Development of a prediction model for infant hospitalisation and death using clinical features assessed by community health workers during routine postnatal home visits in Dhaka, Bangladesh - Summary - MDSpire
Advertisement
Development of a prediction model for infant hospitalisation and death using clinical features assessed by community health workers during routine postnatal home visits in Dhaka, Bangladesh
To develop a prediction model for infant hospitalization and/or death using clinical features assessed by community health workers during home visits.
Approach:
Data Analysis: Analyzed data from 1906 generally healthy infants assessed at 11 scheduled CHW visits from 3 to 60 days of age.
Model Development: Developed two models—time-varying Cox regression and random forest—using 45 clinical features, including WHO danger signs and additional covariates.
Performance Evaluation: Evaluated discrimination (C-statistic) and calibration, comparing performance with a Cox model using only WHO danger signs.
Key Findings:
9.2% of infants had an event (173 hospitalizations, 3 deaths).
Best-performing Cox model (C-statistic=0.71; 95% CI 0.68 to 0.75) included baseline covariates and visit-specific features.
Random forest model (C-statistic=0.69; 95% CI 0.64 to 0.73) did not outperform the Cox model.
Adding four visit-specific features to WHO danger signs improved predictive performance.
Interpretation:
Aggregative features and random forest did not enhance prediction compared to a Cox model using baseline covariates and visit-specific features.
Limitations:
The study was limited to generally healthy infants and may not generalize to all populations.
The predictive models may require further validation in diverse settings.
Conclusion:
Adding visit-specific clinical features to WHO danger signs may improve predictive performance in identifying infants needing referral.
by Alastair Fung, Marimuthu Sappani, Cole Heasley, Chun-Yuan Chen, Shaun K Morris, Peter J Gill, Diego G Bassani, Davidson H Hamer, Prakesh S Shah, S M Abdul Gaffar, Sultana Yeasmin, Shafiqul A Sarker, Shamima Sultana, Joseph Beyene, Daniel E Roth
Federal prosecutors allege that a Florida physician and research staff fabricated clinical trial records that were submitted into database systems used to evaluate investigational drugs.
