Prevalence, incidence, and mortality of inflammatory bowel disease in the Netherlands: development and external validation of machine learning models - Summary - MDSpire
Advertisement
Prevalence, incidence, and mortality of inflammatory bowel disease in the Netherlands: development and external validation of machine learning models
To develop and validate machine learning models for identifying inflammatory bowel disease (IBD) cases in administrative data, while determining the prevalence, incidence, and mortality of IBD in the Netherlands.
Key Findings:
The random forest model for identifying IBD cases achieved an AUC of 0.97 (95% CI [0.96; 0.97]).
The gradient-boosted trees model for subtype classification had an accuracy of 0.95 (95% CI [0.94; 0.95]).
The incidence of IBD was 20.1 per 100,000 in 2020, stable over time.
The prevalence of IBD was 577.6 per 100,000 as of December 31, 2020 (95% CI [566.7; 586.2]).
Mortality rates for IBD patients rose to 11.6 per 1,000 in 2020 (95% CI [10.5; 11.8]).
Interpretation:
The prevalence of IBD in the Netherlands is increasing at a slower rate than expected, indicating a potential trend towards Prevalence Equilibrium, which may have significant implications for healthcare planning.
Limitations:
The study relied on administrative data, which may have inherent inaccuracies that could affect the findings.
External validation was limited to one hospital cohort, which may not represent the entire population.
Conclusion:
Machine learning models can accurately identify IBD cases using administrative data, providing valuable insights into the epidemiology of IBD in the Netherlands.
by Reinier C A van Linschoten, Nikki van Leeuwen, David van Klaveren, Marieke J Pierik, Rob Creemers, Evelien M B Hendrix, Jan A Hazelzet, C Janneke van der Woude, Rachel L West, Desirée van Noord
Large claims analysis finds no significant differences in serious infections, blood clots, or major cardiovascular events across biologics and a Janus kinase inhibitor.
