Analysis for Large-scale clustering of longitudinal faecal calprotectin and C-reactive protein profiles in inflammatory bowel disease
About
This website presents the analyses carried out for Large-scale clustering of longitudinal faecal calprotectin and C-reactive protein profiles in inflammatory bowel disease by Constantine-Cooke et al. In this work, we expand upon our previous research which clustered faecal calprotectin (FC) profiles for a subset of Crohn’s disease patients (Constantine-Cooke et al. 2023) by clustering FC and C-reactive protein (CRP) profiles. Moreover, we extend our modelling to include subjects with Crohn’s disease, ulcerative colitis, and inflammatory bowel disease unclassified identified via the Lothian IBD Registry.
The analysis pipeline for this project can be categorised into three stages:
- Before model fitting, where data obtained from the NHS are reformatted and any data quality issues are dealt with. Baseline data (i.e data available at diagnosis) were also explored in descriptive analyses as part of this stage.
- Model fitting, where latent class mixed models are fitted with differing numbers of assumed clusters and the most appropriate models are chosen for FC and CRP.
- After model fitting, where post hoc associations between baseline data and cluster membership are explored. The relationship between FC and CRP clustering is also analysed.
Using this website
The navigation menu at the top of the page will allow you to navigate through the steps of the analysis pipeline. The code button at the top of each page can be used to show all code blocks instead of clicking on the code buttons for each individual block of code. Moving the mouse pointer over any citations will produce a pop up box with reference details. Clicking on the citation will link to the bibliography at the bottom of the page.
Software
All analyses have been generated using R version 4.4.2.
Please see the Session Information sections at the end of each page for the R packages and the respective versions used.
This website has been generated using the Quarto scientific publishing system built on top of pandoc. A list of major components used by Quarto can be found here