Psychometric Evaluation of the ASCOT-SCT4, a Preference-Based Quality of Life Instrument for Social Care Service Users in Hong Kong: An Application of Item Response Theory
Author(s)
Siyue Yu, MPH1, Judy Sze, BSc1, Annie Wai-ling Cheung, MPhil1, Richard Xu, PhD2, Eliza L. Wong, PhD1;
1The Chinese University of Hong Kong, Hong Kong, China, 2The Hong Kong Polytechnic University, Hong Kong, China
1The Chinese University of Hong Kong, Hong Kong, China, 2The Hong Kong Polytechnic University, Hong Kong, China
Presentation Documents
OBJECTIVES: The Adult Social Care Outcomes Toolkit-Self-Completion Tool (ASCOT-SCT4) is designed to measure quality of life (QoL) among social care users. While it has been validated using classical test theory, limited studies have used item response theory (IRT) to evaluate its item-level psychometric properties. This study aimed to evaluate the Hong Kong Chinese ASCOT-SCT4 by using IRT.
METHODS: A cross-sectional study was conducted from June 2023 to April 2024, involving 312 participants aged 60 or above who were receiving community care and support services (mean age = 76.0 years, SD = 8.64; 70.2% female). The ASCOT-SCT4 comprises eight domains assessed through eight items, each item with four response options. Confirmatory factor analysis (CFA) was used to evaluate the unidimensionality of the instrument. A two-parameter graded response model (GRM) was applied to evaluate item discrimination (a) and item difficulty (b1, b2, b3).
RESULTS: CFA confirmed the unidimensionality of the ASCOT-SCT4, with acceptable model fit indices. Factor loadings exceeded 0.40 for all items except ‘Dignity’ (0.27). GRM result revealed that six of the eight items exhibited high discrimination (a ≥1.35), ranging from 1.35 (‘Personal cleanliness’) to 2.02 (‘Social contact’), indicating strong ability to differentiate participants across levels of QoL. However, two items had lower discrimination: ‘Personal cleanliness (a=0.99), and ‘Dignity’ (a=0.46). Difficulty parameters for most items were well-distributed across the latent trait scale (-0.12 to 4.12), except for the ‘Dignity’ item, which had an extreme difficulty threshold (b3=8.78). The ‘Social contact’ item had the highest item information, while ‘Dignity’ had the least.
CONCLUSIONS: The Hong Kong Chinese ASCOT-SCT4 demonstrates good psychometric properties, supporting its use as a valid and reliable instrument for measuring QoL among social care service users. However, the ‘Dignity’ item warrants further investigation due to its low discrimination and extreme difficulty threshold.
METHODS: A cross-sectional study was conducted from June 2023 to April 2024, involving 312 participants aged 60 or above who were receiving community care and support services (mean age = 76.0 years, SD = 8.64; 70.2% female). The ASCOT-SCT4 comprises eight domains assessed through eight items, each item with four response options. Confirmatory factor analysis (CFA) was used to evaluate the unidimensionality of the instrument. A two-parameter graded response model (GRM) was applied to evaluate item discrimination (a) and item difficulty (b1, b2, b3).
RESULTS: CFA confirmed the unidimensionality of the ASCOT-SCT4, with acceptable model fit indices. Factor loadings exceeded 0.40 for all items except ‘Dignity’ (0.27). GRM result revealed that six of the eight items exhibited high discrimination (a ≥1.35), ranging from 1.35 (‘Personal cleanliness’) to 2.02 (‘Social contact’), indicating strong ability to differentiate participants across levels of QoL. However, two items had lower discrimination: ‘Personal cleanliness (a=0.99), and ‘Dignity’ (a=0.46). Difficulty parameters for most items were well-distributed across the latent trait scale (-0.12 to 4.12), except for the ‘Dignity’ item, which had an extreme difficulty threshold (b3=8.78). The ‘Social contact’ item had the highest item information, while ‘Dignity’ had the least.
CONCLUSIONS: The Hong Kong Chinese ASCOT-SCT4 demonstrates good psychometric properties, supporting its use as a valid and reliable instrument for measuring QoL among social care service users. However, the ‘Dignity’ item warrants further investigation due to its low discrimination and extreme difficulty threshold.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
PCR253
Topic
Patient-Centered Research
Topic Subcategory
Health State Utilities, Instrument Development, Validation, & Translation, Patient-reported Outcomes & Quality of Life Outcomes
Disease
SDC: Geriatrics