.4) | |
2 | 3 (3.0) | 30 (12.5) | |
3 | 4 (4.0) | 38 (15.8) | |
≥ 4 | 2 (2.0) | 9 (3.8) | |
Ever tested for HIV | | | |
Yes | 80 (80.6) | 230 (96.0) | 0.02** |
No | 20 (20.0) | 10 (4.0) | |
HIV serological testing frequency in the last 12 months | | | 0.001** |
00 | 48 (50.0) | 34 (14.0) | |
1 | 8 (8.3) | 46 (19.2) | |
2 | 9 (9.4) | 52 (21.7) | |
3 | 20 (20.8) | 75 (31.3) | |
≥ 4 | 11 (11.5) | 33 (13.8) | |
History of HIV | | | N/A |
I don’t know | 38 (38.0) | 8 (3.3) | |
No | 59 (59.0) | 232 (96.7) | |
Yes | 3 (3.0) | | |
History of Syphilis | | | N/A |
I don’t know | 54 (54.0) | 12 (5.0) | |
No | 20 (20.0) | 91 (37.9) | |
Yes | 26 (26.0) | 137 (57.1) | |
History of Gonorrhoea | | | |
I don’t know | 4 (4.0) | 13 (5.4) | |
No | 28 (28.0) | 175 (72.9) | |
Yes | 68 (68.0) | 52 (21.7) | |
Table 3
Integrated Change Model Scores for Syphlis and HIV testing
Variable | Male (N = 97) | Female (N = 240) | P-value |
SYP_INT | | | 0.001 |
Score ≥ 9, median | 18 (19) | 152 (63.3) | |
Score < 9, median | 79 (81) | 88 (36.7) | |
HIV_INT | | | 0.001 |
Score ≥ 15, median | 23 (23.7) | 158 (65.8) | |
Score < 15, median | 74 (76.3) | 82 (34.2) | |
SYP_ATT | | | 0.001 |
Score ≥ 8, median | 28 (28.9) | 155 (64.6) | |
Score < 8, median | 69 (71.1) | 85 (35.4) | |
HIV_ATT | | | 0.001 |
Score ≥ 25, median | 11 (11.3) | 180 (75.0) | |
Score < 25, median | 86 (88.7) | 49 (25.0) | |
SYP_MN | | | 0.001 |
Score ≥ 8, median | 32 (33.0) | 162 (67.5) | |
Score < 8, median | 65 (67.0) | 78 (32.5) | |
HIV_MN | | | 0.001 |
Score ≥ 16, median | 18 (18.6) | 167 (69.6) | |
Score < 16, median | 79 (81.4) | 73 (30.4) | |
HIV_SE | | | 0.001 |
Score ≥ 30, median | 37 (38.1) | 143 (59.6) | |
Score < 30, median | 60 (61.9) | 97 (40.4) | |
SYP_SE | | | 0.001 |
Score ≥ 3, median | 39 (40.0) | 202 (84.2) | |
Score < 3, median | 58 (60.0) | 38 (15.8) | |
SYP_INT - intention to seek 3-monthly syphilis serological testin; HIV_INT - intention to seek 6-monthly HIV testing; SYP_ATT - attitude towards 3-monthly syphilis testing; HIV_ATT - attitude towards 6-monthly HIV testing; SYP_MN - perceived prevalence of the practice of 3-monthly testing; HIV_MN - perceived prevalence of the practice of 6-monthly HIV testing; HIV_SE - perceived self-efficacy to seek 6-monthly HIV testing |
Table 4
Negative binomial multivariable model for HIV testing in the prior 12 months
| Crude ratio (PR) | Adjusted ratio (PR) | |
| PR (SE) | 95% CI | p-value | PR (SE) | 95% CI | p-value | |
Age category in years | |
17–19 | Reference | |
20–24 | 1.18 (.3) | 0.75–1.84 | 0.47 | 1.21 (.3) | 0.78–1.89 | 0.39 | |
25–29 | 1.24 (.3) | 0.80–1.92 | 0.34 | 1.19 (.3) | 0.76–1.86 | 0.45 | |
30–34 | 1.28 (.3) | 0.81–2.03 | 0.29 | 1.12 (.3) | 0.69–1.82 | 0.65 | |
35 + | 1.07 (.2) | 0.63–1.79 | 0.80 | 0.93 (.2) | 0.54–1.60 | 0.79 | |
Level of education | |
None | Reference | |
primary | 0.97 (.2) | 0.71–1.32 | 0.83 | 1.04 (.2) | 0.76–1.43 | 0.79 | |
secondary | 0.89 (.1) | 0.65–1.20 | 0.44 | 1.04 (.2) | 0.755–1.44 | 0.81 | |
Higher education | 1.08 (.2) | 0.76–1.53 | 0.65 | 1.66 (.3) | 1.09–2.50 | 0.02* | |
Marital status | |
Married | Reference | |
