Twelve months of electronic monitoring (MEMS®) in the Swedish COAST-study: a comparison of methods for the measurement of adherence in schizophrenia.

The primary aim was to compare objective and subjective measures of adherence in a naturalistic cohort of schizophrenia outpatients over 12 months between October 2008 and June 2011. Antipsychotic medication adherence was monitored in 117 outpatients diagnosed with schizophrenia or schizophrenia-like psychosis according to DSM-IV criteria in a naturalistic prospective study. Adherence was determined by the Medication Event Monitoring System (MEMS®), pill count, plasma levels and patient, staff, psychiatrist and close informant ratings. The plasma level adherence measure reflects adherence to medication and to lab visits. Relationships between MEMS® adherence and other measures were expressed as a concordance index and kappa (K). Non-adherence (MEMS® ≤0.80) was observed in 27% of the patients. MEMS® adherence was highly correlated with pill count (concordance= 89% and K=0.72, p<0.001). Concordance and K were lower for all other adherence measures and very low for the relationship between MEMS® adherence and plasma levels (concordance=56% and K=0.05, p=0.217). Adherence measures were also entered into a principal component analysis that yielded three components. MEMS® recordings, pill count and informant ratings had their highest loadings in the first component, plasma levels alone in the second and patient, psychiatrist and staff ratings in the third. The strong agreement between MEMS® and pill count suggests that structured pill count might be a useful tool to follow adherence in clinical practice. The large discrepancy between MEMS® and the adherence measure based on plasma levels needs further study in clinical settings.