Issues regarding simulation study and conclusions (Jonathan Bartlett, 28 July 2014)

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.

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...
read full comment

Comment on: Egbewale et al. BMC Medical Research Methodology, 14:49

Correction to funding acknowledgement for this paper (Annabelle Gourlay, 25 June 2014)

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.
read full comment

Comment on: Gourlay et al. BMC Medical Research Methodology, 14:21

Software (Ulrike Krahn, 28 March 2014)

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 CRAN repository.
read full comment

Comment on: Krahn et al. BMC Medical Research Methodology, 13:35

Ridge parameter (Stef van Buuren, 15 February 2013)

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...
read full comment

Comment on: Hardt et al. BMC Medical Research Methodology, 12:184

erratum (John McGrath, 02 August 2012)

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...
read full comment

Comment on: Gaffikin et al. BMC Medical Research Methodology, 7:36

Correction (Therese Andersson, 02 August 2012)

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.
read full comment

Comment on: Andersson et al. BMC Medical Research Methodology, 11:96

Update on methodology for NSW Ministy of Health telephone surveys (Margo Barr, 24 June 2012)

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.
read full comment

Comment on: Liu et al. BMC Medical Research Methodology, 11:159

Quantum Biophysical Semeiotics plays a central role in predicting intracranial findings on CT-scans. (Sergio Stagnaro, 02 December 2011)

Editors, in order to recognize, among individuals involved by suspected traumatic brain damage, those who really is suffering from such a disorder, physicians nowadays can utilise Quantum Biophysical Semeiotics (1-4). For instance, in health, light digital pressure, applied upon the closed eye, brings about gastric aspecific reflex (= both stomach fundus and body dilates) after a latency time of 8 sec. On the contrary, in presence of a pathological condition latency time results smaller, in relation to the seriousness of underlying disorder. An awful number of other signs allow doctor to make a correct differential diagnosis.

References

1) Stagnaro S., Percussione Ascoltata degli Attacchi Ischemici Transitori. Ruolo dei Potenziali...
read full comment

Further early references to sample sizes with fixed number of clusters (Michael Campbell, 12 September 2011)

Hemming et al (2011) give a useful review of sample size calculations with a fixed number of clusters. As they acknowledge their equations (17) to (19) are derived from Donner and Klar (200). I would like to point out the issue of fixed number of clusters was also discussed by Campbell (2000) that the authors’ equation (13) was first given as equation (2) in that paper and as equation (9.7) in Machin and Campbell (2005).

Mike Campbell

References Campbell MJ Cluster randomized trials in general (family) practice research. Statistical Methods in Medical Research 2000, 9; 81-94

Donner A and Klar N. Design and analysis of cluster randomized trials. London, Arnold 2000

Machin D and Campbell MJ Design of Studies for Medical...
read full comment

Comment on: Hemming et al. BMC Medical Research Methodology, 11:102

Two errors in this article? (Emma Friesen, 04 July 2011)

I can't find a definition for FORM in the article. it appears for the first time in the Method section on page 3 but does not have a definition. Is there one?

Also, there appears to be an error in Table 1, in the cell combining "Consistency" and "D Poor". It says 'Evidence is consistent' however it should read 'Evidence is inconsistent'.
read full comment

Comment on: Hillier et al. BMC Medical Research Methodology, 11:23

Standard measures of differences between outcome rates are problematic for identifying subgroup effects (James Scanlan, 08 June 2011)

White and Elbourne[1] address the way that interaction tests are affected by whether one compares relative changes in risk of an outcome, relative changes in risk of the opposite outcome, absolute changes in outcome rates, or odds ratios. They recommend a conservative approach to identifying interaction that involves examining the measure that is least likely to show a statistically significant subgroup effect.

The fact that, for example, when an intervention reduces one adverse outcome rate from 12.7% to 5.0% and another from 21.7% to 10.0%, whether an interaction test finds a statistically significant difference between the two changes may depend on what measure of change is employed suggests that something may be amiss with interaction tests generally.

