Beyond the prescription: unraveling knowledge, attitudes, and the socio-demographic tapestry of antibiotic misuse among university students in Egypt

BACKGROUND: Antimicrobial resistance (AMR) is a global health crisis. University students represent a critical population for AMR prevention, yet traditional knowledge, attitude, and practice (KAP) studies have repeatedly shown that knowledge alone does not predict behavior. AIM: To assess antibiotic-related knowledge, attitudes, and practices among Suez University students and to identify the prevalence and predictors of knowledge-practice discordance. METHODS: A cross-sectional study was conducted from January 2025 to March 2025 at Suez University, Egypt. Using stratified random sampling, 361 undergraduate students completed an Arabic language questionnaire that underwent pilot testing for face validity and demonstrated acceptable internal consistency, assessing antibiotic knowledge (6 items), attitudes (10 items), and self-reported practices (12 items). Knowledge-practice discordance was defined as adequate knowledge (score ≥ 60%) combined with inappropriate practice (score below the median). Latent class analysis (LCA) identified distinct behavioral profiles. Mediation analysis (PROCESS macro) was used to test whether attitudes mediate the knowledge-practice relationship. Multivariable logistic regression identified predictors of adequate knowledge, appropriate practices, and discordance. RESULTS: The mean knowledge score was 62.4%, with 221 (61.2%) of students demonstrating adequate knowledge. However, only 165 (45.7%) reported appropriate practices. A substantial discordant subgroup 87 (24.1%) of the total sample, 87 (39.4%) of knowledgeable students exhibited adequate knowledge but inappropriate practices. LCA revealed four distinct behavioral profiles: Informed Adherents 98 (27.1%), Knowledgeable Non-Compliers 87 (24.1%), Indifferent 112 (31.0%), and System Reliant 64 (17.7%). Mediation analysis, interpreted with caution given the cross-sectional design, showed that attitudes statistically accounted for 71% of the total association between knowledge and practices. The direct association between knowledge and practices was non-significant after accounting for attitudes, suggesting that the relationship operates primarily through attitudes rather than directly. Prior antibiotic use within the past 12 months was a strong independent predictor of discordance (aOR = 2.87, 95% CI:1.52 to 5.42, p = 0.001), with a similar magnitude of effect as social media reliance (aOR = 2.23). Social media as a primary information source increased discordance odds (aOR = 2.23, 95% CI:1.31 to 3.79, p = 0.003), while physician as an information source was protective (aOR = 0.51, 95% CI:0.29 to 0.89, p = 0.02). Female students had higher knowledge scores (66.2% vs. 58.1%, p < 0.001) but were paradoxically more likely to belong to the discordant subgroup (aOR = 1.78, 95% CI:1.04 to 3.05, p = 0.04). CONCLUSION: Over one-third of knowledgeable Egyptian university students do not translate their understanding into appropriate antibiotic practices. This knowledge-practice discordance is strongly associated with attitudes, prior antibiotic experiences, and information sources, while knowledge was not independently associated with practices after adjustment for these factors. Latent class analysis identified a distinct "Knowledgeable Non Complier" profile comprising nearly a quarter of the sample. Interventions targeting factual knowledge alone are unlikely to succeed. Future programs should address attitudinal barriers, counter social media misinformation, and leverage physician-led counseling, with particular attention to female students and those with prior antibiotic use experience.