Separated | 1.24 (.2) | 0.84–1.83 | 0.28 | 1.43 (.3) | 0.95–2.14 | 0.08 | |
Widow | 1.24 (.4) | 0.69–2.23 | 0.42 | 1.27 (.4) | 0.69–2.35 | 0.45 | |
Single | 1.00 (.2) | 0.68–1.49 | 0.96 | 1.22 (.3) | 0.82–1.83 | 0.33 | |
Have a regular boy friend | 1.36 (.3) | 0.92–2.00 | 0.13 | 1.35 (.3) | 0.90–2.03 | 0.14 | |
Town location | |
Kampala | Reference | |
Mbarara | 0.95 (.1) | 0.81–1.11 | 0.53 | 1.02 (.1) | 0.86–1.22 | 0.81 | |
Gender | |
Male | |
Female | 1.56 (.1) | 1.29–1.89 | 0.001* | 0.97 (.2) | 0.70–1.33 | 0.84 | |
Condom use | | | | | | | |
Yes | | | Reference | | | | |
No | 0.66 (.1) | 0.54–0.79 | 0.001* | 0.86 (.1) | 0.69–1.08 | 0.19 | |
HIV_INT | | | | | | | |
Score ≥ 15, median | | | Reference | | | | |
Score < 15, median | 0.46 (.04) | 0.39 - 0.55 | 0.001* | 0.61 (.1) | 0.49 - 0.74 | 0.001* | |
HIV_ATT | | | | | | | |
Score ≥ 25, median | | | Reference | | | | |
Score < 25, median | 0.55 (.05) | 0.46 - 0 .65 | 0.001* | 0.88 (.1) | 0.69–1.12 | 0.30 | |
HIV_MN | | | | | | | |
Score ≥ 16, median | | | Reference | | | | |
Score < 16, median | 0.54 (.04) | 0.45–0.64 | 0.001* | 0.75 (.1) | 0.60–0 .92 | 0.007* | |
HIV_SE | | | | | | | |
Score ≥ 30, median | | | Reference | | | | |
Score < 30, median | 0.44 (.03) | 0.37 - 0 .52 | 0.001* | 0.71 (.1) | 0.56 - 0.89 | 0.005* | |
Likelihood-ratio test of alpha = 0: chibar2 (01) = 0.0e + 00 Prob > = chibar2 = 0.500, Log likelihood = -517.0 |
Table 5
Negative binomial multivariable model for syphilis testing in the prior 12 months
| Crude Ratio (PR) | Adjusted Ratio (PR) |
| PR (SE) | 95% CI | p-value | PR (SE) | 95% CI | p-value |
Age category in years |
17–19 | Reference |
20–24 | 1.59 (.8) | 0.62–4.11 | 0.34 | 1.54 (.6) | 0.73–3.27 | 0.26 |
25–29 | 1.59 (.8) | 0.62–4.05 | 0.33 | 1.38 (.5) | 0.65–2.91 | 0.40 |
30–34 | 1.49 (.7) | 0.56–3.97 | 0.42 | 1.12 (.4) | 0.49–2.52 | 0.78 |
35 + | 1.26 (.7) | 0.42–3.78 | 0.68 | 1.00 (.5) | 0.41–2.46 | 0.98 |
Level of education |
None | Reference |
primary | 0.88 (.3) | 0.47–1.65 | 0.69 | 0.98 (.2) | 0.62–1.56 | 0.95 |
secondary | 0.75 (.2) | 0.40–1.39 | 0.35 | 0.97 (.2) | 0.61–1.55 | 0.91 |
Higher education | 0.71 (0.3) | 0.33–1.49 | 0.35 | 1.39 (.4) | 0.73–2.66 | 0.32 |
Marital status |
Married | Reference |
Separated | 0.84 (.3) | 0.40–1.75 | 0.63 | 1.23 (.3) | 0.71–2.13 | 0.45 |
Widow | 1.17 (.7) | 0.37–3.73 | 0.79 | 1.63 (.7) | 0.71–3.74 | 0.25 |
Single | 0.82 (.3) | 0.39–1.69 | 0.59 | 1.23 (.3) | 0.72–2.11 | 0.45 |
Have a regular boy/girl friend | 1.12 (.4) | 0.53–2.34 | 0.77 | 1.34 (.4) | 0.79–2.27 | 0.28 |
Study site |
Kampala | Reference |
Mbarara | 0.69 (.1) | 0.49–0.97 | 0.03* | 1.06 (.2) | 0.79–1.41 | 0.71 |
Gender | | | | | | |
Male | | | Reference | | | |
Female | 3.0 (.6) | 1.99–4.55 | 0.001* | 1.08 (.3) | 0.64–1.84 | 0.75 |
Condom use | | | | | | |
Yes | | | Reference | | | |
No | 0.53 (.1) | 0.37–0.78 | 0.001 | 0.89 (.2) | 0. 64–1.27 | 0.54 |
SYP_INT | | | | | | |
Score ≥ 9, median | | | Reference | | | |
Score < 9, median | 0.16 (.02) | 0.11–0.22 | 0.001 | 0.32 (.1) | 0.22–0.47 | 0.001* |
SYP_ATT | | | | | | |
Score ≥ 8, median | | | Reference | | | |
Score < 8, median | 0.22 (.03) | 0.15–0.31 | 0.001 | 0.43 (.1) | 0.30–0.61 | 0.001* |
