This package contains a number of vignettes, each one walking through an example analysis. The table below gives an overview.
Many of these examples recreate analyses from the series of Technical Support Documents published by the NICE Decision Support Unit (Dias et al. 2011). The exceptions are atrial fibrillation (Cooper et al. 2009), white blood cell transfusion (Turner et al. 2012), and plaque psoriasis multilevel network meta-regression (Phillippo et al. 2020).
Title | Outcome type | Likelihood | Link function | Notable features |
---|---|---|---|---|
Blocker | Counts | Binomial | logit | Pairwise MA |
Dietary fat | Rates | Poisson | log | Analysis of log rate ratios from rate data |
Diabetes | Counts with time at risk | Binomial | cloglog | Analysis of log hazard ratios from count data with time at risk |
Parkinson’s | Continuous | Normal | Identity | Analysis of arm-based data, contrast-based data, and a mixture of both |
HTA plaque psoriasis | Ordered | Multinomial (ordered) | probit | Analysis of ordered categorical outcomes |
Statins | Counts | Binomial | logit | Meta-regression with subgroups |
BCG vaccine | Counts | Binomial | logit | Meta-regression with a continuous covariate, predictive distributions |
Smoking cessation | Counts | Binomial | logit | Assessing inconsistency with unrelated mean effects and node-splitting models |
Thrombolytics | Counts | Binomial | logit | Assessing inconsistency with unrelated mean effects and node-splitting models |
Atrial fibrillation | Counts | Binomial | logit | Meta-regression with shared class interactions |
WBC transfusion | Counts | Binomial | logit | Informative log-Normal prior on \(\tau^2\) |
ML-NMR plaque psoriasis | Binary (IPD) and counts (AgD) | Bernoulli (IPD) and two-parameter Binomial (AgD) | probit | Multilevel network meta-regression combining IPD and AgD |