BMC Medical Research Methodology - Latest Comments
http://www.biomedcentral.com/bmcmedresmethodol//comments
The latest comments on all articles published by BMC Medical Research Methodology2015-03-04T15:00:34ZError in equation 1
http://www.biomedcentral.com/1471-2288/13/139/comments#2490701
<p>There is an error in equation 1 (sample size formula for control arm). The first set of brackets should be squared, i.e. (z_{1-\alpha_i}+z_{\omega_i})^2</p>Daniel Bratton2015-03-04T15:00:34Zhttp://www.biomedcentral.com/1471-2288/13/139Bratton et al.BMC Medical Research Methodology13139Thu Nov 14 00:00:00 GMT 2013WinBUGS code available
http://www.biomedcentral.com/1471-2288/14/88/comments#2360698
<p>We are pleased to inform that we have developed a WinBUGS code for fitting the "Latent Class Bivariate Model". The code will be provided upon request with email to the corresponding author.It is possible to estimate sensitivity and specificity (with credible intervals) within each latent class and obtain the relative classification probability for each study.</p>
<p> </p>Paolo Eusebi2014-11-17T17:32:55Zhttp://www.biomedcentral.com/1471-2288/14/88Eusebi et al.BMC Medical Research Methodology1488Fri Jul 11 00:00:00 BST 2014The meaning of MAR, is MAR(X) really MAR, and bias of complete case analysis
http://www.biomedcentral.com/1471-2288/12/184/comments#2340701
<p>The paper by Hardt, Herke and Leonhart is a very useful investigation into the performance of multiple imputation in small samples and how it varies with the inclusion of auxiliary variables. Their recommendations and conclusions are welcome, particularly as researchers increasingly face datasets with larger and larger p relative to n.</p>
<p>I have comments on the definition of MAR, whether the authors MAR(X) mechanism is really MAR, and on the bias of complete case analysis.</p>
<p>1. The meaning of MAR and terminology</p>
<p>On page 3, the authors write "No full MAR mechanism was applied here, because inreal data, finding a variable that completely explains theprocess of missingness is unlikely.", and then describe their MAR(Y) mechanism, where missingness is determined by the value of Y+c, where c~N(0,1), as 50% MCAR and 50% MAR, following an earlier paper by Allison. MAR does not mean that missingness follows a deterministic mechanism which is a function of observed values. Rather it means that conditional on the observed values, missingness in a variable no longer depends on the value of that variable. Thus I would label the authors' MAR(Y) mechanism as simply an MAR mechanism, rather than 50% MCAR and 50% MAR, which I fear might lead to confusion.</p>
<p>2. Is MAR(X) really MAR?</p>
<p>If I understand the authors' data generation scheme correctly, I do not believe their MAR(X) mechanism is truely MAR. I understand the scheme used to be that missingness in X1 and Z_a was determined by the value of d1=X2+c, and missingness in X2 depended on X1+c, where these steps were "carried out separately for each variable", which I presume means the c~N(0,1) variable was newly generated for determining missingness in X2. If I have understood the data generation correctly, the missing data are not MAR. For example, missingness in X2 depends on X1, but X1 is sometimes itself missing. More formally, in non-monotone missingness settings, MAR means that the probability of a pattern being realised only depends on the observed values in that particular pattern - see Robins and Gill, Statistics in Medicine, 1997 16:39-56. As a result, MI would be expected to give some bias under the authors' MAR(X) mechanism. I have tried simulating a large (n=100,000) dataset, with no auxiliary variables, making missingness in X1 and X2 according to what I understand the MAR(X) mechanism to be. MI then gave estimates with a slight downward bias for the coefficients of X1 and X2, so perhaps in this particular setup the problem (that the data are not really MAR) is not such a big deal.</p>
<p>3. Why is CC biased under MAR(X)?</p>
<p>As the authors note in their introduction, complete case analysis (CC) can be unbiased in certain situations - specifically when missingness is independent of the outcome variable, conditional on the covariates. In the MAR(X) mechanism, missingness is generated in a way which only depends on X1 and X2, and not on Y, so CC ought to be unbiased here. Yet Table 2 shows that CC is biased under MAR(X) - can the authors shed any light on why this is the case (in my large simulated dataset under MAR(X), I got no bias in CC).</p>Jonathan Bartlett2014-11-04T16:18:26Zhttp://www.biomedcentral.com/1471-2288/12/184Hardt et al.BMC Medical Research Methodology12184Wed Dec 05 00:00:00 GMT 2012Issues regarding simulation study and conclusions
http://www.biomedcentral.com/1471-2288/14/49/comments#2134698
