Abstract
Background: The glutamatergic modulator ketamine has been shown to result in rapid reductions in both suicidal ideation (SI) and depressive symptoms in clinical trials. There is a practical need for identification of pre-treatment predictors of ketamine response. Previous studies indicate links between treatment response and body mass index (BMI), depression symptoms and previous suicide attempts. Our aim was to explore the use of clinical and demographic factors to predict response to serial doses of oral ketamine for chronic suicidal ideation.
Methods: Thirty-two participants completed the Oral Ketamine Trial on Suicidality (OKTOS). Data for the current study were drawn from pre-treatment and follow-up time-points of OKTOS. Only clinical and sociodemographic variables were included in this analysis. Data were used to create a proof of concept Bayesian network (BN) model of variables predicting prolonged response to oral ketamine, as defined by the Beck Scale for Suicide Ideation (BSS).
Results: The network of potential predictors of response was evaluated using receiver operating characteristic (ROC) curve analyses. A combination of nine demographic and clinical variables predicted prolonged ketamine response, with strong contributions from BMI, Social and Occupational Functioning Assessment Scale (SOFAS), Montgomery-Asberg Depression Rating Scale (MADRS), number of suicide attempts, employment status and age. We evaluated and optimised the proposed network to increase the area under the ROC curve (AUC). The performance evaluation demonstrated that the BN predicted prolonged ketamine response with 97% accuracy, and AUC = 0.87.
Conclusions: At present, validated tools to facilitate risk assessment are infrequently used in psychiatric practice. Pre-treatment assessment of individuals’ likelihood of response to oral ketamine for chronic suicidal ideation could be beneficial in making more informed decisions about likelihood of success for this treatment course. Clinical trials registration number ACTRN12618001412224, retrospectively registered 23/8/2018.
Beaudequin, D., Can, A. T., Dutton, M., Jones, M., Gallay, C., Schwenn, P., Yang, C., Forsyth, G., Simcock, G., Hermens, D. F., & Lagopoulos, J. (2020). Predicting therapeutic response to oral ketamine for chronic suicidal ideation: a Bayesian network for clinical decision support. BMC psychiatry, 20(1), 519. https://doi.org/10.1186/s12888-020-02925-1
