The development and validation testing of a comprehensive frailty assessment in women with breast cancer.

Abstract

Background: Women with breast cancer are known to suffer from disease and treatment, and the generic measurement tools may underestimate their frailty. A specific instrument comprehensively measuring frailty among women with breast cancer has not yet been developed. This study aims to develop and validate the tool of breast cancer comprehensive frailty scale (BCCFS).

Methods: A descriptive and explorative study design was used. We collected the data through systematic literature and modified Delphi method. After an initial search and screening process, a total of 33 articles were included for review and consideration in the item design. Ten experts were invited to generate and validate initial items. The validity was assessed using a sample of 205 women with breast cancer in Taiwan. Its validity was then tested using item analysis, exploratory factor analysis, confirmatory factor analysis, criterion-related validity and areas under the receiver-operating characteristic, while its reliability was evaluated through internal consistencies and test-retest analyses.

Results: A three-factor solution with 16 items was chosen and accounted for approximately 58.57% of the total variance by exploratory factor analysis (KMO = 0.85; Bartlett's Test of Sphericity: χ2 = 2881.34, p < 0.001). The factors were interpreted as (1) deterioration of body and mobility, (2) negative emotions, and (3) cognitive impairment. The goodness of fit indices of the confirmatory factor analysis were as follows: chi-square = 234.498 (p < 0.01), normed chi-square = 2.322, SRMR = 0.055, RMSEA = 0.08, CFI = 0.930, and LI = 0.917. The Cronbach's alpha calculated for the BCCFS (16 items) was 0.91 (95% confidence interval: 0.89 to 0.93), and the test-retest reliability coefficient was 0.60. Using the G8 screening tool as a standard indicator of frailty, analysis of receiver operating characteristic curve showed that 31.5 was the best cut point (area under curve = 0. 816, 95% confidence interval: 0.757 to 0.874) with a sensitivity of 63.5% and specificity of 84.4%.

Conclusion: The instrument exhibited acceptable psychometric properties, proving it to be a valuable tool for evaluating frailty in women with breast cancer. Further assessments of its reliability, validity, and generality from health providers' views in different contexts and cultures are recommended.