Description Usage Arguments Details Value References See Also Examples

Functions here are to take the orginized data (output from
`preprocess`

) and apply the three stages Bayesian Hierarchical Model.
See details for model description and difference between each function.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 | ```
Hier_history(aedata, n_burn, n_iter, thin, n_adapt, n_chain,
alpha.gamma = 3, beta.gamma = 1, alpha.theta = 3, beta.theta = 1,
mu.gamma.0.0 = 0, tau.gamma.0.0 = 0.1, alpha.gamma.0.0 = 3,
beta.gamma.0.0 = 1, lambda.alpha = 0.1, lambda.beta = 0.1,
mu.theta.0.0 = 0, tau.theta.0.0 = 0.1, alpha.theta.0.0 = 3,
beta.theta.0.0 = 1)
sum_Hier(hierraw)
Hier(aedata, n_burn, n_iter, thin, n_adapt, n_chain, alpha.gamma = 3,
beta.gamma = 1, alpha.theta = 3, beta.theta = 1,
mu.gamma.0.0 = 0, tau.gamma.0.0 = 0.1, alpha.gamma.0.0 = 3,
beta.gamma.0.0 = 1, lambda.alpha = 0.1, lambda.beta = 0.1,
mu.theta.0.0 = 0, tau.theta.0.0 = 0.1, alpha.theta.0.0 = 3,
beta.theta.0.0 = 1)
Hiergetpi(aedata, hierraw)
Hierplot(hierdata, ptnum = 10, param = "risk difference",
OR_xlim = c(0, 5))
Hiertable(hierdata, ptnum = 10, param = "risk difference")
Compareplot(modeldata, ptnum = 50, param = "risk difference")
``` |

`aedata` |
output from function |

`n_burn` |
integer, number of interations without saving posterior samples |

`n_iter` |
integer, number of interations saving posterior samples with every |

`thin` |
integer, samples are saved for every |

`n_adapt` |
integer, number of adaptations |

`n_chain` |
number of MCMC chains |

`hierdata` |
utput from function |

`ptnum` |
positive integer, number of AEs to be selected or plotted, default is 10 |

`param` |
a string, either "odds ratio" or "risk difference", indicate which summary statistic to be based on to plot the top AEs, default is "risk difference" |

`modeldata` |
output from function |

`alpha.gamma...` |
alpha.gamma and other parameters are hyperparameters for prior distribution, see the reference for the meaning of each parameters |

`OR_ylim` |
a numeric vector of two elements, used to set y-axis limit for plotting based on "odds ratio" |

**Model**:

Here the 3-stage hierarchical bayesian model was
used to model the probability of AEs. It is model 1b (Bayesian Logistic
Regression Model with Mixture Prior on Log-OR) in H. Amy Xia , Haijun Ma &
Bradley P. Carlin (2011) Bayesian Hierarchical Modeling for Detecting
SaBCIy Signals in Clinical Trials, Journal of Biopharmaceutical Statistics,
21:5, 1006-1029, DOI: 10.1080/10543406.2010.520181)

** Hier_history**:

This function takes formatted Binomial data and output Gibbs sample of the defined parametes. The output is a dataframe with each column represent one parameter and each row is the output from one sampling/one iteration. Diff, OR, gamma, and theta are the parameters recorded.

The result for Diff, OR, gamma, and theta are ordered by j and then by b. For example the result is like Diff.1.1, Diff.2.1, Diff.3.1, Diff.4.1 Diff.1.2, Diff.2.2, Diff.3.2 and so on

`sum_Hier`

This function takes the output from

`Hier_history`

and return the summary
statistics for each parameter recorded by `Hier_history`

. The summary function is applied on each column. `Hier`

This function takes the same input as

`Hier_history`

and calculate the summary statistics
for output from `Hier_history`

.
It outputs the summary statistics for each AE, combining with raw data, and also the Raw risk difference, Raw odds ratio
calculated from raw data.`Hiergetpi`

This function calculates pit (incidence of AE in treatment group) and pic (incidence of AE in control group) from the output of

`Hier_history`

The output is used for Loss function. `Hierplot`

first selects the top `ptnum`

(an integer) AE based on the selected statistic (either "odds ratio" or "risk difference").
Then it plots the mean, 2.5
color. `Hiertable`

creates a table for the detailed information for AE plotted in `Hierplot`

.
`Compareplot`

creates a plot to compare the mean risk difference (or odds ratio meadian) from posterior distribution with raw
risk difference (or row risk difference) from raw data. This function can be used for both bayesian models.
`Hier_history`

It returns a dataframe with each column represent one parameter
and each row is the output from one sampling/one iteraction (like the output of `coda.samples`

)

`sum_Hier`

It returns the summary statistics for each parameter recorded by `Hier_history`

.

`Hier`

It returns the summary statistics for each AE, combining with raw data.
The summary statistics including:
summary statistics for incidence rate difference (mean, 2.5% and 97.5% percentile);
summary statistics for odds ratio (mean, 2.5% and 97.5% percentile).
The other columns include SoC, PT, Nt, Nc, AEt, and AEc.
** Hiergetpi**:

This function calculates pit (incidence of AE in treatment group) and pic (incidence of AE in control group) from the output of

`Hier_history`

.H. Amy Xia , Haijun Ma & Bradley P. Carlin (2011) Bayesian Hierarchical Modeling for Detecting Safety Signals in Clinical Trials, Journal of Biopharmaceutical Statistics, 21:5, 1006-1029, DOI: 10.1080/10543406.2010.520181)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 | ```
## Not run:
data(ADAE)
data(ADSL)
AEdata<-preprocess(adsl=ADSL, adae=ADAE)
HIERRAW<-Hier_history(aedata=AEdata, n_burn=1000, n_iter=1000, thin=20, n_adapt=1000, n_chain=2)
HIERRAW2<-Hier_history(aedata=AEdata, n_burn=1000, n_iter=1000, thin=20, n_adapt=1000, n_chain=1)
# you can specify the hyperparameter as shown below
HIERRAW3<-Hier_history(aedata=AEdata, n_burn=1000, n_iter=1000, thin=20, n_adapt=1000, n_chain=1,
alpha.gamma=5, beta.gamma=1, alpha.theta=3, beta.theta=1, mu.gamma.0.0=0.1, tau.gamma.0.0=0.1, alpha.gamma.0.0=5,
beta.gamma.0.0=1, lambda.alpha=0.1, lambda.beta=0.1, mu.theta.0.0=0.1, tau.theta.0.0=0.1,alpha.theta.0.0=3, beta.theta.0.0=1)
HIERDATA<-Hier(aedata=AEdata, n_burn=1000, n_iter=1000, thin=20, n_adapt=1000, n_chain=2)
HIERPI<-Hiergetpi(aedata=AEdata, hierraw=HIERRAW)
Hierplot(HIERDATA)
Hierplot(HIERDATA, ptnum=15, param="odds ratio")
HIERTABLE<-Hiertable(HIERDATA)
HIERTABLE2<-Hiertable(HIERDATA, ptnum=15, param="odds ratio")
Compareplot(HIERDATA)
# user can use a very big number(bigger than total PTs in dataset) to plot out all the PTs
Compareplot(HIERDATA, ptnum=5000, param='odds ratio')
## End(Not run)
``` |

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.