Evaluating construct and criterion validity of NeuroScreen in assessing neurocognition among hospitalized Ugandan first-episode psychosis patients

Introduction: Neurocognitive impairment (NCI) is commonly exhibited among patients experiencing their first episode of psychosis. However, there are few resources in many low-income countries, such as Uganda, that allow for the administration of extensive neurocognitive test batteries for the detection of NCI. NeuroScreen is a brief tablet-based neurocognitive assessment battery that can be administered by all levels of healthcare staff. We examined the validity of NeuroScreen to assess neurocognition and detect NCI in first-episode psychosis (FEP) patients in Uganda. Methods: We enrolled 112 participants FEP patients and matched controls at Butabika Mental Referral Hospital. Each participant completed NeuroScreen and a traditionally administered neurocognitive battery: the MATRIC Consensus Cognitive Battery (MCCB). We examined correlations between participant performance on NeuroScreen and the MCCB. A ROC curve determined sensitivity and specificity of NeuroScreen to detect NCI as determined by MCCB criterion. Results: There was a large, statistically significant correlation between overall performance on NeuroScreen and the MCCB [r(112) = 0.64, p < .001]. Small to large correlations were found between tests in the MCCB and NeuroScreen batteries. The ROC curve of NeuroScreen performance to detect MCCB-defined NCI had an area under curve of 0.80 and optimal sensitivity and specificity of 83 % and 60 %, respectively. Conclusion: There was a moderate positive correlation between overall performance on both batteries. NeuroScreen shows promise as a valid assessment battery to assess neurocognition and detect NCI in FEP patients in Uganda. Further studies of NeuroScreen in healthy individuals and in a range of mental disorders are recommended.