Unadjusted and adjusted odds ratios (ORs) for consistent condom use with clients, 2007.

(a)The category ‘rural direct non-venue-based FSWs’ was excluded from this model due to its small sample size (n = 11). A model, which included this sub-group, resulted in large estimates signalling sparse-data bias within this categorical stratum, before and after adjustment (unadjusted OR: 1.9E−9 95% CI 0 - ∞, p = 0.999; adjusted OR: 2.6E−9 95% CI 0 - ∞, p = 0.999). This made it difficult to draw any definitive conclusions. Further, given the detail of the ethnography, there is reason to suspect that the rural non-venue-based FSW sample is not representative in statistical terms.(b)The categories divorced/separated and widowed were aggregated due to their similarity in the direction of effect.(c)1 USD = ca. IDR 9, 560.00 on 6 Sept 2012 according to http://www.xe.com/ucc/convert/?Amount=1&From=USD&To=IDR. Here converted from IDR*1000 to USD (rounded up to IDR 10 000 per 1 USD).