Ecological validity of executive function tests in moderate traumatic brain injury in Ghana

Abstract

Objective: While executive functioning (EF) tests are frequently administered in several Sub-Saharan African countries, studies examining their predictive relationships with real-world behaviors (i.e. ecological validity) are nonexistent. The present study investigated the predictive relationship between the Stroop Test, Controlled Oral Word Association Test, and Trail Making Test (TMT), a general cognitive screening test, Revised Quick Cognitive Screening Test (RQCST), and measures of activities of daily living, quality of life, and cognitive failures in Ghana. Method: A total of 50 literate urban dwellers who were diagnosed with moderate traumatic brain injury (TBI) were administered the neuropsychological tests and the self-report measures stated above. The informant version of the Cognitive failure questionnaire (CFQ) was completed by 50 ‘significant other’ who knew the patients very well. Results: There was no statistically significant difference between the self and informant versions of the CFQ. Some EF test scores, specifically the Stroop Test, TMT and EF composite scores, correlated significantly with the outcome measures, with correlations ranging from .29 to .55. The RQCST explained 40–49% variance in the outcome measures, while the addition of the EF composite score not only resulted in 57–62% variance accounted for but also added incremental validity to the RQCST in predicting the behavioral measures, with the exception of cognitive failures. Conclusion: This study has shown that although EF test scores, specifically the Stroop Test, TMT and EF composite scores, can be used to predict real-world behavior after moderate TBI in Ghana, such predictions are likely to be limited. The general implication for cross-cultural neuropsychology is that the (limited) ecological validity of EF tests may not necessarily be affected by whether the tests were administered in settings where they have not been standardized. This argument is, however, tenable granted that the test taker’s backgrounds are similar to those on which the tests have been standardized.