Prospective prediction of childhood body mass index trajectories using multi-task Gaussian processes
An erratum to this article was published (Int J Obes (Lond), Feb 2025 ↗)
Authors' conclusion
MagmaClust provides a unified, probabilistic, non-parametric framework to model, cluster, and prospectively predict childhood BMI trajectories and overweight/obesity risk. The proposed method offers a convenient tool for clinicians to monitor BMI growth in children, allowing them to prospectively identify children with high predicted overweight/obesity risk and implement timely interventions.
Study context
- Population
- Children of a mother-offspring cohort with longitudinal body mass index (BMI) measurements between birth and age 10 y
- Comparison
- MagmaClust vs cubic B-spline and multilevel Jenss-Bayley models
- Outcome
- Prediction of overweight/obesity status at age 10 y from BMI data between birth and age 2 y
- Follow-up
- Age 10 years
Methodological detailsmother-offspring cohort, longitudinal, multi-task Gaussian process (MagmaClust)
- Mother-offspring cohort
- Longitudinal
- Multi-task Gaussian process (MagmaClust)
Performance
As reported in the abstract
| Test or model | Specificity | NPV | Accuracy |
|---|---|---|---|
| MagmaClust | 0.94 | 0.89 | 0.86 |
MagmaClust
Further findings
BMI trajectory clusters · 0 to 10 y
- Distinct patterns of BMI trajectories identified by MagmaClust
- 5
Accuracy of retrospective BMI trajectories and robustness to missing data
- MagmaClust outperformed cubic B-spline and multilevel Jenss-Bayley models, while more robust to missing data
- up to 90%
Prospective forecasting of BMI trajectories · 2 to 8 years into the future
- MagmaClust better at prospectively forecasting BMI trajectories using historic BMI data
Original publication
Composed by AI from the published abstract and checked against it. How Briefs are made →