Introduction
Methods
Search Strategy
Although grey literature can be useful, its inclusion can introduce difficulties in ensuring that the search is systematic and that the studies included are methodologically sound. We therefore chose to restrict our review to capture the highest quality peer-reviewed evidence available using an easily replicable search strategy.Published, peer-reviewed articles on cross-sectional studies, cohort studies or randomised control trials (RCTs) that reported data on FSW from which it was possible to extract or calculate the proportion practising AI and/or the number of AI and UAI acts over any recall period.
Data Extraction
N = 129 | Sources | |
---|---|---|
(A) Outcomes and key study characteristics | ||
Outcomes reporteda | ||
AI prevalence | 123 | |
UAI prevalence onlyb | 5 | |
AI frequency | 13 | |
AI prevalence recall perioda | ||
Lifetime | 30 | |
12 Months | 6 | |
6 Months | 10 | |
3 Months | 7 | |
2 Months | 1 | [113] |
1 Month | 18 | |
15 days | 1 | [65] |
7 days | 9 | |
1 day | 1 | [67] |
With last client | 1 | [154] |
Current primary partner | 3 | |
Not stated | 52 | |
AI practice reported by partner typea | ||
With any type | 63 | |
Clientsc | 62 | |
One-time or new clients | 3 | |
Regular clients | 3 | |
Primary or non-paying partner[s) | 15 | |
Continenta | ||
Africa | 34 | |
Asia | 53 | |
Europe | 23 | |
South America | 10 | |
North America | 14 | |
Mean agea,d | ||
< 28 years | 71 | |
28+ years | 57 | |
Not stated | 6 | |
Survey yeard | ||
Pre-2003 | 64 | |
2003 onwards | 67 | |
Workplacea | ||
Indoors | 33 | |
Outdoors | 12 | |
Mixed indoors and outdoors | 38 | |
Not stated | 53 | |
Mean number of clients per weeka,d | ||
< 10 | 45 | |
10 + | 46 | |
Not stated | 40 | |
(B) Study quality and potential for bias | ||
Interview methoda | ||
ACASI | 10 | |
SAQ | 5 | |
SAQ or FTFIe | 2 | |
FTFI | 111 | |
Coital diary | 4 | |
Polling box | 1 | [23] |
Study design | ||
Cross-sectional | 116 | |
Cohortf | 11 | |
Randomised-controlled trialf | 4 | |
Sampling method | ||
Convenience | 96 | |
Simple-randomised sampling | 5 | |
Cluster-randomised sampling | 7 | |
Respondent-driven sampling | 19 | |
Time-location sampling | 4 | |
Response rate | ||
< 90% | 9 | |
90%+ | 12 | |
Not stated | 110 | |
Place in paper where AI is first mentioned | ||
Title | 11 | |
Abstract | 32 | |
Text | 88 |
Data Synthesis and Statistical Methods
Prevalence Data
Frequency Data
Dealing with Bias
Results
Search Results
Study and Participant Characteristics
Study Quality and Potential Bias
Meta-analysis of AI Prevalence
Sub-group Analysis of AI Prevalence
Study characteristics | Ever | Past month | ||||||
---|---|---|---|---|---|---|---|---|
N | Pooled estimate (95% CI) | I2 | N | Pooled estimate (95% CI) | I2 | |||
Participant characteristics | ||||||||
Partner type | ||||||||
Any | 25 | 14.8% | (11.0–18.6) | 99 | 15 | 15.1% | (8.8–21.6) | 99 |
Clients | 6 | 19.7% | (11.3–28.0) | 97 | 6 | 24.0% | (13.9–34.1) | 99 |
New clients | 0 | – | – | – | 2 | 20.3% | (8.7–32.0) | 90 |
Regular clients | 0 | – | – | – | 2 | 24.8% | (10.0–39.5) | 94 |
