cps.sw.count: Power simulations for cluster-randomized trials: Stepped...

In clusterPower: Power Calculations for Cluster-Randomized and Cluster-Randomized Crossover Trials

Description

This set of functions utilize iterative simulations to determine approximate power for stepped wedge cluster-randomized controlled trials. Users can modify a variety of parameters to suit the simulations to their desired experimental situation.

Usage

cps.sw.count(
  nsim = NULL,
  nsubjects = NULL,
  nclusters = NULL,
  c0 = NULL,
  c1 = NULL,
  steps = NULL,
  sigma_b_sq = NULL,
  alpha = 0.05,
  family = "poisson",
  analysis = "poisson",
  negBinomSize = 1,
  method = "glmm",
  quiet = FALSE,
  allSimData = FALSE,
  poorFitOverride = FALSE,
  lowPowerOverride = FALSE,
  timelimitOverride = TRUE,
  opt = "L-BFGS-B",
  seed = NULL
)

Arguments

nsim	Number of datasets to simulate; accepts integer (required).
nsubjects	Number of subjects per cluster; accepts either a scalar (equal cluster sizes) or a vector of length nclusters (user-defined size for each cluster) (required).
nclusters	Number of clusters; accepts non-negative integer scalar (required).
c0	Expected outcome count in arm 1. Accepts scalar (required).
c1	Expected outcome count in arm 2. Accepts scalar (required).
steps	Number of crossover steps; a baseline step (all clusters in arm 1) is assumed. Accepts positive scalar (indicating the total number of steps; clusters per step is obtained by nclusters / steps) or a vector of non-negative integers corresponding either to the number of clusters to be crossed over at each time point (e.g c(2,4,4,2); nclusters = 10) or the cumulative number of clusters crossed over by a given time point (e.g. c(2,4,8,10); nclusters = 10) (required).
sigma_b_sq	Between-cluster variance; accepts non-negative numeric scalar (indicating equal between-cluster variances for both arms) or a vector of length 2 specifying treatment-specific between-cluster variances (required).
alpha	Significance level. Default = 0.05.
family	Distribution from which responses are simulated. Accepts Poisson ('poisson') or negative binomial ('neg.binom') (required); default = 'poisson'
analysis	Family used for regression; currently only applicable for GLMM. Accepts c('poisson', 'neg.binom') (required); default = 'poisson'
negBinomSize	Only used when generating simulated data from the negative binomial (family = 'neg.binom'), this is the target for number of successful trials, or the dispersion parameter (the shape parameter of the gamma mixing distribution). Must be strictly positive but need not be integer. Defaults to 1.
method	Analytical method, either Generalized Linear Mixed Effects Model (GLMM) or Generalized Estimating Equation (GEE). Accepts c('glmm', 'gee') (required); default = 'glmm'.
quiet	When set to FALSE, displays simulation progress and estimated completion time; default is FALSE.
allSimData	Option to output list of all simulated datasets; default = FALSE.
poorFitOverride	Option to override stop() if more than 25% of fits fail to converge; default = FALSE.
lowPowerOverride	Option to override stop() if the power is less than 0.5 after the first 50 simulations and every ten simulations thereafter. On function execution stop, the actual power is printed in the stop message. Default = FALSE. When TRUE, this check is ignored and the calculated power is returned regardless of value.
timelimitOverride	Logical. When FALSE, stops execution if the estimated completion time is more than 2 minutes. Defaults to TRUE.
opt	Option to fit with a different optimizer (using the package optimx). Default is 'L-BFGS-B'.
seed	Option to set.seed. Default is NULL.

Details

Runs power simulations for stepped wedge cluster-randomized controlled trials with a count outcome. The stepped wedge trial design is a type of cross-over design in which clusters change treatments in waves. Initially all the clusters recieve the same standard treatment, and at the end of the trial all of the clusters will be recieving the treatment of interest. More than one cluster can change treatments in a wave, but the order in which clusters change treatments is randomly determined. The outcome of interest is assessed in each cluster during each wave.

Users must specify the desired number of simulations, number of subjects per cluster, number of clusters per arm, expected absolute difference between arms, between-cluster variance; significance level, distributional family for data generation, analytic method, progress updates, and simulated data set output may also be specified.

Value

A list with the following components

• Character string indicating total number of simulations and simulation type

• Number of simulations

• Data frame with columns "Power" (Estimated statistical power), "lower.95.ci" (Lower 95 "upper.95.ci" (Upper 95

• Analytic method used for power estimation

• Significance level

• Vector containing user-defined cluster sizes

• Vector containing user-defined number of clusters

• Data frame reporting ICC, within & between cluster variances for both arms at each time point

• Vector containing expected difference between groups based on user inputs

• Data frame containing mean response values for each arm at each time point

• Matrix showing cluster crossover at each time point

• Data frame with columns: "Estimate" (Estimate of treatment effect for a given simulation), "Std.err" (Standard error for treatment effect estimate), "Test.statistic" (z-value (for GLMM) or Wald statistic (for GEE)), "p.value", "sig.val" (Is p-value less than alpha?)

• If allSimData = TRUE, a list of data frames, each containing: "y" (Simulated response value), "trt" (Indicator for arm), "time.point" (Indicator for step; "t1" = time point 0) "clust" (Indicator for cluster), "period" (Indicator for at which step a cluster crosses over)

If nofit = T, a data frame of the simulated data sets, containing:

• "arm" (Indicator for treatment arm)

• "cluster" (Indicator for cluster)

• "y1" ... "yn" (Simulated response value for each of the nsim data sets).

Author(s)

Alexander R. Bogdan

Alexandria C. Sakrejda ([e-mail])

Ken Kleinman ([e-mail])

Examples

# Estimate power for a trial with 3 steps and 12 clusters in arm 1
# (often the standard-of-care or 'control' arm) at the initiation
#of the study. Those clusters have 15 subjects each, with sigma_b_sq = 1.
# We have estimated arm outcome counts of 4 and 5 in the first and second arms,
# respectively, and 100 simulated data sets analyzed by the GLMM method. Using seed = 123,
# the resulting power should be 0.85.

## Not run: 
count.sw.rct = cps.sw.count(nsim = 100, nsubjects = 15, nclusters = 12,
                              c0 = 4, c1 = 5, steps = 3, sigma_b_sq = 1,
                              alpha = 0.05, family = 'poisson', analysis = 'poisson',
                              method = 'glmm', seed = 123)

## End(Not run)

clusterPower documentation built on Jan. 29, 2021, 1:06 a.m.