Article Text


Epidemiology poster session 4: Methodological aspects
P1-S4.05 Quantifying social desirability biases in reported condom use among female sex workers in Southern India
  1. M Pickles1,
  2. B M Ramesh2,
  3. R Adhikary3,
  4. S Rajaram2,
  5. S Isac2,
  6. R Washington4,
  7. M C Boily1,
  8. J Bradley5,
  9. S Moses6,
  10. M Alary7
  1. 1Imperial College, London, UK
  2. 2Karnataka Health Promotion Trust, Bangalore, India
  3. 3Family Health International, India
  4. 4St John's Research Institute, Bangalore, India
  5. 5CHARME-India Project, Bangalore, India
  6. 6University of Manitoba, Winnipeg, Canada
  7. 7Centre hospitalier affilié universitaire de Québec, Quebec, Canada


Background As part of the Avahan HIV initiative in Southern India, surveys were carried out in female sex workers (FSWs) and their clients to quantify condom use. We examine reported condom use measured using different survey methodologies, and consistency of response between FSWs and clients to quantify the effect of social desirability bias.

Methods We use data from 15 districts with IBBA face-to-face interviews (FTFIs) for FSWs and clients. Three of these districts also had special behavioural survey (SBS) FTFIs, informal confidential voting interviews (ICVIs), and polling booth surveys (PBSs) for FSWs. ICVI/ PBS survey methodologies increase anonymity, reducing reporting bias of sensitive questions eg, condom use, and are analysed in more detail. The IBBA and SBS FTFIs differed as blood samples were taken in the IBBA to measure HIV prevalence. We use questions about condom use in last act with regular and occasional FSWs/clients. For the 15 IBBA districts the FSW: client ratio of reported condom use at last act was calculated.

Results In IBBAs, condom use in last act with occasional FSWs reported by clients in the 15 districts is 26% (range 13−40%) lower than FSW condom use with last occasional client, and 28% (3−46%) lower for last act with regular FSW/client. In the three districts with extra surveys, from Abstarct P1-S4.05 table 1, FSW reported condom use with occasional clients is broadly comparable, and the maximum difference with the more anonymous ICVI is 5% with regular clients in Belgaum. On average the IBBA FTFI was only 2% lower that other methods. Reported condom use with occasional clients was15−26% lower than that reported in the IBBA, but the difference between IBBA and PBS was smaller for use with regular clients.

Abstarct P1-S4.05 Table 1

Reported condom use in last act by type of act for FSW surveys in the three districts where all surveys were carried out

Discussion There is a substantial difference in reported condom use in last act with occasional and regular partner for FSWs and clients. However, the samples of FSW and clients may not be comparable because low-risk clients may be undersampled in hot spots. FSWs report lower condom use with occasional clients in the more anonymous PBS, suggesting this method may reduce social desirability bias more than FTFIs and ICVIs. Based on FSW data only and using a conservative estimate from the largest difference, the FSW IBBA may over-estimate condom use with occasional clients by up to 26% (mean 17%) and with regular clients by 13% (mean 2%).

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.