A cross-sectional survey of 423 students and staff at one Northeastern Thai university classified 337 (79.7%) as inappropriate users because they answered 'sometimes' or 'regularly' to at least one of five risk behaviours. Sharing analgesics was reported by 71.2%, concurrent multi-analgesic use by 23.6%, dose escalation by 16.1%, excessive frequency by 15.4% and use beyond seven days by 11.3%. Each one-point increase in safer-attitude score was associated with lower odds (aOR .919, 95% CI .871–.970); analgesic use in the previous six months with higher odds (aOR 2.889, 95% CI 1.485–5.620). Knowledge was not independently associated. This is not a clinical-harm rate and does not show that changing attitudes causes safer behaviour.
Key findings
- 337 of 423 respondents (79.7%) met at least one risk-behaviour criterion; the operational definition counted 'sometimes' and 'regularly' responses equally for classification.
- Analgesic sharing was reported by 71.2%, concurrent products by 23.6%, excessive dose by 16.1%, excessive frequency by 15.4% and use beyond seven days by 11.3%.
- Mean knowledge was 6.19±2.26/10. NSAID contraindication knowledge was weak: 37.4% correct for ulcer/GERD, 40.2% for pregnancy and 45.2% for kidney disease.
- In the adjusted model, each attitude point was associated with lower odds (aOR .919, 95% CI .871–.970, p=.002), and recent use with higher odds (aOR 2.889, 95% CI 1.485–5.620, p=.002); Nagelkerke R² was .075.
- Knowledge, sex, student/staff status and a healthcare-professional relative/contact were not independently associated. Staff knew more than students (6.73 vs 5.99, p=.005), but classification did not differ.
Why this matters globally
Medication sharing, product duplication and NSAID contraindications are internationally relevant campus-safety targets. The findings suggest that fact-only campaigns may be insufficient and justify testing risk communication, duplicate-ingredient checks, pharmacist consultation and no-sharing norms. Interventions must measure real behaviour and adverse events, not only attitudes or a survey classification.
Thai researcher contribution
Ubon Ratchathani University pharmaceutical-science researchers developed and piloted a locally relevant instrument and reported item-level behaviour rather than only a composite score. The data expose both widespread medicine sharing and NSAID knowledge gaps that can inform campus pharmacy and health-service trials.
Limitations to consider
This was a one-university cross-sectional convenience/self-selected email survey reliant on self-report, with recall and social-desirability bias. Women were 75.9% and students 73.8%, limiting representativeness. The 79.7% definition captures any one behaviour, not frequency, severity, harm or potentially appropriate context. Behaviour α=.70 was modest, model explanatory power low (Nagelkerke R²=.075), and recent users structurally had more opportunity to report behaviour, so this is not a causal factor. No dispensing/purchase records or clinical outcomes corroborated responses.