Introduction: An electronic clinical record (ECR) coding for all variables of CF and capturing “real time” data from 2007 was used as a basis for this study. From this data, longitudinal patterns of clinical variation have emerged that suggest a relationship between variation in lung function and adherence. Our aim was to determine a) the accuracy of self-report adherence and its relationship with clinical variation, and b) whether objective measures, including clinical variation can predict adherence. Methods: Patients aged 16 years and over attending an adult regional CF centre were invited to complete an adherence questionnaire (CFQ-R) and consent to pharmacy script data collection (previous 6 months). Coefficient of variation for FEV1 (CoV FEV1) was calculated from all contacts within the previous year. Age, gender, microbiology, disease severity (banding status), medication, lung function and weight were noted at baseline. Self-reported adherence was calculated against prescribed medication (using ECR) and patients classified into one of 3 categories: low (<60%), moderate (60-80%), good (>80%) according to % calculated adherence. Ordinal regression was used to determine the contribution of age, gender, microbiology status, disease severity (Banding status), medication, genotype and CoV FEV1 to self-reported adherence. Results: Patients (n=250) [age 29.7 (±9.2) y, 58.6% (M), baseline BMI 22.5 kg/m2 (±3.8), FEV1 61.3% (±25.1), FVC 79.4% (±23.1)] completed the study. Pharmacy collection data were available for 106 (42%) patients. Pharmacy script collection was strongly correlated to self-reported adherence (Table). CoV FEV1 was inversely related to self-reported adherence (Table 1). Regression analysis revealed that CoV FEV1 [OR = 0.95; CI: 0.92-0.98, p=0.006], number of types of medication [OR = 1.18; CI: 1.11- 1.26, p<0.001], and age [OR = 1.03; CI: 1.01 to 1.06, p=0.026] together explained 19% of the variance in the model. Conclusion: Self-report adherence consistently exceeds medication collection by an average of 14%. Coefficient variation of FEV1 may be an indirect measure of adherence and contributed to the final adherence model. Accurate longitudinal measures captured in real time can aid in examining adherence and warrant further investigation. Supported by a grant from Gilead Sciences.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)