Comment on: White et al. BMC Medical Research Methodology, 5:15

Competing Interests? (William Anderson, 31 May 2011)

I wonder if it is correct for the authors to state they have no competing interests given the strong association with Complementary and Alternative Medicine
read full comment

Comment on: Walach et al. BMC Medical Research Methodology, 6:29

Law & Kaldor method (Ian R White, 25 March 2011)

This article quotes me as saying (in reference 13) that the Law & Kaldor method is likely to be biased towards the null. But in fact reference 13 illustrates a far worse situation: data were simulated with no treatment effect, yet the Law & Kaldor method estimated an "adjusted" hazard ratio of 1.48 (95% CI 1.44 to 1.52). By contrast, methods based on a structural model have the highly desirable property of being unbiased under the null.
read full comment

Comment on: Morden et al. BMC Medical Research Methodology, 11:4

regression method of "Peter" (Rainer Beier, 15 March 2011)

It is always a very strong problem for analysing meta-analysis correctly additionally for heterogeneity. In my oppinion the regression model of Peter is the best way to detect publication bias for binary data because it depends on the sample size of the studies, where the sample size is weighted by a function of sample size as described in this article. All other regression models for detecting publication bias do not consider the problem of sample size. But it is very important for investigating meta analysis to consider the sample sizes of the studies included in a meta analysis where small studies with a greater effect could be included compared with larger studies included with a lower effect. Additionally these detecting models do not work if high heterogeneity is given, but in this...
read full comment

Comment on: Moreno et al. BMC Medical Research Methodology, 9:2

Correction to typographical error in figure 1 (Sarah Cockayne, 18 October 2010)

Since the paper has been published the authors have noted that there is a typographical error in figure one. The words 'intervention' and 'control' have accidentally been put in the wrong box. It should read allocated to control (n=250) allocated to intervention (n=788).
read full comment

Comment on: Cockayne et al. BMC Medical Research Methodology, 5:34

Comparisons of the sizes of health inequalities must be based on measures that are unaffected by the prevalence of an outcome (James Scanlan, 22 September 2010)

Jackson et al.[1] explore some complex issues concerning the possibility that differences in the administrative structures of populations may confound cross-country comparisons of geographic health inequalities. But the authors overlook a fundamental problem with standard comparisons of health inequalities that exists irrespective of the issues they raise – specifically, that the measures underlying those comparisons tend to be affected by the overall prevalence of an outcome. Most notably, for reasons inherent in the shapes of distributions of factors associated with experiencing an outcome, the rarer the outcome, the greater tends to be the relative difference in experiencing it and the smaller tends to be the relative difference in avoiding it.[2-7] Thus, other things being...
read full comment

Comment on: Jackson et al. BMC Medical Research Methodology, 10:74

Clarification - Regulatory cases (Fujian Song, 31 August 2010)

Thanks Dr. Eric Turner (Portland VA Medical Center) for sending me (Fujian Song) an email to explain US FDA regulatory process. I found it very helpful. After obtaining Eric's agreement, this clarification is pasted as below for other readers of the article:

"Fujian, I was just looking back at the paper you first-authored in BMC Research Methodology and noticing Figure 1. It shows regulatory authorities first receiving information on a clinical trial after inception and after the data have been analyzed and written up. That's *partly* true. In the US, the FDA gets involved at two points in time, once before inception, and once after trial completion.

The before-inception stage is called an IND application (which stands for Investigational New Drug). Before a drug...
read full comment

Comment on: Song et al. BMC Medical Research Methodology, 9:79

RCTs and Meta-analysis as Knowledge Sources (Vance W Berger, 03 August 2010)

Mickenautsch’s article [1] provides an interesting endorsement of systematic reviews as a method to unite analytic and synthetic knowledge acquisition. There are a few issues that merit further discussion, such as 1) clarifying the basis for the criticism of evidence-based medicine (EBM), 2) considering whether or not all trials should be grouped together as providing equally compelling information and 3) questioning the merits of meta-analysis itself.

