1 1 2 0 1 9 T H S T R E E T , N W | S U I T E 6 0 0 W A S H I N G T O N , D C 2 0 0 3 6 P S I . O R G

P S I ■ ■

If you aren't counted, you don't count: Estimating the number of female sex

workers in Mandalay and Yangon, Myanmar Si Thu Thein*, Tin Aung*, Hla Myo Kyaw*, Anne Lancelot*, Willi Mcfarland** *Population Services International Myanmar, Strategic Information Department, Yangon, Myanmar

**San Francisco Department of Public Health, San Francisco, United States

BACKGROUND

CONCLUSIONS

Estimates of the size of key affected populations, including female sex workers (FSW), are

necessary for advocacy as well as for planning, implementation, and evaluation of HIV prevention

and care programs. However, these populations are usually hidden and hard to reach. This study

aims to provide population size estimates (PSE) of FSW in the two most populous cities of Myanmar

(Yangon and Mandalay) using multiple methods.

Population size estimation methods, when applied properly, could produce plausible size

estimates of hidden, hard to reach populations. Using multiple methods with data from various

sources could produce more reliable results. These results have immediate applications for

establishing the reach of current programs and the need for scale up of services for FSW in

Myanmar, and the adaptation and application of these methods will be useful for estimating the

size of hidden populations, for which the conventional population data is scarce.

METHOD

The study was conducted in Yangon and Mandalay from October to December 2013. Stratified time-

location cluster sampling was used to recruit 450 FSW in Yangon and 328 in Mandalay. Clusters

were selected using probability proportionate to size in the first stage and fixed numbers of FSW

were selected randomly from each cluster in the second stage. Four PSE methods were used to

calculate estimates for each city: unique object, unique event, service multipliers, and wisdom of the

crowd. Strata-adjusted and unadjusted estimates were calculated for the multiplier methods and

compared respectively. The estimated proportion of FSW among adult female population in each

city was calculated and compared against the published estimates.

Figure 1. Map of Yangon and

Mandalay Cities, Myanmar

Total Population1 – 4.3 million Adult Female Population1 – 1.1 million Female sex workers – unknown

Total Population1 – 2.1 million Adult Female Population1 – 580,000 Female sex workers – unknown

METHODS

Figure 2. Unique objects

(Hair bands) distributed

to FSW one month prior

to the survey

N = n1 * 1/p

• N = population size estimate

• n1 = no. of tokens distributed/no. of clients

• p = proportion of survey respondents who received the token/taken the test/visited DiC (adjusted for variations among strata)

Variance = n1n2(n1-u)(n2-u)/u3

• n2 = total no. of respondents

• u = adjusted count of respondents who received/taken/visited

Multiplier Method2

Wisdom of the Crowd Method

Ask respondents “Their Best Guess”

of the number of FSW in small

geographic area

Compute summary measures for each small geographic

area

Combine the summaries to get a total estimate for

the whole city

RESULTS

Method Number of objects /

attendees / clients (n)

Adjusted % (p)

Population size

estimate 95% CI

Estimated population prevalence

95% CI for estimated

population prevalence

Yangon City (N=450)

Unique object multiplier 490 10.02% 4889 4231-5547 0.44% 0.38-0.50%

Unique event multiplier 113 2.93% 3855 2872-4838 0.35% 0.26-0.44%

Service multipliers

(1) DiC Visit within 6 months 1458 28.62% 5094 4732-5457 0.46% 0.43-0.49%

(2) STI treatment within 12 months 842 12.07% 6975 6117-7834 0.63% 0.55-0.71%

(3) HIV counseling & testing within 12 months 1607 28.88% 5565 5170-5960 0.50% 0.47-0.54%

Wisdom of the crowd N/A N/A 4278 1441-15889 0.39% 0.13-1.43%

Mandalay City (N=328)

Unique object multiplier 480 14.90% 3222 2819-3625 0.55% 0.48-0.62%

Unique event multiplier 107 4.90% 2182 1693-2671 0.37% 0.29-0.45%

Service multipliers

(1) DiC Visit within 6 months 1202 35.27% 3408 3166-3650 0.58% 0.54-0.62%

(2) STI treatment within 12 months 1005 22.86% 4397 3968-4826 0.75% 0.67-0.82%

(3) HIV counseling & testing within 12 months 1274 35.36% 3603 3346-3859 0.61% 0.57-0.66%

Wisdom of the crowd N/A N/A 219 67-574 0.04% 0.01-0.10%

Table 1. Population size estimation of female sex workers in Yangon and Mandalay, Myanmar (Only adjusted estimates were shown.)

Wisdom of the crowd method produced population size estimates consistent with multiplier methods for Yangon but very low estimates for Mandalay.

Estimates from the multiplier methods were consistent among each other, with medians of ~5000

in Yangon (Range: 3500-7000); and ~3300 in Mandalay (Range: 1600-4400). Adjusted proportions

produced slightly higher estimates, by ~200-500. The estimated proportion of FSW among the

adult female population (15-49 years) was 0.45% (0.26-0.71%) in Yangon and 0.56% (0.29-0.82%)

in Mandalay, falling within published estimates for the Asia region (0.2-2%)3.

0

1000

2000

3000

4000

5000

6000

7000

8000

Po

pu

lati

on

siz

e e

stim

ate

s

Figure. 3. Median Population Size Estimates from Multiplier Methods

Unique object multiplier

Unique event multiplier

Service multiplier (1) - Visits

Service multiplier (2) - STI Rx

Service multiplier (3) - HIV test

Mandalay

Yangon

Extrapolated PSE from prior national estimate4 (0.40%)

Literature cited 1. Statistical Year Book 2011, Central Statistical Organization, Myanmar. 2. UNAIDS/WHO Working Group on HIV/AIDS/STI Surveillance. Guidelines on estimating the size of populations most at

risk to HIV. 2010. http://data.unaids.org/ pub/Manual/2010/guidelines_popnestimationsize_en.pdf 3. Vandepitte J, et al. Estimates of the number of female sex workers in different regions of the world. Sex Transm Infect

2006;82(Suppl III):iii18–iii25 4. HIV Estimates and Projections Myanmar 2008-2015, National AIDS Programme, Myanmar.

Acknowledgments We thank all the members of FSW community in Yangon and Mandalay who agreed to participate in the study. Funding for this project was provided by PEPFAR and USAID.

Poster No. WEPDC0105