Psychometric properties and measurement equivalence of the English and Chinese versions of the Beck Anxiety Inventory in patients with breast cancer

• Published in: Supportive care in cancer 2017-02, Vol.25 (2), p.633-643
• Main Authors: Ke, Yu, Ng, Terence, Yeo, Hui Ling, Shwe, Maung, Gan, Yan Xiang, Chan, Alexandre
• Format: Article
• Language: English
• Subjects: Analysis; Anxiety; Anxiety - psychology; Asian Continental Ancestry Group - psychology; Breast cancer; Breast Neoplasms - psychology; Cancer patients; Female; Humans; Language; Measurement; Medical research; Medicine; Medicine & Public Health; Middle Aged; Neurophysiology; Nursing; Nursing Research; Oncology; Original Article; Pain Medicine; Prospective Studies; Psychometrics - methods; Quality of Life - psychology; Quantitative psychology; Rehabilitation Medicine; Reproducibility of Results; Surveys and Questionnaires; Validation studies
• ISSN: 0941-4355; 1433-7339
• DOI: 10.1007/s00520-016-3452-3

Background There is a lack of psychometric data for both the English and Chinese versions of Beck Anxiety Inventory (BAI) to support its usage among breast cancer patients. This study examined the psychometric properties and measurement equivalence of the English and Chinese versions of BAI among breast cancer patients in Singapore. Methods Patients were recruited from two major cancer centers in Singapore. The criterion and construct validity of BAI was assessed by its correlation strength with (1) the emotional functioning subdomain of EORTC QLQ-C30 and (2) constructs related to anxiety, namely fatigue, dyspnea, and quality of life. The known-group validity was assessed according to the patients’ breast cancer stage, religious beliefs, and emotional functioning levels. The internal consistency of the BAI domains was evaluated using Cronbach’s alpha coefficient. Regression analysis was performed to compare the BAI total and domain scores between the two language versions. Results Data from 244 patients (144 English-speaking and 100 Chinese-speaking) were analyzed. For both language versions, the BAI total scores correlated moderately with the EORTC QLQ-C30 emotional functioning subdomain ( r  = −0.655 and −0.601). Correlations with fatigue, quality of life, and dyspnea were moderate (| r | = 0.456–0.606). Patients with poorer emotional functioning reported higher anxiety levels, establishing known-group validity. All BAI domains demonstrated satisfactory internal consistencies ( α  = 0.74–0.87), except for the panic domain ( α =  0.57–0.61). Possible measurement equivalence between the language versions was established. Conclusion Both English and Chinese versions of BAI are valid, reliable, and possibly equivalent for future use.