After quoting sources asserting that EBM through randomized control trials (RCTs) provides the best evidence, Mickenautsch raises the criticism that EBM claims to have “unique access to absolute scientific truth” which “devalues and replaces knowledge sources of other types.” This criticism does not...
read full comment

Comment on: Mickenautsch BMC Medical Research Methodology, 10:53

proof of assumption done in the letter posted on 14th of April (Rainer Beier, 23 July 2010)

In addition to my posted letters on 14th of April the proof of the assumption is given as follows:

Let p1= P(E+/K+) and p2=P(E-/K+) be the being the well known probabilties of diseased patients being exposed and not exposed.Then the following 95% CI for the true value log(NNE) = log(1/(p1-p2)) is given as follows:

Addition to my letter "confidence interval for NNE (Rainer Beier, 23 July 2010)

In the following a simulation of the problem explained in my first letter to this article posted on 14th of April: The simulation is done with the statistical program R where k is the difference of the upper ends of my suggestion for calculating confifidence intervals and done in the article and v is the difference of the length of the two kinds of calculating confidence intervals with one million trials on the interval (0,1) as follows:

Please Download the Latest Version (Tom Trikalinos, 09 July 2010)

The installer included with this publication as supplementary material is dated; please download the latest version at: http://tuftscaes.org/meta_analyst/. Thank you.
read full comment

Comment on: Wallace et al. BMC Medical Research Methodology, 9:80

Clarification of methodology (Sue Walker, 21 June 2010)

We wish to clarify the selection of countries we included in our analysis of drowning deaths. To ensure statistical stability in the calculation of percentages, we included those countries with 20 or more reported unintentional drowning deaths, not more than 20 as indicated in the methodology.
read full comment

Comment on: Lu et al. BMC Medical Research Methodology, 10:30

confidence intervals for NNE (Rainer Beier, 14 April 2010)

The confidence intervals for NNE computed in this article were computed by inverting the lower and upper end confidence intervals for ARI. This confidence intervals are very wide. My suggestion is to compute the confidence intervals like these for the RR using (log(1/ARI^) as an estimate for log NNE. Then you have to compute the confidence interval for this log estimate and after this using the exponentialfunction to compute the upper and lower ends for the confidence intervals for NNE. E.g in table 5 we get a 95% CI as follows: 101-148. This confidence interval is much more closer to 122.37 instead of 69.55-508.69 . The estimate log(1/ARI^) could be handled like a continious function of two estimates.
read full comment

GREAT ( AND WORTHY) EFFORT (Vinícius Ynoe Moraes, 22 March 2010)

This is a very interesting manuscript, very creative and relevant. It is a fact that regarding to treatment issues we do not have the access to an up-to-date systematic review, in some cases. Therefore, this paper strenghtes a personal belief that we can rely on the largest trial, specially if it respects the treatment effect issues and (what i believe to be THE ISSUE) HAS AN EXPLICT AND CLEAR prevous power anaysis. Thanks for this pleasant reading and congratulations.
read full comment

Comment on: Glasziou et al. BMC Medical Research Methodology, 10:23

Previous IPD diagnostic and prognostic reviews (Bob Phillips, 17 March 2010)

Diagnostic tests and prognostic variables have been subject to previous IPD meta-analysis. Some of these include:

Royston, P., M.K. Parmar, and R. Sylvester, Construction and validation of a prognostic model across several studies, with an application in superficial bladder cancer. Stat Med, 2004. 23(6): p. 907-26.

Fibrinogen Studies Collaboration, Plasma Fibrinogen Level and the Risk of Major Cardiovascular Diseases and Nonvascular Mortality: An Individual Participant Meta-analysis. JAMA, 2005. 294(14): p. 1799-1809.

Trivella, M., et al., Microvessel density as a prognostic factor in non-small-cell lung carcinoma: a meta-analysis of individual patient data. Lancet Oncol, 2007. 8(6): p. 488-99.

Mant, J., et al., Systematic review and individual...
read full comment

Comment on: Broeze et al. BMC Medical Research Methodology, 9:22

RSS

## Latest comments

## Issues regarding simulation study and conclusions (Jonathan Bartlett, 28 July 2014)

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.

