Validation and Extension of a Risk Calculator to Predict Mood Recurrence in Young People with Bipolar Disorder
Avolio, A.; Merranko, J.; Gill, M. K.; Levenson, J. C.; Goldstein, T. R.; Hafeman, D.; Birmaher, B.
Show abstract
ObjectiveGiven the episodic nature of bipolar disorder (BD) and the variability in mood episode recurrence across individuals, accurate recurrence prediction is critical. The original COBY recurrence risk calculator (RC) was developed in a longitudinal youth cohort to estimate threshold recurrence risk. However, its accuracy for predicting subthreshold recurrences had not been fully evaluated. The objective of this study was to extend the previously developed COBY mood recurrence RC to predict both threshold and subthreshold mood recurrences and evaluate its performance in an independent sample. MethodAdolescents and young adults with BD-I/II (N= 51; BD-I: 38, BD-II: 13; 14-24 years old) were interviewed with standard instruments at intake and during the follow-up on average every 6 months for a median of 54 weeks. We assessed the degree to which the COBY RC predicted mood recurrence (threshold or subthreshold) in this independent sample. Discrimination was measured using the area under the receiver operating characteristic curves (AUC); calibration and variable importance were also assessed. ResultsThe model demonstrated good prediction of any recurrence within the next six months (any threshold recurrence AUC = 0.72, any subthreshold or worse recurrence AUC = 0.77). Calibration analysis demonstrated the model tended to overestimate risk in the external sample, plausibly attributable to differences in recurrence ascertainment strategy (prospective vs retrospective) or the significant difference in prior remission length, a key predictor. Recalibration greatly improved calibration without loss of discrimination. ConclusionThe mood recurrence RC demonstrated good discrimination for both threshold and subthreshold mood recurrences in an independent young adult cohort, consistent with prior youth and adult validations. Validation now spans across developmental stages and different degrees of severity of mood symptoms opening the opportunity for clinical implementation to provide personalized monitoring and early intervention for people with BD.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Predicting remission after internet-delivered psychotherapy in patients with depression using machine learning and multi-modal data 96%
- Distinguishing clinical and genetic risk factors for suicidal ideation and behavior in a diverse hospital population 93%
- Differential and spatial expression meta-analysis of genes identified in genome-wide association studies of depression 92%
Similar papers in this journal
- Low predictive power of clinical features for relapse prediction after antidepressant discontinuation in a naturalistic setting 94%
- Isolating the genetic component of mania in bipolar disorder 93%
- Atypical depression is associated with a distinct clinical, neurobiological, treatment response and polygenic risk profile 93%
Similar papers in this journal
- Sociodemographic, clinical, and genetic factors associated with self-reported antidepressant response outcomes in the UK Biobank 93%
- A computational approach to understanding effort-based decision-making in depression 93%
- Using Major Depression Polygenic Risk Scores to Explore the Depressive Symptom Continuum 92%
Similar papers in this journal
- Latent subtypes of manic or irritable episode symptoms in two population-based cohorts 96%
- Predictors of suicidal thoughts and behavior in children: results from penalized logistic regression analyses in the ABCD study 93%
- Smoking and the risk for bipolar disorder: causal evidence from a bidirectional Mendelian randomization study 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.