Validation of the delirium diagnostic tool-provisional (DDT-Pro) in geriatric medical inpatients with diagnostic permutations of the 3Ds with and without delirium

María Botero Urrea,Maria Carolina González,María Margarita Villa García,Marcela Alviz Núñez,Juan D Velásquez-Tirado,María V Ocampo,Paula T Trzepacz,José G Franco
DOI: https://doi.org/10.1016/j.jpsychores.2024.111880
2024-08-06
Abstract:Objective: Validations of brief delirium tools have not included analysis of psychiatric disorders comorbidities or control groups. We validated the Delirium Diagnostic Tool-Provisional (DDT-Pro) in 422 geriatric inpatients with high incidence of depression and/or dementia. Methods: Cross-sectional study using two delirium reference standards, DSM-5-TR and Delirium Rating Scale-Revised-98 (DRS-R98). We assessed concurrent and construct DDT-Pro validity too. Results: There were 117 (27.7%) delirium cases using DDT-Pro, 104 (24.6%) per DSM-5-TR and 93 (22.0%) per DRS-R98; 133 patients (31.5%) had depression and 105 (24.9%) dementia, some comorbid with delirium. DDT-Pro accuracy (AUC under ROC curve) ranges were 88.3-95.9% vs DSM-5-TR and 92.7-95.0% vs DRS-R98 for whole sample and four diagnostic groups, without statistical differences. DDT-Pro ≤6 had the most balanced sensitivity-specificity for delirium diagnosis against both DSM-5-TR and DRS-R98 with similar specificity but higher sensitivity for DRS-R98 than DSM-5-TR delirium, with the highest values in patients with depression and dementia (≥92% sensitivity, ≥81% specificity). Positive and negative likelihood ratios support diagnostic strength. Concurrent validity was high reflected by significant correlations (p < 0.001) of DDT-Pro total and item scores with DRS-R98 and Delirium Frontal Index scores, highest in groups with comorbid depression and/or dementia. The DDT-Pro represented a single construct for delirium demonstrated by one factor with high item loadings and high internal consistency reliability of its items. Conclusions: The DDT-Pro demonstrated strong performance metrics in general hospital elderly inpatients with preexisting depression and/or dementia, which is unique among brief delirium tools. Its optimized cutoff score was the same as in other populations.
What problem does this paper attempt to address?