Non-paying partners | 2 | 43.9% | (14.7–73.1) | 97 | 5 | 16.5% | (11.4–21.6) | 83 |
Continent | ||||||||
Africa | 10 | 15.1% | (8.8–21.4) | 98 | 7 | 20.4% | (10.1–31.8) | 98 |
Asia | 13 | 14.5% | (10.2–18.8) | 99 | 12 | 14.0% | (6.3–21.6) | 99 |
Europe | 3 | 8.0% | (1.9–14.0) | 86 | 2 | 21.4% | (12.9–29.8) | 64 |
South America | 3 | 22.2% | (14.3–30.2) | 82 | 0 | – | – | – |
North America | 2 | 29.1% | (1.8–56.3) | 95 | 2 | 18.4% | (10.4–26.4) | 0 |
Mean age | ||||||||
< 28 years | 14 | 11.9% | (7.9–15.9) | 98 | 10 | 15.5% | (5.4–25.6) | 99 |
28+ years | 13 | 20.7% | (14.5–26.9) | 99 | 12 | 18.3% | (13.2–24.0) | 95 |
Not stated | 4 | 10.8% | (4.3–17.3) | 98 | 1 | 11.4% | (7.1–15.7) | – |
Survey year | ||||||||
Pre-2003 | 13 | 12.9% | (5.3–19.2) | 99 | 7 | 10.5% | (1.0–19.9) | 99 |
2003 onwards | 18 | 19.2% | (15.4–24.8) | 98 | 16 | 19.4% | (13.2–26.0) | 98 |
Workplace | ||||||||
Indoors | 7 | 21.4% | (12.2–30.5) | 94 | 5 | 14.4% | (0.0–33.8) | 99 |
Outdoors | 2 | 5.5% | (0.0–11.7) | 86 | 1 | 40.6% | (33.6–47.7) | – |
Mixed | 10 | 8.8% | (4.8–12.8) | 98 | 4 | 13.3% | (11.1–16.1) | 1 |
Not stated | 12 | 20.0% | (15.7–24.3) | 97 | 13 | 16.8% | (11.6–22.0) | 96 |
Number of clients/week | ||||||||
< 8 | 12 | 18.6% | (10.5–26.7) | 99 | 5 | 13.6% | (7.1–20.0) | 97 |
8+ | 9 | 13.5% | (10.6–16.5) | 84 | 10 | 19.6% | (9.3–29.9) | 99 |
Not stated | 10 | 14.3% | (9.8–18.8) | 97 | 8 | 15.2% | (9.0–21.5) | 96 |
Study quality and potential for bias | ||||||||
Interview method | ||||||||
ACASI | 3 | 19.3% | (9.8–28.7) | 95 | 2 | 11.3% | (2.7–16.3) | 98 |
SAQ | 0 | – | – | – | 0 | – | – | – |
FTFI | 28 | 15.4% | (11.6–19.1) | 99 | 15 | 17.0% | (10.3–23.6) | 99 |
SAQ/FTFI | 0 | – | – | – | 0 | – | – | – |
Coital diary | 0 | – | – | – | 5 | 15.4% | (2.9–27.9) | 97 |
Polling box | 0 | – | – | – | 1 | 26.0% | (20.8–31.3) | NA |
Study design | ||||||||
Cross-sectional | 26 | 15.4 | (11.4–19.4) | 99 | 14 | 17.5% | (11.4–23.5) | 99 |
Cohort | 3 | 15.0 | (10.3–19.8) | 57 | 1 | 37.0% | (30.3–43.7) | NA |
RCT | 1 | 31.9 | (23.6–40.3) | NA | 1 | 14.1% | (11.7–16.6) | NA |
Recruitment method | ||||||||
Convenience | 16 | 13.2% | (8.3–18.1) | 98 | 16 | 13.9% | (7.3–21.3) | 99 |
Simple randomised | 2 | 36.4% | (30.2–42.5) | 12 | 0 | – | – | – |
Cluster randomised | 5 | 14.8% | (10.9–18.9) | 96 | 3 | 26.9% | (7.8–46.1) | 99 |
Respondent-driven | 5 | 17.8% | (9.9–25.6) | 96 | 6 | 17.1% | (12.5–21.7) | 90 |
Time-location | 3 | 13.7% | (10.2–17.2) | 90 | 0 | – | – | – |
Response rate | ||||||||
< 90% | 2 | 18.9% | (8.3–29.5) | 99 | 1 | 10.2% | (7.0–14.4) | NA |
90+ | 3 | 12.9% | (4.1–21.8) | 99 | 1 | 13.3% | (10.5–16.2) | NA |
Not stated | 25 | 15.3% | (11.6–19.1) | 99 | 16 | 16.6% | (13.5–25.8) | 99 |
AI first mentioned | ||||||||
Title | 4 | 23.9% | (14.0–33.8) | 97 | 3 | 23.8% | (12.8–34.7) | 95 |
Abstract | 10 | 16.9% | (13.4–20.5) | 95 | 5 | 20.1% | (6.0–34.2) | 99 |
Text | 17 | 13.2% | (8.0–18.3) | 99 | 15 | 14.1% | (8.1–20.2) | 99 |
Participant Characteristics
Study Quality and Bias
Comparative Condom use During AI and VI
Frequency of AI Compared to VI