The topic of covariate adjustment in randomised trials is an important one. However, I believe the simulation study and conclusions of Edgewale are flawed, and moreover my concerns mirror those of one of the paper's original reviewer's (Gillian Raab), which appear not to have been dealt with.<br /><br />The authors investigate the performance of three methods for analysing randomised trials with a single continuous outcome and a corresponding baseline measure. They focus on the issue of baseline imbalance, and conduct a simulation study where trial data are generated such that there is, on average, an imbalance at baseline between the two treatment groups. From their simulation results, the authors conclude that ANCOVA is unbiased, whereas analysis of change scores and an unadjusted comparison of outcomes (ignoring baseline) are biased.<br /><br />The problem is that the above approach for generating data does not correspond to how data arise in randomised trials. As the authors explain in the introduction, randomisation guarantees balance in expectation or on average, but not balance in any given study. But in the simulation study conducted, data are generated so that there is systematically (i.e. on average) imbalance at baseline. It is therefore unsurprising that in this case the ANOVA analysis (a t-test of outcomes ignoring baseline) is biased. All this demonstrates is that if one has a confounder, and one does not adjust for the confounder, estimates are biased. Put another way, the simulations show the following: if one performs repeated randomised trials where patients are more likely to be allocated to one of the treatment groups if they have higher baseline values, then a ANOVA or analysis of change scores is biased. But of course this is not how patients are allocated to groups in a simple randomised trial!<br /><br />In a randomised trial, provided that the randomisation procedure is not compromised, all three of the methods considered by the authors are unbiased, at least according to the statistical definition of bias of an estimator as the difference between the expectation of the estimator and the true parameter value. The methods do however differ in terms of precision/efficiency, and previously it has been shown that ANCOVA is superior in this regard to the other two methods. All of these results can be found in the following paper:<br /><br />Yang L, Tsiatis A (2001). Efficiency study of estimators for a treatment effect in a pretest-posttest trial. The American Statistician; 55: 314-321.Jonathan Bartlett2014-07-28T09:50:12Zhttp://www.biomedcentral.com/1471-2288/14/49Egbewale et al.BMC Medical Research Methodology1449Wed Apr 09 00:00:00 BST 2014Correction to funding acknowledgement for this paper
http://www.biomedcentral.com/1471-2288/14/21/comments#2123698
<p>This study was funded through the East Africa International epidemiological Databases to Evaluate AIDS (IeDEA) Consortium by the US National Institutes of Health - the Eunice Kennedy Shriver National Institute Of Child Health & Human Development (NICHD) and the National Institute Of Allergy And Infectious Diseases (NIAID). Grant award 3U01AI069911-06S2.</p>
<p> </p>Annabelle Gourlay2014-06-25T13:30:51Zhttp://www.biomedcentral.com/1471-2288/14/21Gourlay et al.BMC Medical Research Methodology1421Tue Feb 11 00:00:00 GMT 2014Software
http://www.biomedcentral.com/1471-2288/13/35/comments#2030698
The functions for assessing the heterogeneity and inconsistency in network meta-analysis, for producing a net heat plot and a network graph are now implemented in the R package netmeta with version 0.4-0 and are available from the standard <a class="external-link-new-window" href="http://www.cran.r-project.org/">CRAN repository</a>.Ulrike Krahn2014-03-28T16:31:24Zhttp://www.biomedcentral.com/1471-2288/13/35Krahn et al.BMC Medical Research Methodology1335Sat Mar 09 00:00:00 GMT 2013Ridge parameter
http://www.biomedcentral.com/1471-2288/12/184/comments#1279696
<p>The paper by Hardt, Herke and Leonhart is a welcome addition to the literature. It warns against simplistic approaches that throw just anything into the imputation model. While the imputation model is generally robust against including junk variables, the paper clearly demonstrates that we should not drive this to the edge. In general building the imputation model requires appropriate care. My personal experience is that it is not beneficial to include more than -say- 25 well-chosen variables into the imputation model.
<br/>
<br/>In their simulations the authors investigate cases where the number of variables specified in the imputation model exceeds the number of cases. Many programs break down in this case, but MICE will run because it uses ridge regression instead of the usual OLS estimate. The price for this increased computational stability is -as confirmed by Hardt et all - that the parameters estimates will be biased towards zero. It is therefore likely that some of the bias observed by the authors is not intrinsic to PMM, but rather due to the setting of the ridge parameter (the default value 1E-5 may be easily changed as mice(..., ridge = 1E-6)). Would a tighter ridge setting (e.g., 1E-6 or 1E-7) appreciably reduce the bias?
