Presentation Information
[P-19-06]Distinguishing Late-Life Depression from Dementia: A Predictive Model Based on Clinical Presentation
*Quankamon Dejatiwongse Na Ayudhya, Punyisa Prachgosin, Kritta Supanimitamorn, Nipaporn Supaprasert, Parawee Keawjamrus, Kankamol Jaisin (Department of Psychiatry, Faculty of Medicine Siriraj Hospital, Mahidol University(Thailand))
Keywords:
Late-life depression,dementia,predictive model,geriatric psychiatry
Introduction: Differentiating late-life depression from dementia remains a diagnostic challenge due to overlapping clinical features. Older adults with depression often present with cognitive impairment — commonly termed pseudodementia — while those in the early stages of dementia may exhibit depressive symptoms. This overlap may result in misdiagnosis and delays in appropriate treatment.
Objective: This study aimed to develop a predictive model based on the chief complaint at presentation and routinely available clinical data from an outpatient geriatric psychiatry setting to differentiate late-life depression from dementia in older adults.
Methods: We conducted a retrospective chart review of 141 patients aged 60 or older who presented to a geriatric psychiatry outpatient clinic. Chief complaints at initial evaluation were categorized and analyzed as predictors in a binary logistic regression model, with diagnosis of either depression or dementia as the dependent variable. Key covariates included age, sex, and cognitive screening scores. Model performance was evaluated using Nagelkerke R² and receiver operating characteristic (ROC) analysis.
Results: Binary logistic regression analysis revealed that chief complaints involving neurovegetative symptoms — such as sleep disturbance, fatigue, and appetite change — significantly predicted depression (OR = 5.70, 95 percent confidence interval 1.36 to 23.88, p = 0.017), as did depression-related symptoms — including sadness, boredom, and death wish (OR = 8.18, 95 percent confidence interval 1.73 to 38.65, p = 0.008). Cognitive complaints were strongly associated with dementia (OR = 0.06, 95 percent confidence interval 0.01 to 0.27, p < 0.001), while anxiety-related symptoms were not statistically significant in the model (OR = 2.01, 95 percent confidence interval 0.42 to 9.70, p = 0.383). The model showed robust explanatory power (Nagelkerke R² = 0.652) and excellent discriminative accuracy (AUC = 0.935).
Conclusions: These findings emphasize the diagnostic value of chief complaints in distinguishing late-life depression from dementia. Neurovegetative and depression-related symptoms indicate depression, while cognitive complaints suggest dementia. Incorporating these clinical cues into routine assessments may enhance diagnostic accuracy, especially in resource-limited settings.
Objective: This study aimed to develop a predictive model based on the chief complaint at presentation and routinely available clinical data from an outpatient geriatric psychiatry setting to differentiate late-life depression from dementia in older adults.
Methods: We conducted a retrospective chart review of 141 patients aged 60 or older who presented to a geriatric psychiatry outpatient clinic. Chief complaints at initial evaluation were categorized and analyzed as predictors in a binary logistic regression model, with diagnosis of either depression or dementia as the dependent variable. Key covariates included age, sex, and cognitive screening scores. Model performance was evaluated using Nagelkerke R² and receiver operating characteristic (ROC) analysis.
Results: Binary logistic regression analysis revealed that chief complaints involving neurovegetative symptoms — such as sleep disturbance, fatigue, and appetite change — significantly predicted depression (OR = 5.70, 95 percent confidence interval 1.36 to 23.88, p = 0.017), as did depression-related symptoms — including sadness, boredom, and death wish (OR = 8.18, 95 percent confidence interval 1.73 to 38.65, p = 0.008). Cognitive complaints were strongly associated with dementia (OR = 0.06, 95 percent confidence interval 0.01 to 0.27, p < 0.001), while anxiety-related symptoms were not statistically significant in the model (OR = 2.01, 95 percent confidence interval 0.42 to 9.70, p = 0.383). The model showed robust explanatory power (Nagelkerke R² = 0.652) and excellent discriminative accuracy (AUC = 0.935).
Conclusions: These findings emphasize the diagnostic value of chief complaints in distinguishing late-life depression from dementia. Neurovegetative and depression-related symptoms indicate depression, while cognitive complaints suggest dementia. Incorporating these clinical cues into routine assessments may enhance diagnostic accuracy, especially in resource-limited settings.