Related Experiment Video
Updated: Oct 12, 2026
Methods to Quantify Pharmacologically Induced Alterations in Motor Function in Human Incomplete SCI
Published on: April 18, 2011
Medication adherence in epilepsy: Validation of the short medication adherence scale (SMAS-7) and mediation analysis
Mariam Dabbous1,2,3, Fouad Sakr2,3, Pascale Salameh3,4,5,6
1Inserm U1094, IRD UMR270, Univ. Limoges, CHU Limoges, EpiMaCT - Epidemiology of chronic diseases in tropical zone, Institute of Epidemiology and Global Health - Michel Dumas, OmegaHealth, Limoges, France.
Background:
Medication adherence is a key modifiable determinant of outcomes in epilepsy and may represent an important pathway linking psychosocial and treatment-related burdens to quality of life (QOL). However, its role within these pathways remains insufficiently understood, partly due to the lack of brief, multidimensional, validated measures in this population.
Objectives:
This study aimed to validate the Short Medication Adherence Scale (SMAS-7) in adults with epilepsy and to examine whether medication adherence mediates the associations of financial wellbeing, mental health, cognitive difficulties, and antiepileptic drug (AED) adverse effects with QOL.
Methods:
A cross-sectional study was conducted among 649 adults with epilepsy recruited from community pharmacies and a primary healthcare center. Confirmatory factor analysis, measurement invariance, reliability, construct validity, and receiver operating characteristic (ROC) analyses were used to validate the SMAS-7. Mediation analyses were performed using structural equation modeling with 5000 bootstrap resamples, adjusting for age, gender, access to healthcare, seizure characteristics, and seizure control. A Monte Carlo simulation confirmed adequate statistical power to detect indirect effects.
Results:
The SMAS-7 demonstrated excellent structural validity, confirming a three-factor model with excellent fit (χ2/df = 1.437; CFI = 0.999; TLI = 0.998; RMSEA = 0.026; SRMR = 0.015), and showed measurement invariance across gender, seizure type, and seizure control. Internal consistency was high (Cronbach's α = 0.940; McDonald's ω = 0.962). Convergent validity was excellent, and concordance with the parent instrument (LMAS-14) was very high, while divergent and concurrent validity were supported. Medication adherence partially mediated the associations of financial wellbeing (B = 0.028; 95% CI: 0.013-0.046), mental health (B = -0.037; 95% CI: -0.062 to -0.016), and AED adverse effects (B = -0.036; 95% CI: -0.058 to -0.017) with QOL, whereas the indirect effect for cognitive difficulties was not statistically significant. In a mutually adjusted sensitivity model including all four predictors simultaneously, the association between medication adherence and QOL was no longer statistically significant, and all four indirect effects were attenuated to near zero.
Conclusion:
The SMAS-7 is a valid, reliable, and clinically useful tool for assessing medication adherence in adults with epilepsy. When each predictor was considered individually, medication adherence showed partial indirect associations linking financial wellbeing, mental health burden, and AED adverse effects with QOL; however, these associations were attenuated to near zero in a mutually adjusted model including all four predictors simultaneously, indicating that they reflect pathway-specific associations rather than an independent mediating effect of adherence. Medication adherence nonetheless remains a clinically relevant, modifiable target within a multidimensional approach to patient-centered epilepsy care.