<br/>
<br/>The '1 out of 3 of the complete cases' rule is interesting and easily remembered. However, a complication in practice is that there are often no complete cases in real data, especially in merged datasets. What would the authors think of the slightly more liberal rule 'n/3 variables'?
<br/>
<br/>Stef van Buuren</p>Stef van Buuren2013-02-15T14:56:37Zhttp://www.biomedcentral.com/1471-2288/12/184Hardt et al.BMC Medical Research Methodology12184Wed Dec 05 00:00:00 GMT 2012erratum
http://www.biomedcentral.com/1471-2288/7/36/comments#1042696
<p>This commentary serves to point out that in the results section (p.7) of the manuscript, we (the authors) incorrectly described the calculation of specificity of VIA for the 3-class model. The corrected text (below) describes a specificity of 0.65 versus 0.57. The calculation correction is also appended.
<br/>
<br/>"Specificity was 0.645. In that model, class 1 (p = 0.682) and class 2 (p = 0.200) combine to form non-disease. Specificity was calculated as 0.323 (probability of VIA inflammation given class 2) plus 0.117 (probability of VIA normal given class 2) multiplied by the probability of class 2 (0.200), plus the analogous values for Class 1, i.e., 0.429 (probability of VIA inflammation given class 1) plus 0.276 (probability of VIA normal given class 1) multiplied by the probability of class 1 (0.682), the entire sum being divided by the probability of non-disease (0.682 + 0.200)."
<br/>
<br/>In Table 4, the specificity values for the 3-class LCA solution reference standard (#4) are likewise incorrect. These values were given as 0.568, 0.820, 0.568, 0.758, and 0.860 for, respectively, VIA Abnormal, CA; Pap LGSIL+; HPV >= 1.0 RLU; Colposcopy/Biopsy LGSIL+; and Colposcopy/Biopsy HGSIL+. The correct values are all higher and should be listed as 0.645, 0.930, 0.644, 0.860, and 0.975.
<br/>
<br/>In keeping with the above, the following text (pages 6-7) should be corrected as follows (0.645 instead of 0.568):
<br/>
<br/>"In this study, despite apparent imperfections in the reference standard, the conventionally-derived VIA results fell within the range of published data and were relatively consistent between with the 3-class LCA model (0.775 versus 0.744, respectively, for sensitivity and 0.639 and 0.645, respectively for specificity)."
<br/>
<br/>These corrections do not substantively alter the results (as noted above, the corrected LCA-derived specificity value is now almost the same as the conventionally derived value, similar to what was observed for VIA for sensitivity) and the corrections do not affect our conclusions.</p>John McGrath2012-08-02T14:47:45Zhttp://www.biomedcentral.com/1471-2288/7/36Gaffikin et al.BMC Medical Research Methodology736Tue Jul 31 01:34:30 BST 2007Correction
http://www.biomedcentral.com/1471-2288/11/96/comments#1027698
<p>Equation 12 contains errors. In the equation k_min should be replaced by k_1, k_max with k_K and k_K-j with k_K-j+1. Below equation 12, in the description of lambda_j, k_K-j should again be replaced with k_K-j+1. We would like to thank Dr. Finian Bannon at the N. Ireland Cancer Registry for pointing out these errors.</p>Therese Andersson2012-08-02T12:15:04Zhttp://www.biomedcentral.com/1471-2288/11/96Andersson et al.BMC Medical Research Methodology1196Wed Jun 22 00:00:00 BST 2011Update on methodology for NSW Ministy of Health telephone surveys
http://www.biomedcentral.com/1471-2288/11/159/comments#924696
<p>As referenced in the article landline random digit dialling (RDD) have been the method of choice for the telephone based population health survey conducted by the NSW Ministry of Health over the last decade. However because of the increase in mobile phone ownership the methology was modified to include mobile only persons using an overlapping duel-frame design in 2012. The methodology was developed in collaboration with the Centre for Statistical and Survey Methodology at the University of Wollongong. A full description of the methods and preliminary findings will be available soon.</p>Margo Barr2012-06-24T14:38:47Zhttp://www.biomedcentral.com/1471-2288/11/159Liu et al.BMC Medical Research Methodology11159Thu Nov 24 00:00:00 GMT 2011