pb.hybrid.binary {altmeta}

Hybrid Test for Publication Bias/Small-Study Effects in Meta-Analysis With Binary Outcomes

Description

Performs the hybrid test for publication bias/small-study effects introduced in Lin (2020), which synthesizes results from multiple popular publication bias tests, in a meta-analysis with binary outcomes.

Usage

pb.hybrid.binary(n00, n01, n10, n11, data, methods,
                 iter.resam = 1000, theo.pval = TRUE)

Arguments

n00

a numeric vector or the corresponding column name in the argument data, specifying the counts of non-events in treatment group 0 in the collected studies.

n01

a numeric vector or the corresponding column name in the argument data, specifying the counts of events in treatment group 0 in the collected studies.

n10

a numeric vector or the corresponding column name in the argument data, specifying the counts of non-events in treatment group 1 in the collected studies.

n11

a numeric vector or the corresponding column name in the argument data, specifying the counts of events in treatment group 1 in the collected studies.

data

an optional data frame containing the meta-analysis dataset. If data is specified, the previous arguments, n00, n01, n10, and n11, should be specified as their corresponding column names in data.

methods

a vector of character strings specifying the publication bias tests to be included in the hybrid test. They can be a subset of "rank" (Begg's rank test; see Begg and Mazumdar, 1994), "reg" (Egger's regression test under the fixed-effect setting; see Egger et al., 1997), "reg.het" (Egger's regression test accounting for additive heterogeneity), "skew" (the skewness-based test under the fixed-effect setting; see Lin and Chu, 2018), "skew.het" (the skewness-based test accounting for additive heterogeneity), "inv.sqrt.n" (the regression test based on sample sizes; see Tang and Liu, 2000), "trimfill" (the trim-and-fill method; see Duval and Tweedie, 2000), "n" (the regressoin test with sample sizes as the predictor; see Macaskill et al., 2001), "inv.n" (the regressoin test with the inverse of sample sizes as the predictor; see Peters et al., 2006), "as.rank" (the rank test based on the arcsine-transformed effect sizes; see Rucker et al., 2008), "as.reg" (the regression test based on the arcsine-transformed effect sizes under the fixed-effect setting), "as.reg.het" (the regression test based on the arcsine-transformed effect sizes accounting for additive heterogeneity), "smoothed" (the regression test based on the smoothed sample variances under the fixed-effect setting; see Jin et al., 2014), "smoothed.het" (the regression test based on the smoothed sample variances accounting for additive heterogeneity), "score" (the regression test based on the score function; see Harbord et al., 2006), and "count" (the test based on the hypergeometric distributions of event counts, designed for sparse data; see Schwarzer et al., 2007). The default is to include all aforementioned tests.

iter.resam

a positive integer specifying the number of resampling iterations for calculating the p-value of the hybrid test.

theo.pval

a logical value indicating whether additionally calculating the p-values of the tests specified in methods based on the test statistics' theoretical null distributions. Regardless of this argument, the resampling-based p-values are always produced by this function for the tests specified in methods.

Details

The hybrid test statistic is defined as the minimum p-value among the publication bias tests considered in the set specified by the argument methods. Note that the minimum p-value is no longer a genuine p-value because it cannot control the type I error rate. Its p-value needs to be calculated via the resampling approach. See more details in Lin (2020).

Value

This function returns a list containing p-values of the publication bias tests specified in methods as well as the hybrid test. Each element's name in this list has the format of pval.x, where x stands for the character string corresponding to a certain publication bias test, such as rank, reg, skew, etc. The hybrid test's p-value has the name pval.hybrid. If theo.pval = TRUE, additional elements of p-values of the tests in methods based on theorectical null distributions are included in the produced list; their names have the format of pval.x.theo. Another p-value of the hybrid test is also produced based on them; its corresponding element has the name pval.hybrid.theo.

References

Begg CB, Mazumdar M (1994). "Operating characteristics of a rank correlation test for publication bias." Biometrics, 50(4), 1088–1101. <doi: 10.2307/2533446>

Duval S, Tweedie R (2000). "A nonparametric ‘trim and fill’ method of accounting for publication bias in meta-analysis." Journal of the American Statistical Association, 95(449), 89–98. <doi: 10.1080/01621459.2000.10473905>

Egger M, Davey Smith G, Schneider M, Minder C (1997). "Bias in meta-analysis detected by a simple, graphical test." BMJ, 315(7109), 629–634. <doi: 10.1136/bmj.315.7109.629>

Harbord RM, Egger M, Sterne JAC (2006). "A modified test for small-study effects in meta-analyses of controlled trials with binary endpoints." Statistics in Medicine, 25(20), 3443–3457. <doi: 10.1002/sim.2380>

Jin Z-C, Wu C, Zhou X-H, He J (2014). "A modified regression method to test publication bias in meta-analyses with binary outcomes." BMC Medical Research Methodology, 14, 132. <doi: 10.1186/1471-2288-14-132>

Lin L (2020). "Hybrid test for publication bias in meta-analysis." Statistical Methods in Medical Research, 29(10), 2881–2899. <doi: 10.1177/0962280220910172>

Lin L, Chu H (2018). "Quantifying publication bias in meta-analysis." Biometrics, 74(3), 785–794. <doi: 10.1111/biom.12817>

Macaskill P, Walter SD, Irwig L (2001). "A comparison of methods to detect publication bias in meta-analysis." Statistics in Medicine, 20(4), 641–654. <doi: 10.1002/sim.698>

Peters JL, Sutton AJ, Jones DR, Abrams KR, Rushton L (2006). "Comparison of two methods to detect publication bias in meta-analysis." JAMA, 295(6), 676–680. <doi: 10.1001/jama.295.6.676>

Rucker G, Schwarzer G, Carpenter J (2008). "Arcsine test for publication bias in meta-analyses with binary outcomes." Statistics in Medicine, 27(5), 746–763. <doi: 10.1002/sim.2971>

Schwarzer G, Antes G, Schumacher M (2007). "A test for publication bias in meta-analysis with sparse binary data." Statistics in Medicine, 26(4), 721–733. <doi: 10.1002/sim.2588>

Tang J-L, Liu JLY (2000). "Misleading funnel plot for detection of bias in meta-analysis." Journal of Clinical Epidemiology, 53(5), 477–484. <doi: 10.1016/S0895-4356(99)00204-8>

Thompson SG, Sharp SJ (1999). "Explaining heterogeneity in meta-analysis: a comparison of methods." Statistics in Medicine, 18(20), 2693–2708. <doi: 10.1002/(SICI)1097-0258(19991030)18:20<2693::AID-SIM235>3.0.CO;2-V>

See Also

pb.bayesian.binary, pb.hybrid.generic

Examples

## meta-analysis of (log) odds ratios
data("dat.whiting")
# based on only 10 resampling iterations
set.seed(1234)
out.whiting <- pb.hybrid.binary(n00 = n00, n01 = n01,
  n10 = n10, n11 = n11, data = dat.whiting, iter.resam = 10)
out.whiting
# increases the number of resampling iterations to 10000,
#  taking longer time