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... read full comment

Comment on:

Egbewale et al.BMC Medical Research Methodology,14:49## Correction to funding acknowledgement for this paper (Annabelle Gourlay, 25 June 2014)

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. read full comment

Comment on:

Gourlay et al.BMC Medical Research Methodology,14:21## Software (Ulrike Krahn, 28 March 2014)

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 CRAN repository. read full comment

Comment on:

Krahn et al.BMC Medical Research Methodology,13:35## Ridge parameter (Stef van Buuren, 15 February 2013)

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... read full comment

Comment on:

Hardt et al.BMC Medical Research Methodology,12:184## erratum (John McGrath, 02 August 2012)

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... read full comment

Comment on:

Gaffikin et al.BMC Medical Research Methodology,7:36## Correction (Therese Andersson, 02 August 2012)

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. read full comment

Comment on:

Andersson et al.BMC Medical Research Methodology,11:96## Update on methodology for NSW Ministy of Health telephone surveys (Margo Barr, 24 June 2012)

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. read full comment

Comment on:

Liu et al.BMC Medical Research Methodology,11:159## Quantum Biophysical Semeiotics plays a central role in predicting intracranial findings on CT-scans. (Sergio Stagnaro, 02 December 2011)

Editors,

in order to recognize, among individuals involved by suspected traumatic brain damage, those who really is suffering from such a disorder, physicians nowadays can utilise Quantum Biophysical Semeiotics (1-4).

For instance, in health, light digital pressure, applied upon the closed eye, brings about gastric aspecific reflex (= both stomach fundus and body dilates) after a latency time of 8 sec.

On the contrary, in presence of a pathological condition latency time results smaller, in relation to the seriousness of underlying disorder.

An awful number of other signs allow doctor to make a correct differential diagnosis.

References

1) Stagnaro S., Percussione Ascoltata degli Attacchi Ischemici Transitori. Ruolo dei Potenziali... read full comment

Comment on:

van der Ploeg et al.BMC Medical Research Methodology,11:143## Further early references to sample sizes with fixed number of clusters (Michael Campbell, 12 September 2011)

Hemming et al (2011) give a useful review of sample size calculations with a fixed number of clusters. As they acknowledge their equations (17) to (19) are derived from Donner and Klar (200). I would like to point out the issue of fixed number of clusters was also discussed by Campbell (2000) that the authors’ equation (13) was first given as equation (2) in that paper and as equation (9.7) in Machin and Campbell (2005).

Mike Campbell

References

Campbell MJ Cluster randomized trials in general (family) practice research. Statistical Methods in Medical Research 2000, 9; 81-94

Donner A and Klar N. Design and analysis of cluster randomized trials. London, Arnold 2000

Machin D and Campbell MJ Design of Studies for Medical... read full comment

Comment on:

Hemming et al.BMC Medical Research Methodology,11:102## Two errors in this article? (Emma Friesen, 04 July 2011)

I can't find a definition for FORM in the article. it appears for the first time in the Method section on page 3 but does not have a definition. Is there one?

Also, there appears to be an error in Table 1, in the cell combining "Consistency" and "D Poor". It says 'Evidence is consistent' however it should read 'Evidence is inconsistent'. read full comment

Comment on:

Hillier et al.BMC Medical Research Methodology,11:23## Standard measures of differences between outcome rates are problematic for identifying subgroup effects (James Scanlan, 08 June 2011)

White and Elbourne[1] address the way that interaction tests are affected by whether one compares relative changes in risk of an outcome, relative changes in risk of the opposite outcome, absolute changes in outcome rates, or odds ratios. They recommend a conservative approach to identifying interaction that involves examining the measure that is least likely to show a statistically significant subgroup effect.

The fact that, for example, when an intervention reduces one adverse outcome rate from 12.7% to 5.0% and another from 21.7% to 10.0%, whether an interaction test finds a statistically significant difference between the two changes may depend on what measure of change is employed suggests that something may be amiss with interaction tests generally.

More... read full comment

Comment on:

White et al.BMC Medical Research Methodology,5:15## Competing Interests? (William Anderson, 31 May 2011)

I wonder if it is correct for the authors to state they have no competing interests given the strong association with Complementary and Alternative Medicine read full comment

Comment on:

Walach et al.BMC Medical Research Methodology,6:29## Law & Kaldor method (Ian R White, 25 March 2011)

This article quotes me as saying (in reference 13) that the Law & Kaldor method is likely to be biased towards the null. But in fact reference 13 illustrates a far worse situation: data were simulated with no treatment effect, yet the Law & Kaldor method estimated an "adjusted" hazard ratio of 1.48 (95% CI 1.44 to 1.52). By contrast, methods based on a structural model have the highly desirable property of being unbiased under the null. read full comment

Comment on:

