Prevalence Incidence Mixture Models
The R package and webtool fits Prevalence Incidence Mixture models to left-censored and irregularly interval-censored time to event data that is commonly found in screening cohorts assembled from electronic health records. Absolute and relative risk can be estimated for simple random sampling, and stratified sampling (the two approaches of superpopulation and a finite population are supported for target populations). Non-parametric (absolute risks only), semi-parametric, weakly-parametric (using B-splines), and some fully parametric (such as the logistic-Weibull) models are supported.
- PIMixture Help File: PIMixture.pdf
- PIMixture.Predict Help File: PIMixture.predict.pdf
- Example Data: PIData1.pdf
- License Agreement
- Contact DCEG
- Cheung LC, Pan Q, Hyun N, Schiffman M, Fetterman B, Castle PE, Lorey T, Katki HA. Mixture models for undiagnosed prevalent disease and interval-censored incident disease: applications to a cohort assembled from electronic health records. Stat Med 2017; 30:3583-95.
- Hyun N, Cheung, LC, Pan Q, Schiffman M, Katki HA. Flexible risk prediction models for left or interval-censored data from electronic health records. Ann Applied Stat 2017; 11:1063-1084.