Inconsistencies between measures of cognitive dysfunction in childhood acute lymphoblastic leukemia survivors: Description and understanding.

OBJECTIVES: The frequency of cognitive difficulties in childhood cancer survivors varies according to the measurement strategy. The goal of this research is to (a) describe agreements and differences between measures of working memory and attention (b) identify contributors of these differences, such as emotional distress, affects, and fatigue. METHODS: We used data available for 138 adults successfully treated for childhood acute lymphoblastic leukemia (ALL) (PETALE cohort). Working memory and attention were assessed using subtests from the WAIS-IV and self-reported questionnaires (BRIEF-SR and CAARS-S:L). Potential contributors included emotional distress, anxiety, depression (BSI-18), affects (PANAS), and fatigue (PedsQL-MFS). We explored measurement agreements and differences using diagnostic indices and multivariate regression models. RESULTS: The frequencies of working memory and attention deficits were higher when using cognitive tests (15%-21%) than with self-reports (10%-11%). Self-reported questionnaires showed high specificity (median 0.87) and low sensitivity (median 0.10), suggesting they did not reliably identify positive cases on cognitive tests. We identified negative affectivity as a possible contributor to inconsistencies between self-report and test results. CONCLUSIONS: When measuring working memory and attention in childhood ALL survivors, cognitive test results and self-reports should not be considered equivalent. At best, self-report may be used for screening (high specificity), but not to assess prevalence in large samples. Self-reported difficulties are also probably influenced by the negative mood in this population.