Morden et al.BMC Medical Research Methodology,11:4## regression method of "Peter" (Rainer Beier, 15 March 2011)

It is always a very strong problem for analysing meta-analysis correctly additionally for heterogeneity. In my oppinion the regression model of Peter is the best way to detect publication bias for binary data because it depends on the sample size of the studies, where the sample size is weighted by a function of sample size as described in this article. All other regression models for detecting publication bias do not consider the problem of sample size. But it is very important for investigating meta analysis to consider the sample sizes of the studies included in a meta analysis where small studies with a greater effect could be included compared with larger studies included with a lower effect. Additionally these detecting models do not work if high heterogeneity is given, but in this... read full comment

Comment on:

Moreno et al.BMC Medical Research Methodology,9:2## Correction to typographical error in figure 1 (Sarah Cockayne, 18 October 2010)

Since the paper has been published the authors have noted that there is a typographical error in figure one. The words 'intervention' and 'control' have accidentally been put in the wrong box. It should read allocated to control (n=250) allocated to intervention (n=788). read full comment

Comment on:

Cockayne et al.BMC Medical Research Methodology,5:34## Comparisons of the sizes of health inequalities must be based on measures that are unaffected by the prevalence of an outcome (James Scanlan, 22 September 2010)

Jackson et al.[1] explore some complex issues concerning the possibility that differences in the administrative structures of populations may confound cross-country comparisons of geographic health inequalities. But the authors overlook a fundamental problem with standard comparisons of health inequalities that exists irrespective of the issues they raise – specifically, that the measures underlying those comparisons tend to be affected by the overall prevalence of an outcome. Most notably, for reasons inherent in the shapes of distributions of factors associated with experiencing an outcome, the rarer the outcome, the greater tends to be the relative difference in experiencing it and the smaller tends to be the relative difference in avoiding it.[2-7] Thus, other things being... read full comment

Comment on:

Jackson et al.BMC Medical Research Methodology,10:74## Clarification - Regulatory cases (Fujian Song, 31 August 2010)

Thanks Dr. Eric Turner (Portland VA Medical Center) for sending me (Fujian Song) an email to explain US FDA regulatory process. I found it very helpful. After obtaining Eric's agreement, this clarification is pasted as below for other readers of the article:

"Fujian, I was just looking back at the paper you first-authored in BMC Research Methodology and noticing Figure 1. It shows regulatory authorities first receiving information on a clinical trial after inception and after the data have been analyzed and written up. That's *partly* true. In the US, the FDA gets involved at two points in time, once before inception, and once after trial completion.

The before-inception stage is called an IND application (which stands for Investigational New Drug). Before a drug... read full comment

Comment on:

Song et al.BMC Medical Research Methodology,9:79## RCTs and Meta-analysis as Knowledge Sources (Vance W Berger, 03 August 2010)

Mickenautsch’s article [1] provides an interesting endorsement of systematic reviews as a method to unite analytic and synthetic knowledge acquisition. There are a few issues that merit further discussion, such as 1) clarifying the basis for the criticism of evidence-based medicine (EBM), 2) considering whether or not all trials should be grouped together as providing equally compelling information and 3) questioning the merits of meta-analysis itself.

After quoting sources asserting that EBM through randomized control trials (RCTs) provides the best evidence, Mickenautsch raises the criticism that EBM claims to have “unique access to absolute scientific truth” which “devalues and replaces knowledge sources of other types.” This criticism does not... read full comment

Comment on:

MickenautschBMC Medical Research Methodology,10:53## proof of assumption done in the letter posted on 14th of April (Rainer Beier, 23 July 2010)

In addition to my posted letters on 14th of April the proof of the assumption is given as follows:

Let p1= P(E+/K+) and p2=P(E-/K+) be the being the well known probabilties of diseased patients being exposed and not exposed.Then the following 95% CI for the true value log(NNE) = log(1/(p1-p2)) is given as follows:

Log(NNE) 1.96*sqrt(((Var(p^1)+Var(p^2))/(p1-p2)2)

So the 95% confidence interval for NNE is now given as follows:

NNE*exp( 1.96*sqrt(((Var(p^1)+Var(p^2))/(p1-p2)2))

The estimate of NNE was considered as a continious function of two variables f(p1^,p2^) in IR2

