Andrew Cooper1, Helen Aucote. 1. Department of Psychology, Goldsmiths, University of London, New Cross, London, SE14 6NW, UK. a.cooper@gold.ac.uk
Abstract
OBJECTIVES: To use confirmatory factor analysis (CFA) to test the proposed factor structure of the Psychological Consequences Questionnaire (PCQ), a measure of the psychological impact of breast cancer screening. A further aim was to examine the robustness of the proposed factor structure across key demographic and clinical variables. METHOD: Following visits to breast cancer screening clinics, women who received a false-positive diagnosis and a matched sample of women who had received all-clear diagnoses were sent a questionnaire package containing the PCQ and a demographics measure. A total of 220 women returned completed questionnaires. CFA was used to test the factor structure and multiple indicator-multiple cause (MIMIC) models were used to test the robustness of the factor structure across the test result group, age, and family history of breast cancer diagnosis. RESULTS: The CFA results suggested support for both a three- and a one-factor model; a one-factor model was preferred, however, due to the very high covariance between the three latent factors in the three-factor model. A CFA MIMIC model suggested that the test result impacted on the latent factor: women who initially received a false-positive diagnosis showed significantly higher levels of psychological dysfunction after screening. CONCLUSIONS: The PCQ appears to be a promising tool for assessing psychological dysfunction after breast cancer screening; however, a one-factor model received more support than the initially proposed three-factor model. There was little evidence of differential item functioning across key demographic and clinical variables for the PCQ.
OBJECTIVES: To use confirmatory factor analysis (CFA) to test the proposed factor structure of the Psychological Consequences Questionnaire (PCQ), a measure of the psychological impact of breast cancer screening. A further aim was to examine the robustness of the proposed factor structure across key demographic and clinical variables. METHOD: Following visits to breast cancer screening clinics, women who received a false-positive diagnosis and a matched sample of women who had received all-clear diagnoses were sent a questionnaire package containing the PCQ and a demographics measure. A total of 220 women returned completed questionnaires. CFA was used to test the factor structure and multiple indicator-multiple cause (MIMIC) models were used to test the robustness of the factor structure across the test result group, age, and family history of breast cancer diagnosis. RESULTS: The CFA results suggested support for both a three- and a one-factor model; a one-factor model was preferred, however, due to the very high covariance between the three latent factors in the three-factor model. A CFA MIMIC model suggested that the test result impacted on the latent factor: women who initially received a false-positive diagnosis showed significantly higher levels of psychological dysfunction after screening. CONCLUSIONS: The PCQ appears to be a promising tool for assessing psychological dysfunction after breast cancer screening; however, a one-factor model received more support than the initially proposed three-factor model. There was little evidence of differential item functioning across key demographic and clinical variables for the PCQ.