Related Experiment Video
Updated: May 13, 2025
High-speed Video Microscopy Analysis for First-line Diagnosis of Primary Ciliary Dyskinesia
Published on: January 19, 2022
Feasibility of Machine Learning Analysis for the Identification of Patients with Possible Primary Ciliary Dyskinesia
Gully Burns1, Carey Kauffman2, Michele Manion2
1Chan Zuckerberg Initiative, PO BOX 8040, Redwood City, CA 94063.
Insights
Machine learning can screen for primary ciliary dyskinesia (PCD) in children using health claims data. This approach aids early identification of PCD, improving health outcomes for this rare disease.
Area of Science:
- Medical Informatics
- Rare Disease Research
- Machine Learning Applications
Background:
- Primary ciliary dyskinesia (PCD) is a rare genetic disorder often diagnosed late, leading to delayed treatment and poorer health outcomes.
- Scalable screening tools are needed for early identification of pediatric patients with PCD.
Purpose of the Study:
- To evaluate the feasibility of using machine learning (ML) to screen for primary ciliary dyskinesia (PCD) in pediatric populations.
- To assess the utility of claims data for identifying potential PCD cases without a specific diagnostic code.
Main Methods:
- A random forest ML model was developed using data from the PCD Foundation Registry and a national claims database.
- Features were derived from diagnostic, procedural, and pharmaceutical codes associated with PCD and related conditions.
- Models were trained and validated using confirmed PCD cases, patients with related diagnoses (Q34.8+EM), and matched control groups.
Main Results:
- Initial models showed variable performance; expanding the dataset improved screening suitability (PPV 0.51-0.54, sensitivity 0.82-0.90).
- The ML model identified 7,705 potential PCD cases in a cohort of 1.32 million pediatric patients, aligning with estimated prevalence.
- Synthetic data augmentation did not significantly enhance model performance.
Conclusions:
- Machine learning models can effectively screen for primary ciliary dyskinesia (PCD) using readily available health claims data.
- This ML-based screening approach can help identify individuals who may benefit from timely diagnostic evaluation and intervention for PCD.
- The feasibility of using ML highlights potential for improving early detection in rare diseases lacking specific ICD codes.
Background:
Significant diagnostic delays are common in primary ciliary dyskinesia (PCD), a rare disease that is significantly underdiagnosed. Scalable screening methods could improve early identification and health outcomes.
Research Question:
Can machine learning (ML) be used to screen for PCD in pediatric patients?
Study Design And Methods:
We evaluated the feasibility of a random forest model to screen for PCD using data from the PCD Foundation Registry and a national claims database. We identified a cohort of pediatric patients with diagnostic codes indicative of conditions potentially associated with PCD, and studied diagnostic, procedural, and pharmaceutical codes associated with PCD to develop ML features. Models were trained on composite claims data from confirmed patients with PCD, patients with Q34.8 (Specific Congenital Malformation of the Respiratory System) diagnosed within six months of an Electron Microscopy procedure (Q34.8+EM), and a randomly-selected, matched control group. Model performance was tested through 5-fold cross-validation.
Results:
Using 82 confirmed PCD cases and 4,161 matched controls, the model demonstrated variable performance (positive predictive value 0.45-0.73, sensitivity 0.75-0.94). Synthetic data augmentation did not improve results (positive predictive value 0.45-0.67, sensitivity 0.71-1.00). Expanding the dataset to include 319 Q34.8+EM patients and 8,214 controls improved performance (positive predictive value 0.51-0.54, sensitivity 0.82-0.90), suitable for screening. In a cohort of 1.32 million pediatric patients, 7,705 were classified as positive, consistent with the estimated prevalence of PCD (1:7,554).
Interpretation:
This study demonstrates the feasibility of using ML to screen for PCD using claims data, even in the absence of a specific International Classification of Disease (ICD) code. Such screening approaches may aid in the identification of individuals who may benefit from timely diagnostic testing and targeted interventions.
More Related Videos
09:03Nasal Brushing Sampling and Processing Using Digital High Speed Ciliary Videomicroscopy – Adaptation for the COVID-19 Pandemic
Published on: November 7, 2020
11:13Collection, Expansion, and Differentiation of Primary Human Nasal Epithelial Cell Models for Quantification of Cilia Beat Frequency
Published on: November 10, 2021