Female sex worker typology: too complicated to be used pragmatically.

I read with interest the recent article on how to devise a female sex worker typology from Karnataka, India by Raluca Buzdugan et al. The authors demonstrated that combining ‘site of having sex’ along with the ‘site of solicitation’ has significant relations with HIV prevalence and risk behaviour and thus they recommended a modification in devising the existing methodology for greater efficiency of the national HIV programme in India. However, there are a few issues that I would wish to point out. The foremost point of concern is about the criteria for selecting the sample size. The authors have reported using the sample size designed to detect a 10–15% increase in condom use between baseline and endline surveys of Integrated Biological and Behavioural Assessment. The current study appears to be a subgroup analysis attempted out of the data collected in the baseline survey. The authors did not mention whether that sample size had sufficient power to detect the observed changes in the subgroup analysis attempted in their study. The important confounders known to affect HIV prevalence, such as longer duration of sex work, client volume, migration for sex work, etc have been adjusted by authors during analysis by logistic regression. The question remains why this particular subgroup should have higher vulnerability. The authors have given plausible reasons such as compromised level of autonomy and controlled working conditions in lodges and brothels as opposed to independent power while at home or rented rooms, and have also recommended further research to answer the question. My concern however, is with a more basic question linked to definitions for these sundry groups. The authors have missed giving any insinuation of the definition, neither have they given reference where it was published. On searching, I could locate the detailed methods employed for the study, published by Tobi Saidel et al. in a separate article, wherein the operational definition used for sampling of female sex workers (FSW) in Karnataka were as shown in Table 1. Table 1 shows that in Karnataka, of the five districts where FSW were sampled, four districts comprising almost 1600 (67%) of total sample had ‘combined’ FSW. Exclusive groups were present only in Bangalore. Even there, the credibility of the definition of the ‘street’ FSW used in Bangalore is difficult to appreciate, if one tries to decipher what happens if a sex worker is soliciting within 100 m of a lodge. Thus, it reflects that the majority of the samples were actually ‘combined’ by definition and hence, it is very difficult to understand how the authors have carried on with such sophisticated analysis with unclear definitions. It is well recognized that given the mobility in the high risk groups in general and FSW in particular, the size of the risk populations would remain fluid at any given point in time. Moreover, as pointed out by the authors themselves in the article, it is difficult to approach the FSW at the lodges; any attempts to reach them in lodges would yield different results. Let us see how the situation would turn out for programmatic interventions to take shape if we indeed consider the authors’ conclusion. As per the findings, out of the total 2365, maximum sex was happening at home (43%) followed by lodge (24%), brothel (11%) and rented room (11%). Regarding