Purpose: To validate the FACT-G scale for measuring quality of life of patients with cancer in Colombia.
Methods: The analysis included factor analysis, confirmatory analysis, Rasch analysis, convergent validity, internal consistency (473 patients diagnosed with cancer), test-retest reliability (97 patients evaluated at two different time points) and sensitivity to change (25 patients evaluated before and after an intervention).
Results: A four-factor structure has been found ("Physical well-being", "Social-family well-being", "Functional well-being" and "Emotional well-being"). Two subscales ("Emotional well-being" and "Social-family well-being") have misfitting items. Cronbach's alpha was 0.89 for the whole scale. None of the items had significant impact on the scale's alpha when removed. Lin's concordance correlation coefficient indicated test-retest reliability (rho c: 0.64-0.76) adequate to the uses of the tool. Regarding sensitivity to change, repeated measures analysis demonstrated significant change of the score after an intervention [F(3, 72) = 39.89, P = 0.000]. Except for the domain "Social-family well-being", Pearson's correlation coefficient between equivalent domain scores on FACT-G and the EORTC QLQC-30 ranged from 0.5 to 0.7.
Conclusions: The FACT-G scale measures a four-factor construct. Results indicate that the FACT-G scale is an instrument that performs consistently over time, with evidence of responsiveness. The finding of misfitting items in two subscales ("Social-family well-being", and "Emotional well-being") imposes caution in interpreting the scores of these domains.