Étude de la validité de mesures de l'observance du traitement médicamenteux antidiabétique

Authors: Zongo, Arsène
Advisor: Grégoire, Jean-Pierre; Guénette, Line
Abstract: Measurement of adherence remains a challenge in some contexts because the different measurement methods have limitations. The validity of pharmacy-based measures of adherence and self-reported measures of adherence to antidiabetes drug treatment has not been comprehensively studied. The research objectives were thus to assess the validity of 1) four measures of pharmacybased adherence to oral antidiabetes multi-drug treatment and 2) four self-reported measures of adherence. We built a cohort of new users of at least two oral antidiabetes drugs using the Quebec health insurance Board (RAMQ) medico-administrative data. Four pharmacy-based adherence measures were calculated from these data: the proportion of days covered (PDC) by at least one class of drugs, the mean PDC, the PDC by all classes of drugs and the daily polypharmacy possession rate (DPPR). All-cause hospitalization and diabetes-related hospitalization were used as validation criteria. For the second objective, four self-reported measures of adherence were completed during a web survey by a sample of individuals with type 2 diabetes of the Canadian province of Quebec: a 4-item scale (SR-4), a French version of the 8-item Morisky medication adherence scale (MMAS-8), the self-reported proportion of pills missed and a single-item scale (SR-1). Measures of HbA1c taken between 3 and 6 months after the web survey were used as the validation criterion. To assess the validity of the pharmacy-based adherence measures, we conducted ROC curve analyses. In addition, the optimal threshold of adherence to classify individuals as adherent to their treatment was assessed using the method of the "minimum distance". ROC curve analyses, linear regression analyses and factor analyses (MMAS-8 only) were performed to assess the validity of the self-reported measures. Finally, we assessed the internal consistency of the MMAS-8 by calculating Cronbach's alpha. For pharmacy-based adherence to oral antidiabetes multi-drug treatment, the areas under the ROC curves (AUCs), for respectively, the PDC by at least one class of drugs, the mean vi PDC, the PDC by all classes of drugs and the DPPR were 0.54 (95% CI: 0.52-0.56), 0.51 (0.49-0.53), 0.50 (0.48-0.52) and 0.51 (0.49-0.53) with all-cause hospitalization as criterion, and 0.55 (0.53-0.57), 0.53 (0.51-0.55), 0.51 (0.49-0.53) and 0.53 (0.51-0.55) with diabetes-related hospitalization as criterion. The optimal threshold to classify individuals as adherent to their treatment was 95.7% according to the two validation criteria. AUCs for the self-reported measures (low and non-statistically significant) were 0.51 (0.43- 0.59), 0.52 (0.43-0.60), 0.53 (0.45-0.60) and 0.52 (0.44-0.59), for the SR-4, the MMAS-8, the proportion of pills missed and the SR-1, respectively. According to the results of the linear regression analyses in the total sample, all the measures were non-significantly associated with HbA1c, but the association between the SR-4 and HbA1c was close to statistical significance (regression coefficient of -0.25, p-value of 0.07). In the subsample of participants with HbA1c ≥ 7%, the SR-4, the MMAS-8 and the SR-1 were significantly associated with HbA1c. Factor analyses of the MMAS-8 showed that it is made of two subscales: a subscale assessing intentional non-adherence and a subscale assessing nonintentional non-adherence. The internal consistency of the MMAS-8 was low since the Cronbach's alpha of 0.60 was below the suggested threshold value of 0.70. In conclusion, the results suggest that the PDC by at least one class of drugs is the most valid method for calculating pharmacy-based adherence to oral antidiabetes multi-drug treatment. The results also suggest that a minimum level of 95% of PDC should be considered for classifying individuals as adherent to their treatment. For the self-reported measures of adherence, the results support the predictive validity of the SR-1, the SR-4 and the MMAS-8 with a slight superiority of the SR-4. Finally, if the MMAS-8 is used to assess adherence to antidiabetes drug treatment, results suggest that the score of each subscale should be used separately to identify the type of non-adherence in order to provide appropriate intervention to the patient.
Document Type: Thèse de doctorat
Issue Date: 2016
Open Access Date: 8 February 2021
Permalink: http://hdl.handle.net/20.500.11794/68061
Grantor: Université Laval
Collection:Thèses et mémoires

Files in this item:
Description SizeFormat 
32368.pdf2.09 MBAdobe PDFThumbnail
View/Open
All documents in CorpusUL are protected by Copyright Act of Canada.