Now we have to compare the upper and lower limits of the confidence interval above with the well known confidence interval... read full comment

Comment on:

Hildebrandt et al.BMC Medical Research Methodology,6:32## Addition to my letter "confidence interval for NNE (Rainer Beier, 23 July 2010)

In the following a simulation of the problem explained in my first letter to this article posted on 14th of April:

The simulation is done with the statistical program R where k is the difference of the upper ends of my suggestion for calculating confifidence intervals and done in the article and v is the difference of the length of the two kinds of calculating confidence intervals with one million trials on the interval (0,1) as follows:

> x=runif(1000000,0,1)

> y=runif(1000000,0,1)

> n=1000

> m=1000

> a=x

> b=y

> f=x*(1-x)/n

> g=y*(1-y)/m

> j=(1/(a-b))*exp(-1.96*(sqrt((f+g)/(a-b)^2)))-1/((a-b)+1.96*sqrt(f+g))

> k=(1/(a-b))*exp(1.96*(sqrt((f+g)/(a-b)^2)))-1/((a-b)-1.96*sqrt(f+g))

> v=k-j... read full comment

Comment on:

Hildebrandt et al.BMC Medical Research Methodology,6:32## Please Download the Latest Version (Tom Trikalinos, 09 July 2010)

The installer included with this publication as supplementary material is dated; please download the latest version at: http://tuftscaes.org/meta_analyst/. Thank you. read full comment

Comment on:

Wallace et al.BMC Medical Research Methodology,9:80## Clarification of methodology (Sue Walker, 21 June 2010)

We wish to clarify the selection of countries we included in our analysis of drowning deaths. To ensure statistical stability in the calculation of percentages, we included those countries with

20 or more reported unintentional drowning deaths, not more than 20 as indicated in the methodology. read full commentComment on:

Lu et al.BMC Medical Research Methodology,10:30## confidence intervals for NNE (Rainer Beier, 14 April 2010)

The confidence intervals for NNE computed in this article were computed by inverting the lower and upper end confidence intervals for ARI. This confidence intervals are very wide. My suggestion is to compute the confidence intervals like these for the RR using (log(1/ARI^) as an estimate for log NNE. Then you have to compute the confidence interval for this log estimate and after this using the exponentialfunction to compute the upper and lower ends for the confidence intervals for NNE. E.g in table 5 we get a 95% CI as follows: 101-148. This confidence interval is much more closer to 122.37 instead of 69.55-508.69 . The estimate log(1/ARI^) could be handled like a continious function of two estimates. read full comment

Comment on:

Hildebrandt et al.BMC Medical Research Methodology,6:32## GREAT ( AND WORTHY) EFFORT (Vinícius Ynoe Moraes, 22 March 2010)

This is a very interesting manuscript, very creative and relevant. It is a fact that regarding to treatment issues we do not have the access to an up-to-date systematic review, in some cases. Therefore, this paper strenghtes a personal belief that we can rely on the largest trial, specially if it respects the treatment effect issues and (what i believe to be THE ISSUE) HAS AN EXPLICT AND CLEAR prevous power anaysis. Thanks for this pleasant reading and congratulations. read full comment

Comment on:

Glasziou et al.BMC Medical Research Methodology,10:23## Previous IPD diagnostic and prognostic reviews (Bob Phillips, 17 March 2010)

Diagnostic tests and prognostic variables have been subject to previous IPD meta-analysis. Some of these include:

Royston, P., M.K. Parmar, and R. Sylvester, Construction and validation of a prognostic model across several studies, with an application in superficial bladder cancer. Stat Med, 2004. 23(6): p. 907-26.

Fibrinogen Studies Collaboration, Plasma Fibrinogen Level and the Risk of Major Cardiovascular Diseases and Nonvascular Mortality: An Individual Participant Meta-analysis. JAMA, 2005. 294(14): p. 1799-1809.

Trivella, M., et al., Microvessel density as a prognostic factor in non-small-cell lung carcinoma: a meta-analysis of individual patient data. Lancet Oncol, 2007. 8(6): p. 488-99.

Mant, J., et al., Systematic review and individual... read full comment

Comment on:

Broeze et al.BMC Medical Research Methodology,9:22