Abstract
AIMS: This study sought to validate a community-acceptable Substance Use & Sex Index (SUSI) for use in substance use intervention research. SUSI aims to measure behaviours associated with the transmission of blood-borne viruses (BBV) and sexually transmitted infections (STI) among people who use substances and incorporate contemporary sexual and drug practices. DESIGN: Validation of a self-administered online behavioural questionnaire. SETTING AND PARTICIPANTS: An Australian anonymous online questionnaire advertised through health services and social media resulted in 289 respondents with a mean age 35 years (standard deviation [SD] 10.9 years). MEASUREMENTS: A 26-item scale assessing BBV- and STI-associated behaviours based on previous piloting and expert review was assessed for scale structure using exploratory and confirmatory factor analytic approaches. Item Response Theory (IRT) analyses were applied in decisions to retain and categorise items. Item weightings were defined following expert consensus informed by local BBV, STI and HIV epidemiological profiles. Test-retest reliability was examined on a subsample (n = 98) over three to five days. Criterion validity of the new SUSI-2 scale was examined in comparison to the HIV Risk-taking Behaviour subscale of the Opiate Treatment Index (OTI-HRBS). FINDINGS: Factor analysis identified a two-factor model ("sex"; "drugs with sex"), with moderate magnitude correlation (r = 0.38; 95% confidence interval [CI] 0.19-0.54) between factors and acceptable model fit (p = 0.061). IRT discrimination was statistically significant for all items (p < 0.05). Kappa values for test-retest reliability (n = 98 subsample) ranged from 0.66 to 1.00 with high agreement (all above 87%). Free text responses indicated the questionnaire items were acceptable to respondents, with minimal suggestion for improvements. There was a statistically significant positive correlation between the SUSI-2 and OTI-HRBS sex subscales (weighted r = 0.63, p < 0.001) and between the SUSI-2 sex with drugs and OTI-HRBS drug subscale (weighted r = 0.21, p < 0.01). Four additional items were retained to reflect local other BBV and STI transmission risk. CONCLUSIONS: The Substance Use & Sex Index 2 (SUSI-2) appears to be a valid and acceptable two-factor brief scale for the measurement of behaviours associated with blood-borne viruses and sexually transmitted infections for use in substance use interventional research.