si4bayesmeta-package {si4bayesmeta}R Documentation

Sensitivity and Identification for the Bayesian Meta-Analysis

Description

Sensitivity and identification estimates for all the parameters in the Bayesian normal-normal hierarchical model (NNHM) induced by a Half-Normal (HN) and a Half-Cauchy (HC) heterogeneity priors are produced by two d2BC_S_I_HN_raw and d2BC_S_I_HC_raw functions. Six scenarios are considered: target relative latent model complexity (RLMC) values fixed at 0.25, 0.5 and 0.75 with RLMC-adjusted HN and HC heterogeneity priors. Corresponding posterior estimates can be obtained by the functions raw_estimates_HN and raw_estimates_HC. The methodology implemented in this package has been developed in Roos et al. (2020).

Details

The DESCRIPTION file: This package was not yet installed at build time.
Index of help topics:

BC_normal_m1m0          Bhattacharyya coefficient between two normal
                        densities: the location modification
BC_normal_sd1sd0        Bhattacharyya coefficient between two normal
                        densities: the spread modification
BC_normal_total         Bhattacharyya coefficient between two normal
                        densities
HC_A0_2_Al_Au           Epsilon-grid for a HC heterogeneity prior
HN_A0_2_Al_Au           Epsilon-grid for a HN heterogeneity prior
Nmuh1N012h              Hellinger distance between normal and standard
                        normal densities
d2BC                    The second derivative of the Bhattacharyya
                        coefficient
d2BC_S_I_HC_raw         Sensitivity and Identification measure
                        expressed by the second derivative of the
                        Bhattacharyya coefficient for a Bayesian NNHM
                        with a HC heterogeneity prior
d2BC_S_I_HN_raw         Sensitivity and Identification measure
                        expressed by the second derivative of the
                        Bhattacharyya coefficient for a Bayesian NNHM
                        with a HN heterogeneity prior
descr_extract           Extraction of descriptive statistics from a
                        bayesmeta object
h2Nmuh1N01              Calibration of the Hellinger distance
h_choice_all            The choice of the numerical 'h' step value for
                        RLMC perturbations
median_rlmc             Median RLMC
pri_par_adjust_HC       RLMC-based adjustment of the scale parameter
                        for a HC distribution
pri_par_adjust_HN       RLMC-based adjustment of the scale parameter
                        for a HN distribution
raw_estimates_HC        Raw estimates for the Bayesian meta-analysis
                        expressed by a Bayesian NNHM with a HC
                        heterogeneity prior
raw_estimates_HN        Raw estimates for the Bayesian meta-analysis
                        expressed by a Bayesian NNHM with a HN
                        heterogeneity prior
rlmc_scaling_low        RLMC adjusted scaling of the study-specific
                        within-study standard deviation in the
                        likelihood: reduced impact of observations
rlmc_scaling_up         RLMC adjusted scaling of the study-specific
                        within-study standard deviation in the
                        likelihood: increased impact of observations
si4bayesmeta-package    Sensitivity and Identification for the Bayesian
                        Meta-Analysis
sigma_ref               Reference within-study standard deviation

An overview of how to use the package, including the most important functions

Author(s)

Sona Hunanyan [aut, cre], Malgorzata Roos [aut]

Maintainer: Sona Hunanyan <sona.hunanyan@uzh.ch>

References

Roos, M., Hunanyan, S., Bakka, H., Rue, H. (2020). Sensitivity and Identification quantification by a relative latent model complexity perturbation in the Bayesian meta-analysis. Manuscript submitted to Research Synthesis Methods.

Examples

# Acute Graft rejection (AGR) data analyzed in Friede et al. (2017),  
# Sect. 3.2, URL: https://doi.org/10.1002/bimj.201500236
# First study: experimental group: 14 cases out of 61; 
# control group: 15 cases out of 20 
# Second study: experimental group: 4 cases out of 36; 
# control group: 11 cases out of 36 
  rT<-c(14,4)
  nT<-c(61,36)
  rC<-c(15,11)
  nC<-c(20,36)
  
  df = data.frame(y = log((rT*(nC-rC))/(rC*(nT-rT))), # log-odds-ratio
                  sigma = sqrt(1/rT+1/(nT-rT)+1/rC+1/(nC-rC)), # SE(log-odds-ratio)
                  labels = c(1:2))



####---- Settings: hh, t_RLMC, Prior parameter values for mu ----####

# Set the normal prior parameter values for mu according to Roever 2018, bayesmeta
# Suitable for logOR and log-odds scale with the unit information argument
mu_mean<-0
mu_sd<-4


# set a global parameter for RMC (relative model complexity) perturbation
hh0<-0.0044 # new value obtained empirically by adjusting it for the epsilon-local 
# sensitivity approach
# with grid_epsilon=0.00354 for RLMC adjusted HN, HC heterogeneity priors with targets
# (0.25,0.5,0.75) in the case of NNHM 

# Fix the value of the target RMC
t_RLMC_025<-0.25

####---- HN: S-I measure (based on the second derivative d2BC/dRLMC at base/target RLMC0) 
#computation for a NNHM ----####

res_d2BC_S_I_rlmc025_HN_raw <- d2BC_S_I_HN_raw(df=df, hh=hh0, 
  rlmc=t_RLMC_025, mu_mean=mu_mean, mu_sd=mu_sd)
res_d2BC_S_I_rlmc025_HN_raw


####---- HC: S-I measure (based on the second derivative d2BC/dRLMC at base/target RLMC0) 
#computation for a NNHM ----####

res_d2BC_S_I_rlmc025_HC_raw <- d2BC_S_I_HC_raw(df=df, hh=hh0, 
  rlmc=t_RLMC_025, mu_mean=mu_mean, mu_sd=mu_sd)
res_d2BC_S_I_rlmc025_HC_raw


[Package si4bayesmeta version 0.1-1 Index]