Background Measuring consistent condom use (CCU) is crucial in understanding the trajectory of the HIV epidemic and in evaluating HIV preventive interventions that promote condom use. We examine how one measure of CCU varies across different populations, and compare it to condom use in the last sex act.

Methods Using data from cross-sectional behavioural and biological surveys (2005–2007), we compare responses concerning self-reported level of condom use and condom use in the last act by female sex workers (FSWs), their clients, and men who have sex with men (MSM) in districts from Andhra Pradesh, Maharashtra and Tamil Nadu states in India. For FSWs/clients, the question related to sex with occasional clients/FSWs. For MSM, the question was for non-commercial, non-regular male partners. The time period was unspecified. Levels of condom use were defined as every time", “most of the time/sometimes”, and “never”. We define CCU as those who reply “every time”.

Results 95.2% of all FSWs reported using a condom in the last act, compared to 72.9% of clients and 80.0% MSM. However, only 36.7% of clients and 29.6% of MSM reported CCU compared to 80.0% of FSWs. Abstract P1-S5-16 table 1 shows condom use in the last act stratified by reported level of condom use. For all populations reporting CCU “always”, condom use in the last act was >98%. Of those who reported “most of the time/sometimes”, there was a significant difference between condom use in the last act between clients and the other groups (χ² test, p<0.001), with clients using condoms less frequently. Condom use in the last act among those who “never” use a condom was low (<3%).

Conclusions CCU may not accurately reflect number of acts protected by condoms. Among males in particular, dividing condom users into CCU and non-CCU neglects many who almost always use condoms, and so will still have a high degree of protection. These individuals are an important group to consider when evaluating the effect of changes in condom use on the HIV epidemic, especially among MSM. HIV prevention programmes should try to identify reasons explaining lower CCU among MSM and clients who already use condoms “most of the time/sometimes”, and endeavour to increase consistency within this group.