clearanceEstimatorBayes: Bayesian Hierarchical Regression on Clearance Rates

In bhrcr: Bayesian Hierarchical Regression on Clearance Rates in the Presence of Lag and Tail Phases

Description

clearanceEstimatorBayes estimates the parasite clearance rates by using a Bayesian hierarchical model. Moreover, it provides regression analysis of clearance rates on given covariates.

Usage

clearanceEstimatorBayes(data, covariates = NULL, seed = 1234,
  detect.limit = 40, outlier.detect = TRUE, conf.level = 0.95,
  niteration = 1e+05, burnin = 500, thin = 50,
  filename = "output.csv")

Arguments

data	a data frame containing the profiles of patients. This data frame must contain id, time, and count columns, in that order. The first column represents the IDs of patients. The second and third columns contain parasite measurements (per microliter) in different times.
covariates	an optional data frame containing individual level covariates. This argument may be NULL, in which case estimation of clearance rates is of primary interest.
seed	a user-specified number used to initialize a pseudorandom number generator. The default value is set to be 1234 for reproducibility. If seed = NULL, then its value will be automatically obtained from the system clock.
detect.limit	detection limit of the parasite density in blood (parasites per microliter)
outlier.detect	indicator of whether or not to use Flegg's outlier detection method. outlier.detect = TRUE is recommended.
conf.level	required confidence level for reporting credible intervals
niteration	total number of simulations after the burn-in period
burnin	length of the burn-in period in the MCMC used in clearanceEstimatorBayes
thin	step size of the thinning process in the MCMC used in clearanceEstimatorBayes
filename	the name of the csv file used to store some output elements. This file contains id, clearance.mean, lag.median, and tail.median.

Details

This function estimates parasite clearance rates, along with the effect of covariates on them, by using the Bayesian hierarchical model which was introduced in Fogarty et al. (2015). A change point model is used on the log of the parasite densities to account for three potential phases: (1) a constant phase (the lag phase); (2) a phase with a linear decrease (decay phase); (3) another constant phase (the tail phase). Hence the estimation of the parasite clearance rate is only based on observations within the decay phase. The Bayesian approach allows us to treat the delineation between lag, decay, and tail phases within an individual's clearance profile as themselves being random variables, thus taking into account the additional uncertainty of boundaries between phases. Details are in Fogarty et al. (2015).

Value

The function summary (i.e., summary.bhrcr) can be used to obtain a summary of the results. clearanceEstimatorBayes returns an object of class "bhrcr" which is a list containing:

CALL	function call
clearance.post	posterior distributions of clearance rates
clearance.mean	mean values of the posterior distributions of clearance rates
clearance.median	median values of the posterior distributions of clearance rates
intercept.post	posterior distributions of the intercepts (alpha_i's) in the model
gamma.post	posterior distribution of gamma
gamma.post.thin	thinned posterior sample of gamma
gamma.mean	mean values of the posterior distribution of gamma
gamma.median	median values of the posterior distribution of gamma
gamma.CI	Credible intervals for gamma
halflifeslope.post	posterior distribution for the effect of covariates on log half-lives
halflifeslope.mean	mean values of the posterior distribution for the effect of covariates on log half-lives
halflifeslope.median	median values of the posterior distribution for the effect of covariates on log half-lives
halflifeslope.CI	Credible intervals for the effect of covariates on log half-lives
predicted.pce	PCE estimates
eta.post	posterior distribution of eta
changelag.post	posterior distributions of changetime between lag and decay phases
changetail.post	posterior distributions of changetime between decay and tail phases
lag.median	median values of the posterior distributions of changetime between lag and decay phases
tail.median	median values of the posterior distributions of changetime between decay and tail phases
var.epsilon.post	posterior variance of epsilon after simulation
var.error.post	thinned posterior sample of variance of epsilon
var.alpha.post	posterior distribution of variance of alpha
var.beta.post	posterior distribution of variance of beta
index	a list containing each patient's indices in the data
counts	Original parasite counts of all patients
counts.current	Parasite counts of all patients after sampling censored measurements
t.overall	measurement times of all patients
p.lag	posterior value of the priori probability of there being a lag phase after simulation
p.lag.thin	thinned posterior sample of the priori probability of there being a lag phase
p.tail	posterior value of the priori probability of there being a tail phase after simulation
p.tail.thin	thinned posterior sample of the priori probability of there being a tail phase
var1.post	posterior distribution of c^2
var2.post	posterior distribution of d^2
mu1.post	posterior distribution of a
mu2.post	posterior distribution of b
detect.limit	the detection limit of parasitemia
lag.post	posterior distributions of index of changetime between lag and decay phases
lag2.post	posterior distributions of index of changetime between decay and tail phases
theta.post	posterior distributions of log-parasite-count's mean in lag phase
theta2.post	posterior distributions of log-parasite-count's mean in tail phase
burnin	length of the burn-in period

Author(s)

Colin B. Fogarty <[e-mail]>, Saeed Sharifi-Malvajerdi <[e-mail]>, Feiyu Zhu <[e-mail]>

References

Flegg, J. A., Guerin, P. J., White, N. J., & Stepniewska, K. (2011). Standardizing the measurement of parasite clearance in falciparum malaria: the parasite clearance estimator. Malaria journal, 10(1), 339.

Fogarty, C. B., Fay, M. P., Flegg, J. A., Stepniewska, K., Fairhurst, R. M., & Small, D. S. (2015). Bayesian hierarchical regression on clearance rates in the presence of "lag" and "tail" phases with an application to malaria parasites. Biometrics, 71(3), 751-759.

Examples

data("pursat")
data("pursat_covariates")
out <- clearanceEstimatorBayes(data = pursat, covariates = pursat_covariates, outlier.detect = TRUE,
                               niteration = 200, burnin = 50, thin = 10)

bhrcr documentation built on May 1, 2019, 8:41 p.m.