A short questionnaire (IRQ) to assess injecting risk behaviour

Publisher Website
Google Scholar
Abstract
A bstract A im . To develop a short injecting risk questionnaire (IRQ) to measure sharing of injecting equipment. D esign. Matrix design with quota assignment, designed to compare the questionnaire when used by interview and self‐completion, in agency and community settings, by agency staff and fieldworkers, with different injectors (age 26 vs. 26 + ; male vs. female, opiate vs stimulant injectors), and in different geographical areas. Settings. Drug treatment and helping agencies, and community settings, in England. Participants. Drug users who had injected in the last 4 weeks. M easurem ents. Questions measured different aspects of equipment sharing. Questionnaire performance assessed by question acceptability, test‐retest (parallel forms) reliability, inter‐rater reliability, inter‐instrument reliability, internal reliability, construct validity and internal collateral validity. Statistical tests included product moment correlation, principal components analysis, and Cronbach's alpha. Findings. The questionnaire was highly acceptable. Test‐retest correlations were all high and significant, questions performed well in all conditions, with no differences by site (agency vs. out‐of‐contact), order (interview or self‐completion first), administration (staff vs. fieldworker), elapsed time or subject characteristics. The questionnaire had high internal reliability (Cronbach's alpha measured a similar domain with all questions loading highly ( 0.32) on a single factor which accounted for 42% of the variance. The complete IRQ elicited higher reports of equipment sharing (77%) than a single question (56%). C onclusions. IRQ performs well in a variety of settings, when administered in different ways to different kinds of IDUs. A single question on 'sharing' elicits fewer positive responses than the use of multiple questions about different sharing practices. + 0.86), and items

This publication has 0 references indexed in Scilit: