Meta-Analysis - Science method

Meta-analysis is a statistical method used to combine and analyze the results of multiple independent studies on a particular topic, to provide a more comprehensive and reliable assessment of the evidence. In the context of treatment approaches for pain management, a meta-analysis can be conducted to synthesize the findings from various studies that have investigated the effectiveness of different interventions for alleviating pain.

What is Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA)?

I have retrieved a study that reports a logistic regression, the OR for the dichotomous outcome is 1.4 for the continuous variable ln(troponin) . This means that the Odds increase 0.4 every 2.7-fold in the troponin variable; but, is there any way of calculating the OR for a 1 unit increase in the Troponin variable?

I want to meta-analyze many logistic regressions, for which i need them to be in the same format (i.e some use the variable ln(troponin) and others (troponin). (no individual patient data is available)

I'm very interested in meta-analysis. If anyone has experience in conducting meta-analyses and is open to collaboration (Forestry Sector), I would love to work together on exploring this research method further. Please feel free to reach out!"

I have the OR of a logistic regresion that used the independent variable as continuous. I also have the ORs of 2x2 tables that dichotomized the variable (high if >0.1, low if < 0.1).

Is there anyway i can merge them for a meta-analysis. i.e. can the OR of the regression (OR for 1 unit increase) be converted to OR for High vs Low?

In a meta-analysis, there is just one study that reported the incident rate ratio, can I consider it as HR?

We are trying to conduct a meta-analysis. One of the studies is providing Nagelkerke's R2 and p-value (and sample sizes for two groups) but not the actual effect size. Is there a way to convert this data to an effect size that we can use in a meta-analysis? Thanks!

Q 1. Which model should be preferred (FE/RE) when conducting a meta-analysis for pooling prevalence?

Q2. In the RE model, why do all studies receive equal weight irrespective of sample size or confidence intervals?

Q3. When we use the FE model, all the studies receive weight according to their sample size or CI. But my question is, is it correct to use the Fixed Effect Model?

I am researching systematic reviews and meta-analyses of radon risk exposure from drinking water. The summary of the random effects models of 222Rn concentration is 25.01, and the 95% confidence intervals (CI) are 7.62 and 82.09) and displayed heterogeneity of (I2 = 100%; P < 0.001) with residual heterogeneity of (I2 = 62 %, p = 0.01). Can anyone interpret the result for me? Why I2 = 100% in this context? what is the significance of the residual heterogeneity?

I am pooling the effect size of a specific intervention. I am calculating the mean difference and 95% confidence interval scores. I came across an outcome measure expressed in Meters in some of these studies and expressed in Feet in one study.

Is it correct if I convert the mean and the standard deviation to Meters and I calculate the mean difference or should I calculate the standard mean difference as an alternative?

During writing a review, usually published articles are collected from the popular data source like PubMed, google scholar, Scopus etc.

My questions are

1. how we can confirm that all the articles that are published in a certain period (e.g.,2000 to 2020) are collected and considered in the sorting process(excluding and including criteria)?

2. When the articles are not in open access, then how can we minimize the challenges to understand the data for the metanalysis?

Hello friends, how many paper should include in a meta analysis??

How it is different from systematic review??

Is it possible to conduct a Meta-SEM (meta-analysis of structural equation models) using the SmartPLS 4 software?

I would very much like you all to answer this question of mine. I found it in a 2009 paper that reported sample sizes in the form of N=7-8, N=4-8, and so on. Meanwhile I can't find its supplementary material and raw data. Can tell me how to deal with this situation in meta-analysis?

Its research data provides the mean, SE.

Hello,

I am currently conducting a systematic review and meta-analysis in the field of epidemiology. However, I am facing challenges in selecting databases for literature collection due to limited access through my university.

The databases that I have personal or university access to include PubMed, EBSCO MEDLINE, Scopus, ScienceDirect, and Google Scholar.

I do not have access to databases such as Embase, Ovid MEDLINE, and Web of Science.

Since I cannot access some of the popular databases typically used for systematic reviews, I am seeking guidance on how to select and create a combination of three databases for my research.

Thank you in advance for your assistance.

Forest plot formation for meta-analysis.

Q 1. Which model should be preferred (FE/RE) when conducting a meta-analysis for pooling prevalence?

Q2. In the RE model, why do all studies receive equal weight irrespective of sample size or confidence intervals?

Q3. When we use the FE model, all the studies receive weight according to their sample size or CI. But my question is, is it correct to use the Fixed Effect Model?

Hi, I am conducting a meta-analysis on the overall survival and disease-free survival of TACE in Large HCC (dichotomous: 1-3-5 year survival). In inputting data from studies, which do we input as the event for a dichotomous survival meta-analysis like this? The number of deaths or the number of survivors?

thank you

much appreciated

Indah

Could someone explain to me why the p-value in the right column of the forest plot is different than the p-value in the test for effect in the subgroup?

I thought that these two p.values should be the same.

If anyone know to how to calculate the pooled prevalence rate in Meta analysis. In my dataset I have sample population and total population. Using metaprop command the output will come in proportion, how we get the output in percentage?

I am performing a meta analysis in which there are some studies which are RCTs and cohorts. There are two studies which are non-randomized controlled trials. Can I include those two studies in my systematic review and meta analysis? Is it fine to pool RCTs and non-RCTs?

hello! we are currently working on a systematic review containing 14 studies - however only 8 studies are similar enough to conduct the meta-analysis on and we are planning to do that. Can we still call it a meta-analysis then?

I am looking for your suggestion, options etc.

Dear colleagues

I have recently performed a meta-analysis project however one of the journal reviewers asked me to perform GRADE scoring for the results. Do you know any special software helping me?

I'm conducting a systematic review where I specified that I'll be taking studies from LMICs (Low-Middle income countries). During search, I came across a study which was conducted in high income country but the sample was from low income families, would this study be an eligible study?, please answer considering my LMICs criteria. I know superficially it seems that the study needs to be discarded but please think in terms of heterogeneity that we (explicitly or implicitly) assumed that heterogeneity will be coming from socio-economic status more so if a study already uses it, what is the issue in taking it.

I hope, I'm able to explain my query.

Help me please, I'm just learning to do meta-analysis using the R program. When I use summary measure with mean differences (MD) mode "metacont(Ne, Me, Se, Nc, Mc, Sc, data = data1, studlab = paste(Author), sm = "MD", common = T, random = T, hakn = T, prediction = T, subgroup = Sbg)" everything works normally.

However, when I change the summary measure to standardized mean differences (SMD) mode "metacont(Ne, Me, Se, Nc, Mc, Sc, data = data1, studlab = paste(Author), sm = "SMD", common = T , random = T, hakn = T, prediction = T, subgroup = Sbg)" an error message appears as follows "Error in (function (yi, vi, sei, weights, ai, bi, ci, di, n1i, n2i, x1i , : Fisher scoring algorithm did not converge. See 'help(rma)' for possible remedies."

Thank you

I want to develop an understanding for conducting Meta-Analysis studies in Psychology. how many studies to select, how to screen them, any tools which are used for the purpose. kindly explain with references how to conduct a meta-analysis

Greetings,

for my systematic review I have 23 research articles 21 of which I got from PubMed, CINAHL, OVID and WOS. I got 2 articles from British library, the explore further option and these articles are cited as peer-reviewed.

can I put British Library as a database in my Prisma Flow diagram. Or do I indicate that two articles were gotten from british library in my methodology? your answers are highly appreciated.

In some meta-analysis, the authors manually added some reference, but why were these studies not included in their initial database search.

I'm wondering if there is any difference between a meta-analysis in medical sciences and economic sciences. In addition, If you have any guidelines or research about how to conduct a meta-analysis in economic sciences, let me know, please.

We were planning about meta analysis and when doing literature review, we came across this condition:

In such condition, from what I have known, we have to use Standardized mean Difference as units are different. But real problem is in that first article with 2 different approach. How to decide which approach out of two, should I use?

Hi everyone, do you know any formulae to calculate the combined SD while knowing the M and SD of each group from the same population in the meta-analysis?

For example, in our meta-analysis, one study reported the Mean and SD for each facets of the Self-compassion scale in participants without reporting the overall score. Our study only wanted to investigate the overall score. We could calculate the Mean score easily by combining the mean score of both groups since they are the same population. However, we could not calculate the SD.

I know that the Cochrane Handbook of Systematic Reviews of Interventions offered formulae for Combining groups but it seems that this formulae can only used for group with different population.

J is a bias correction factor that used to remove the small-sample-size bias of the standardized differences of means.

Dear Colleagues,

We are currently conducting a qualitative meta-analysis and looking for qualitative research that is focused on transgender, non-binary, and gender diverse people’s experiences of gender identity development and how their gender functions to support coping with minority stress as well as increasing resilience and flourishing. We are looking specifically at articles that focus on participants in the United States. We are contacting scholars to ask if they have authored or are aware of studies that may meet our inclusion criteria, particularly any new or unpublished studies. Our team will screen the articles to determine if they meet our inclusion criteria. If you know of any studies that may be relevant to our qualitative meta-analysis, please contact or send them to Kelsey Kehoe at kelsey.kehoe001@umb.edu.

Thank you!

Heidi Levitt & Kelsey Kehoe

Do you know if I am able to conduct a comparison meta-analysis with correlation coefficient and sample size?

And if yes, Would you be able to recommend any software?

I am working on a meta-analysis where i have extracted the data directly from KM curves via web plot digitizer to calculate HRs for the studies that reported only KM Curves. One of the study has three curves and web plot digitizer would give me a total of three groups. I was wondering if it is appropriate to combine the data for two of those groups and calculate an overall HR for a meta-analysis? Keeping in view that it is a time-event data and there's censoring too. I tried using the method elaborated by Cochrane but it gave me a really wide confidence Interval.

Anyone having any lead of how to deal with this?

Can someone suggest me the best method for meta-analysis of proportions where there is a high heterogeneity. I am using random effects model to estimate the pooled proportion. I have done the pooled proportion and subgroup analysis with both logistic regression and dersimonian Laird method. Both yielded a varying result. Which one should I take into consideration?

I am planning on conducting a systematic review, which will focus on the impact randomised lifestyle interventions (RCTs) on intergenerational health.

Initially, I considered the ROBINS-I tool, which is often used for non-randomised cohort follow-up studies, but the non-randomised element does not align with the randomised nature of RCTs. My next thought was the RoB 2 tool, but it seems more geared towards evaluating the outcome of the trial (e.g. weight loss during trial) rather than the longer-term effects on the cohort.

I have looked online and have not found a specific tool tailored to assessing risk of bias in long-term follow-up cohort studies from RCTs. As it stands, I'm leaning towards the ROBINS-I tool, but any expertise would be appreciated.

Thanks.

I am trying to meta-analyze studies that report Overall Survival (OS) with studies that report only the Hazard Ratio for OS. How can I interconvert both variables to a comparable unit to execute the meta-analysis?

Thanks in advance!

The paper (therapy/intervention study) includes an assessment of quality and risk of bias, however does not using a GRADE approach? Would this be considered a weakness?

I just want to know if someone can provide me an example input data for microbial association network analysis using SPIEC- EASI. Also, if possible, an R script how to do this network analysis will be very helpful. Thank you very much in advance.

Update 1: I have received the official RevMan installer links via email, which are still up on Cochrane's website.

Update 2: The links referred to above no longer work, as Cochrane have taken down the files, so I have removed them to avoid confusion. Please see Ingrid Arevalo-Rodriguez's answer below for further details.

We are accepting colleague for a meta-analysis study regarding the accuracy of nuclear medicine procedures. The only important point having skills in this regard and publishing at least two articles related to PET, PET/CT methods. Thanks

How to run a proper meta-analysis when the selected studies only report the median and 95% Confidence Interval of a parameter of survival?

Should I treat the data in the same way as a mean, and apply the same meta-analysis techniques?

Some proposed methods such as Hozo and Wan address this question, but either IQR, Quartiles or full data distribution are required for estimating mean and SD, which trials do NOT provide. Please assume I DO NOT have the range, so estimating the mean and SD would not be possible also.

I have already emailed the authors for the full data, though.

Can anyone with experience in the topic help?

Thanks in advance!

In random-effects meta-analysis, I2 is very low (I2 = 1%). What does this mean? Is it inappropriate to do meta-analysis under such condition?

Besides, how to pool the HR with OR/RR in meta-analysis? Is there any possible conversion?

I am doing a meta-analysis in which I have extracted data on ORs as the effect measure for presence of a condition against absence (reference category) of the condition. I one particular study, 4 levels of the variable of interest are presented instead of 2 as in the rest of the studies. ORs were estimated for 3 levels against one level as the reference category. By definition, adding the first and last two categories will give the two levels I need as in the other studies. However, the paper did not report frequencies that resulted in the OR estimates so the only thing left is the OR and 95% CI. How can I combine the ORs of the first and last two categories to produce the presence and absence comparison I need?

In doing a meta-analysis using articles done with time-to-event analysis, I faced the problem that some authors reported the effect size (incidence density) with person-year observations, while others reported person-month observations. Some also reported the incidence density for only the exposed group (single arm). Others reported the rate ratio (the ratio of IR in the exposed group to IR in the unexposed group).

In most cases, we may obtain odds ratios or relative risks from the given beta coefficients if the DV is categorical to estimate the pooled effect sizes.

What if the DV is continuous and report beta coefficients with corresponding 95% CI or p-values?

For example, if we have chronotype score from MEQ (19-item), rMEQ (5-item) and CMS ?

I have 3 papers suitable for inclusion in my systematic review looking at high versus low platelet to red call ratio in TBI, and want advice as to whether I can combine their estimates of effect in a meta-analysis.

One RCT which provides an unadjusted odds ratio and adjusted odds ratio of 28-day mortality for two groups (one intervention (high ratio) and one control (low ratio), adjusted for differences in baseline characteristics).

One retrospective cohort study which provides absolute unadjusted 28-day mortality data for two groups (one exposed to high ratio, and another exposed to a low ratio). They have also performed a sophisticated propensity analysis to adjust for the few differences between the groups and multivariate cox regression to adjust for factors associated with mortality, and presented hazard ratios.

Finally, a post-hoc analysis of a RCT, which compares outcomes for participants grouped according to presence/absence of haemorrhagic shock (HS) and TBI. This generates 4 groups - neither HS nor TBI, HS only, TBI only and TBI + HS. I am interested in the latter two as they included patients with TBI. One group was exposed to a high ratio, whereas the other a lower ratio. The authors provided unadjusted mortality data for all groups, and they adjust for differences in admission characteristics, to generate odds ratio of 28-day mortality. However, they present these adjusted odds ratios of death at 28days for the HS only, TBI only and TBI + HS groups compared to the neither TBI nor HS group, not to each other.

I could analyse unadjusted mortality in a meta-analysis, but want to know if I can combine all or some of the adjusted outcome measures, I have described instead? Any help greatly appreciated.

I am currently preparing to conduct a meta-analysis that aims to incorporate correlations from various studies. I have observed that different studies employ different types of correlations, such as Pearson's correlation coefficient, Spearman's correlation coefficient, Heterotrait-Monotrait ratio of correlations, correlations of latent variables, and so on so forth. I am seeking guidance on whether it is appropriate to include all of these correlation types in my meta-analysis or if I should focus on a specific type. Any insights would be greatly appreciated.

Dear My fellow researchers,

I'm conducting a systematic review for a nutrigenomic topic, and found out, I can't proceed with meta-analysis due to the heterogeneity of the included studies (different study design, different reporting style, multi-arm study).

My question: is there any other statistical analysis or any method that will allow me to quantitatively investigate the effect of the intervention to the outcome? The only similar data between all those included studies are mean and confidence interval.

Thank you in advance. God bless.

Greetings fellow researchers,

I am embarking on a meta-analysis project to study the efficacy of a certain drug, let's call it "Drug A". My challenge lies in the limited and varied nature of the clinical trials available. So far, I have identified only three clinical trials conducted on Drug A. Two of these trials are double arm (comparing Drug A against a placebo), while one is a single arm trial.

I am grappling with how to best proceed with the analysis given the different study designs. I would like to harness as much data as possible from these trials to ensure my meta-analysis is robust. One approach I am considering is to extract single arm data from all three trials and analyze this, but I'm unsure if this approach is methodologically sound or if it introduces biases.

Does anyone have advice on whether this is an appropriate approach or if there are alternative strategies I should consider? Could I potentially combine the single-arm and two-arm trials in some way? And if so, what statistical methods or adjustments should I be aware of to correctly handle the different types of data?

Any insights or guidance would be highly appreciated.

Best regards.

Dr. Moiz Ahmed

for example: its a comparison study, and each time control group is same but test group is different.

Hi,

I have a set of studies that looked at the association of sex w.r.t to multiple variables. The majority of the studies reported regression variables such as beta, b values, t-stats, and standard errors. Is it possible to run a meta-analysis using any of the above-mentioned variables? If so, which software would be more meaningful to perform a meta-analysis? I did a wee bit of research and found out that Metafor in R would be the better choice to perform these kinds of meta-analyses.

Any help would be highly appreciated!

Thanks!

Hi,

I have a set of studies that looked at the association of sex w.r.t to multiple variables. The majority of the studies reported regression variables such as beta, b values, t-stats, and standard errors. Is it possible to run a meta-analysis using any of the above-mentioned variables? If so, which software would be more meaningful to perform a meta-analysis? I did a wee bit of research and found out that Metafor in R would be the better choice to perform these kinds of meta-analyses.

