The Questionnaire Survey Method in Medicine on the Example of Treatment Adherence Scales

Abstract

Aim. Development, testing and validation of the original questionnaire “Adherence Scale” (AS) in the PRIORITY and ANTEY observational studies (OS).Materials and methods. The OS PRIORITY assessed adherence to statins in 298 patients with high and very high cardiovascular risk for 3 months. The OS ANTEY assessed adherence to oral anticoagulants in 201 patients with non-valvular atrial fibrillation for 1 year. Adherence was assessed using the original AS questionnaire, for which external validation was performed (with the calculation of the Cohen's Kappa coefficient). The reference methods were the validated questionnaire and direct medical interview. And internal validation was performed (consistency of questions on the AS using Spearman's correlation analysis). The sensitivity, specificity (ROC analysis) and retest reliability of the adherence scale (Cronbach's alpha) were also determined.Results. In the OS PRIORITY Cohen's kappa for the AS and the reference method of direct medical survey was 0.76 (high consistency), and for the AS and the reference method of the validated questionnaire=0.28 (low consistency). High internal consistency of the questionnaire questions (correlation coefficient=0.78, p<0.0001) confirms the internal validity of the adherence scale. Evaluation of the main characteristics of the modified AS in the OS ANTEY showed high consistency between the results of the AS and the validated questionnaire: Cohen's kappa=0.94 (high external validity of the AS). The retest reliability of the AS was 0.76 (Cronbach's alpha). The internal consistency of the questionnaire was confirmed by a strong and statistically significant correlation between the test questions: Spearman's correlation coefficient=0.80, p<0.0001. The sensitivity of the test, determined using the ROC analysis, was 89%, and the specificity was 62%.Conclusion. The developed and tested new original questionnaire (modified version) – the AS – showed high indicators of reliability, validity and sensitivity. This ensures its reliability and ease of use for assessing various types of adherence and determining the leading factors of non-adherence, and also allows its use in scientific studies and clinical practice.