Background: We examine the extent to which an existing sex work typology captures human immunodeficiency virus (HIV) risk in Karnataka and propose a systematic approach for devising evidence-based typologies.
Methods: The proposed approach has four stages: (i) identifying main places of solicitation and places of sex; (ii) constructing possible typologies based on either or both of these criteria; (iii) analysing variations in indicators of risk, such as HIV/sexually transmitted infection (STI) prevalence and client volume, across the categories of the typologies; and (iv) identifying the simplest typology that captures the risk variation experienced by female sex workers (FSWs) across different settings. Analysis is based on data from 2312 participants in integrated biological and behavioural assessments of FSWs conducted in Karnataka, India. Logistic regression was used to predict HIV/STI status (high-titre syphilis, gonorrhea or chlamydia) and linear regression to predict client volume.
Results: Our analysis suggests that the most appropriate typology in Karnataka consists of the following categories: brothel to brothel (i.e. solicit and have sex in brothels) (11% of sampled FSWs); home to home (32%), street to home (11%), street to rented room (9%), street to lodge (22%), street to street (9%) and other FSWs (8%). Street to lodge FSWs had high HIV (30%) and STI prevalence (27%), followed by brothel to brothel FSWs (34 and 13%, respectively).
Conclusions: The proposed typology identifies street to lodge FSWs as being at particularly high risk, which was obscured by the existing typology that distinguishes between FSWs based on place of solicitation alone.