Any help would be highly appreciated!

Thanks!

I want to conduct a trial sequential analysis for my meta-analysis, which evaluates the impact of exposure on survival (EFS and OS). The effect size (Hazard Ratio) is extracted directly from each study included in the meta-analysis.

TSA software allows the evaluation of dichotomous outcomes, but it requires the specification of the total number of patients and the number of events, while I already have the calculated HR values. Does the TSA Software allow working with HR?

Let's talk about topic selection for meta analysis. How should a researcher approach for meta analysis?

Suppose there is a drug "X" which has shown to be effective for a disease, let's say for many years. Then is it a good idea to test effectiveness of some other drug "Y" over "X" through meta-analysis?

Four studies investigated Salmonella in chicken meat but one of them (Somda et al., 2016) assessed this pathogen in different chicken products (grilled chicken, fumed chicken, flamed chicken and chicken prepared around fire). We entered Somda et al 2026 as four entries (just treated as if it were different studies/articles) while conducting the meta-analysis. What is your opinion/advice on this approach? See snipped part of the forest plot for clarity (attached as file). Thank you for your expertise and time.

You are interested in pooling the effect of multiple observational studies that used ordinal measures as outcomes. What are your options?

Models, for biochar estimatimation and prediction for Africa use by using mata analysis.

I am performing a meta-analysis using the Comprehensive Meta-Analysis software. However, I am not sure how to enter DID estimates. Anyone here who knows which exact option it is in CMA's comprehensive list of ES's? Thanks!

I am a final year medical student and have already participated in systematic reviews, case reports and other types of articles. I am interested in expanding my experience as a researcher

If anyone is preparing a study and wants to give me the opportunity to participate, just contact me.

Hey there, I‘m currently working on a surivival meta-analysis and I use SPSS for my statistics. The problem is that I have a lot of data regarding mean OS / PFS, but no belonging standard deviations… does anyone have an idea how to solve this problem? I mean I can‘t make a meta-analysis with only the „mean“ data… thank you already very much!

How can i extract data from Kaplan-Meier curves for meta-analysis?

I am currently performing a meta-analysis, and I am using the The Restricted Maximum Likelihood ("REML") method to estimate tau2. However for one of my subgroups (33 studies) I have a tau2 of 0, but an I2 of 81%. I realise they are different measure of heterogeneity Tau2 (Distribution of true effect sizes about the mean​) and I2 (proportion of variance that is true (due to differences about effect size)), however I am struggling to understand how none of the heterogeneity can be accounted to true differences - what mathematically has gone on for this to be true. Hope this makes sense, thank you!

Hello Everyone,

I'm looking for a statistician to collaborate under a meta-analysis manuscript. Interested to join our research team? Please let me know on priv: michal.karbownik@umed.lodz.pl

I am a fourth year medical student in Brazil. I am interested in conducting an SR with meta analysis, but I still have little experience in any field. So i would like to participate in a research to learn in practice to in the future, make my own.

I have a particular interest in neurological surgery, but I am open to any ideas.

Hi to the all ResearchGate community,

I would appreciate some advices for conducting a meta-analysis on RECs!

All suggestions are accepted and helpful for me!

I am a 3rd medical student and I'm interested in working on a meta-analysis/ systematic review, however, I am a fledgling when it comes to this realm of research. I have published an LTE and I'm presently working on a cross-sectional study. If anyone would like some help with their meta-analysis/ systematic review, please reach out to me. I'm not particularly looking for authorship, I'd love to get some experience + direction.

I am not able to construct the funnel plot. As well as should we use even values or event rate value?

SD can be calculated using range, if yes which formula

max range-min range/4

0r

max range-min range/6

I was recently trying to figure out R vs. Revman for meta-analysis. Revman as a beginner, looked easy and user-friendly. But on the contrary to that, I have been reading R gives you much more flexibility. If I had to use R, which package should I use for metanalysis?

I am doing a meta-analysis and multiple studies report an unstandardized beta without an SD, but with a 95%CI. Now I've read that you can compute the SD using the following formulas:

SE = (upper CI - lower CI)/3.92

SD = SE / sqrt (1/n1 +1/n2)

Does anyone know if this is the correct methods? I feel like I am getting quite large SDs this way.

We performed a meta analysis recently and SAS popped out an i=squared of exactly zero. I know this is theoretically possible but it doesn't make sense to me as the outcomes in the different studies in the analysis did not have the exact same means and variances. Can anyone shed light on this?

Thanking you all in advance for your thoughts.

Gary