SYP_MN | | | | | | |
Score ≥ 8, median | | | Reference | | | |
Score < 8, median | 0.28 (.04) | 0.19–0.39 | 0.001 | 0.63 (.1) | 0.45–0.88 | 0.007* |
SYP_SE | | | | | | |
Score ≥ 3, median | | | Reference | | | |
Score < 3, median | 0.12 (.03) | 0.064–0.21 | 0.001 | 0.39 (.1) | 0.21–0.74 | 0.004* |
Likelihood-ratio test of alpha = 0: chibar2(01) = 0.17 Prob > = chibar2 = 0.340, Log likelihood = -344.47 |
Syphilis and HIV testing
Compared to FSW, self-report of syphilis was less likely among MSW (26% vs. 57%; p = 0.001), as was testing for syphilis in the prior 6 months (6% vs. 19%; p = 0.001). Reasons cited by MSW for not testing - fear (33%), not feeling sick (30%) and not beneficial (20%) – differed from FSW, for whom not being aware (40%), never thought about it (35%) and have no signs of illness (20%) influenced non-testing behavior. Similarly, preferences for syphilis testing venues differed by gender, with 78% of MSW preferring private clinics while FSW tested at public health clinics (49%), private clinics (35%) or during outreach campaigns (10%). HIV testing in the prior 12 months was less likely among MSW (50% vs 86%; p = 0.001), as was ever testing for HIV (80% vs. 96%; p = 0.02). In contrast to syphilis testing, MSW preferences for HIV testing were diverse: public clinics (34%), private clinics (29%), outreach campaigns (23%), and self-testing (8%). FSW preferred to test at public clinics (53%) and private clinics (45%).
Psychosocial Influences of regular Syphilis and HIV testing
We found that relative to FSW, MSW were significantly less likely to test for syphilis in the next 3 months (19% vs. 63%), test for HIV in the next six months (24% vs. 66%), believe in the benefits of regular testing for syphilis (29% vs. 65%) or HIV (11% vs. 75%), have self-efficacy to seek regular syphilis (40% vs. 84%) or HIV testing (38% vs. 60%) or perceive that regular syphilis (33% vs. 68%) or HIV testing (19% vs. 70%) was normative for SW (Pearson Chi-square p < 0.001 for all comparisons).
Associations with Syphilis and HIV Testing
We examined factors associated with syphilis and HIV testing in the prior 12 months. In multivariate analysis, after adjustment for age, level of education, marital status, study site, attitudes to testing and condom use practices, attainment of higher education (adjusted prevalence ratio [aPR] 1.66; 95% confidence interval [CI] 1.09–2.50; p = 0.02), poor intention to seek HIV testing (aPR 1.64; 95% CI: 1.35–2.04; p < 0.001), perception that 6-monthly HIV testing was not common (aPR 1.33; 95% CI: 1.09–1.67; p = 0.007) and poor self-efficacy (aPR 1.41; 95% CI: 1.12–1.79; p = 0.005) were associated with HIV non-testing.
In the multivariable model adjusting for age, level of education, marital status, and study site, low intention to seek syphilis testing (aPR 3.13; 95% CI: 2.13–4.55; p < 0.001), negative testing attitudes (aPR 2.33; 95% CI: 1.64–3.33; p = 0.004), perception that regular testing was not normative (aPR 1.59; 95% CI: 1.14–2.22; p < 0.001), and low self-efficacy (aPR 2.56; 95% CI: 1.35–4.76; p = 0.007) were associated with syphilis non-testing.