Psychometric evaluation of the Chinese version of the Patient Self-Advocacy Scale using classical test theory and item response theory

TL;DR

This study validated a Chinese version of the Patient Self-Advocacy Scale for head and neck cancer patients, showing it is reliable and useful for improving patient care.

Contribution

The study provides a validated Chinese version of the PSAS with strong psychometric properties for head and neck cancer patients.

Findings

The Chinese PSAS showed good item discrimination and internal consistency (Cronbach’s α = 0.942).

The three-factor model demonstrated good fit indices (χ2/df = 2.595, RMSEA = 0.090).

Rasch analysis confirmed good model fit and reliability (person/item separation index > 1.5).

Abstract

Background: Patient self-advocacy plays a crucial role in improving cancer patients’ quality of life, but there is no validated instrument to assess this concept among Chinese head and neck cancer patients. This study aimed to cross-culturally translate the Patient Self-Advocacy Scale (PSAS) and evaluate its psychometric properties using classical test theory and item response theory. Methods: The PSAS underwent cross-cultural adaptation based on Brislin’s translation model and a cross-sectional survey of 302 head and neck cancer patients at a tertiary hospital in Tianjin was conducted from November 2023 to August 2024. Classical test theory was used for item analysis and validation of reliability (internal consistency, test-retest reliability) and validity (content validity, construct validity). Item response theory was applied to evaluate model fit, reliability, item difficulty, and measurement invariance. Results: Classical test theory analysis demonstrated good item discrimination with item-total correlations ranging from 0.776 to 0.942 and critical ratios from 13.269 to 33.170 (p < 0.05), as well as good internal consistency (Cronbach’s α = 0.942 for the total scale) and test-retest reliability (ICC = 0.840 for the total scale, p < 0.001). I-CVI values ranged from 0.80 to 1.00, with an S-CVI of 0.95. The three-factor model demonstrated good fit (χ2/df = 2.595, RMSEA = 0.090, SRMR = 0.072, CFI = 0.966, IFI = 0.966, TLI = 0.956). Rasch analysis indicated a good model fit and reliability (person/item separation index > 1.5, person/item reliability coefficient > 0.9). The Wright map showed good matching between item difficulty and person ability. Differential item functioning (DIF) analysis revealed no significant differences across gender. Conclusion: The Chinese version of PSAS demonstrates satisfactory psychometric properties among head and neck cancer patients and provides healthcare providers with a tool to assess patients’ self-advocacy, potentially facilitating patient-centered care and self-management in clinical practice and improving patients’ health and quality of life outcomes.

Figures