Comparison of longitudinal and cross-sectional approaches in studies on knowledge, attitude and practices related to non-medical tranquilizer use

Abstract

Research about the association of knowledge and attitudes with practices (KAP) of non-medical tranquilizer use is scarce. We compared findings from cross-sectional and longitudinal approaches in a KAP-based study on non-medical tranquilizer use in Spain using data collected from the same population. Eight-hundred forty-seven participants completed a validated KAP questionnaire at baseline and were then followed-up bimonthly for one year for episodes of non-medical tranquilizer use. Non-medical use was defined as unprescribed use, non-adherence to treatment, storage/sharing of tranquilizers, or a combination of those practices. Adjusted odds ratios (ORs) and their 95% confidence intervals (CIs) were estimated using logistic regression from cross-sectional data and generalized linear mixed models for repeated measures in the longitudinal approach. Only the longitudinal approach showed that limited knowledge about the effect of tranquilizers on behaviour [OR: 3.24 (95% CI: 1.12–9.38)] and about the negative effect of their excessive consumption [OR: 4.12 (95% CI: 1.5–11.33)] is associated with storing/sharing tranquilizers. Both cross-sectional and longitudinal analyses indicated that personal attitudes towards tranquilizers and attitudes towards healthcare providers are associated with non-medical tranquilizer use, yet with different magnitude of associations. Differences between the two approaches were also observed for individual types of non-medical use. Certain discrepancies exist between findings from longitudinal and cross-sectional approaches on KAP of non-medical tranquilizer use. KAP studies are the backbone for designing and evaluating prevention programs on non-medical tranquilizer use, and hence choosing a proper study design, scrutinizing the associated biases, and carefully interpreting findings from those studies are required.