Country | N | Interview method | Partner type | AI prevalence (recall period) | Number of acts/month | Original recall period | % acts that are: | % acts condom protected during: | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
% | AI | VI | UAI | UVI | AI | UAIc | AI | VI | ||||||
Intercourse acts reported among sub-sample who report practicing AI | ||||||||||||||
Van Damme [10] | Multiplea | 765 | Coital diary | Any | 14 (1 month) | 8.7 | NS | NS | NS | 1 week | NS | NS | NS | NS |
Schwandt [29] | Kenya | 147 | FTFI | Any | 41 (ever) | 3.4 | NS | NS | NS | 1 month | NS | NS | NS | NS |
Markosyan [32] | Armenia | 98 | FTFI | Any | 28 (1 month) | 7.4 | NS | 6.2 | NS | 1 month | NS | NS | 83.8 | NS |
Bradley [30] | India | 2394 | FTFI | Any | 10 (ever) | 8.5 | NS | 2.6 | NS | 1 week | NS | NS | 30.9 | NS |
Hladik [28] | Uganda | 942 | ACASI | Any | 15 (1 month) | 3.0 | NS | NS | NS | 1 month | NS | NS | NS | NS |
Tucker [9] | India | 555 | FTFI | Any | 13 (1 month) | 1.8 | NS | 0.2 | NS | 1 month | NS | NS | 11.1 | NS |
Marek [31] | Hungary | 34 | SAQ | Clients | 50 (service) | 27.8 | NS | NS | NS | 1 day | NS | NS | NS | NS |
Maheu-Giroux [3] | Cote d’Ivoire | 466 | FTFI | Any | 19 (1 month) | NS | NS | NS | NS | 1 week | 17.0 | NS | NS | NS |
Intercourse acts reported among whole sample (i.e. including also those who report no AI) | ||||||||||||||
Van de Perre [34] | Rwanda | 33 | FTFI | Any | NA | 1.1 | 43.9 | NS | NS | past 5–10 sexual encounters | 2.4 | NS | NS | NS |
Van Damme [10]b | South Africa | 187 | Coital diary | Any | 41 (1 month) | 4.0 | NS | 1.0 | NS | 1 month | NS | NS | 25.0 | NS |
Ramjee [26] | South Africa | 52 | Weekly FTFI | Any | NS | 3.0d,e | 12.6 | NS | NS | 1 week | 19.4 | NS | NS | NS |
25 | Daily FTFI | Clients | NS | 3.5d,f | 75.4 | NS | NS | 1 day | 4.4 | NS | NS | NS | ||
25 | Daily FTFI | Primary | NS | 0.9d,g | 14.7 | NS | NS | 1 day | 5.6 | NS | NS | NS | ||
25 | Coital diary | Clients | NS | 16.9d,h | 89.3 | NS | NS | 1 day | 15.9 | NS | NS | NS | ||
25 | Coital diary | Primary | NS | 4.3 | 28.6 | NS | NS | 1 day | 13.1 | NS | NS | NS | ||
Voeten [35] | Kenya | 64 | Coital diary | Any | NS | 1.5 | 37.5 | NS | NS | 2 weeks | 4.0 | NS | NS | NS |
Markosyan [32] | Armenia | 98 | FTFI | Any | 28 (1 month) | 2.0 | NS | 1.7 | NS | 1 month | NS | NS | 85.0 | NS |
Carney [36] | South Africa | 457 | FTFI | Any | NS | 2.6 | 30.0 | 1.0 | 9.6 | 3 months | 8.0 | 3.1 | 38.6 | 32.0 |
Bradley [33] | India | 223 | Telephoneb | Any | 19 (ever) | 2.9 | 47.0 | NS | NS | 1 day | 5.9 | NS | NS | NS |
Maheu-Giroux [3] | Cote d’Ivoire | 466 | FTFI | Any | 19 (1 month) | 4.3di | 138.6 | NS | NS | 1 week | 3.0 | NS | NS | NS |
Discussion
Recommendations for Future Reporting of AI Practice
-
Have you had AI in the past 12 months?
-
How many VI acts have you had in the past week with (a) clients and (b) non-paying partners?
-
Was a condom used throughout your last VI act with (a) a client and (b) a non-paying partner
-
How many AI acts have you had in the past week with (a) clients and (b) non-paying partners?
-
Was a condom used throughout your last AI act with (a) a client and (b) a non-paying partner