Patient experiences with general practitioners: psychometric performance of the generic PEQ-GP instrument among patients with chronic conditions.

<bold>Background: </bold>Most generic patient experience instruments have not been validated specifically for persons with chronic health problems, even though they are the dominant user of GPs/family physicians.<bold>Objectives: </bold>To assess the psychometric properties of the generic Patient Experiences with GP Questionnaire (PEQ-GP) instrument (five scales: assessment of GP, coordination, patient enablement, accessibility, and practice) in persons with chronic conditions, and to develop a short version to maximize response rates and minimize respondent fatigue in future applications.<bold>Methods: </bold>Secondary analysis of data from a national survey of patient experiences with general practitioners in 2018-2019 (response rate: 42.6%). The psychometric properties of PEQ-GP were assessed with exploratory factor analysis and Cronbach's alpha, supplemented with confirmatory factor analysis (CFA) and item response theory (IRT). A short version was constructed and evaluated based on item performance.<bold>Results: </bold>Nine hundred and seventy persons reported a chronic condition(s), the most frequent being "musculoskeletal, arthritis, other back and joints" (n = 473, 48.8%). Factor analysis identified three scales with adequate psychometric results: GP (15 items; Cronbach's alpha: 0.96), practice (3 items; Cronbach's alpha: 0.87), and accessibility (2 items; Cronbach's alpha: 0.77). Evaluation of item performance identified a 7-item short version, including a 5-item GP scale with scores with strong concordance with the 15-item scale (Intraclass Correlation Coefficient: 0.97, P < 0.001).<bold>Conclusions: </bold>The generic PEQ-GP exhibits adequate psychometric performance for persons with chronic conditions. Three empirically derived PEQ-GP scales cover evaluation of the GP, accessibility, and practice. The 7-item short form minimize respondent burden, but further validation work is warranted before large-scale use. [ABSTRACT FROM AUTHOR]