RESUMO
Purpose: Understanding the treatment-related attributes influencing medication-taking behaviors in people with type 2 diabetes (T2D) is important for delivery of patient-centered care. This review aimed to identify and summarize studies in which people with T2D (PwD) directly indicated the treatment-related attributes associated with medication-taking behaviors or intentions. Materials and Methods: EMBASE and PubMed were searched for studies (Jan 2005-May 2021) reporting the link between PwD-expressed diabetes treatment-related attributes and the decision to initiate, adhere to, or discontinue a T2D medication. Eligible studies reported attributes associated with oral antidiabetes drugs or injectables (not insulin). Studies not explicitly exploring the link between attributes and indicators of behaviors (eg most discrete-choice experiments [DCE] and those interrogating electronic medical records or claims databases) were excluded, as were studies where the link between attribute and behavior came from anyone but the PwD. Results: Of the 6464 studies identified, 16 were included. Studies were conducted across multiple countries; the USA was most represented (n = 8 studies). The impact of treatment attributes was described on indicators of initiation (n = 3), adherence (n = 12), and discontinuation (n = 4). Some studies evaluated multiple behaviors. PwD perspectives were solicited by structured questionnaires (n = 10), qualitative approaches (n = 4), or DCE explicitly exploring the link to medication-taking behaviors (n = 2). Closed- (n = 9) and open-ended questions (n = 7) were employed. Across studies, several factors including glycemic efficacy (n = 9), weight change (n = 9), dosing frequency (n = 9), hypoglycemia (n = 8), gastrointestinal adverse events (n = 8), regimen complexity (n = 6), route of administration (n = 3), and cardiovascular risk (n = 1) were reported as influencing behaviors, being motivators or barriers to initiation, adherence, or discontinuation. Conclusion: Several attributes influence how PwD take their medications. Insights gained directly from PwD have the potential to assist stakeholders in making more informed, patient-centered, treatment decisions, thus choosing and managing medications that PwD are comfortable initiating and persisting with over the longer term.
RESUMO
AIMS: We sought to create a new research and clinical instrument -the Influence and Motivation for Patient ACTivation in Diabetes care (IMPACT-D™) - to measure the degree to which patients with type 2 diabetes (T2DM) value health and believe they can influence it. METHODS: Candidate items were generated via a literature review, expert opinion, and qualitative interviews and focus groups with T2DM patients in Chicago, IL and Chapel Hill, NC. Psychometric testing guided by item response theory was conducted among an online panel of 500 English-speaking adults with T2DM. Differential item functioning analyses evaluated item performance across key participant characteristics. To determine construct validity, IMPACT-D™ scores were compared to other general measures of personality and patient activation. A second study among 300 patients recruited from two internal medicine clinics further assessed associations between the IMPACT-D™ and health outcomes. Cognitive interviews confirmed patient understanding of IMPACT-D™ items and instructions. RESULTS: Exploratory factor analyses revealed a single-factor solution that included 6 items. The IMPACT-D™ demonstrated adequate reliability (α = 0.72) and moderate construct validity with patient activation (r = 0.51, p < 0.001) and personality-conscientiousness subscales (r = 0.29, p < 0.001). Higher scores on the IMPACT-D™ were associated with better physical health (r = 0.17, p = 0.003) and lower severity of depressive (r = -0.35, p < 0.001) and anxiety (r = -0.28, p < 0.001) symptoms. There were no significant differences by blood pressure (r = -0.0021, p = 0.9) or haemoglobin A1c (r = -0.069p = 0.2). CONCLUSIONS: The IMPACT-D holds potential for use in both clinical care and research applications. Future studies should evaluate how to best operationalize its use for both settings.
