Science topic

Randomized Control Trials - Science topic

Explore the latest questions and answers in Randomized Control Trials, and find Randomized Control Trials experts.
Questions related to Randomized Control Trials
  • asked a question related to Randomized Control Trials
Question
4 answers
Hi, I would like to calculate the sample size for a 3 armed randomised control trial to study the effect of a particular intervention in a clinical population. The primary outcome is a continuous variable and the effect size is of 0.34 (obtained froma recent meta-analysis). I will have pre- and post- measures for participants in each of the three study arms. Could someone kindly guide me with this sample size calculation please?
Relevant answer
Answer
Group 1 will be usual treatment only, group 2 (intervention) will be usual treatment + intervention being investigated, group 3 (control) will be usual treatment and sequential calls from a healthcare provider (To make sure any change in outcome is secondary to the intervention not to the therapeutic alliance). all groups will receive treatment as usual for obvious ethical reasons
  • asked a question related to Randomized Control Trials
Question
3 answers
I understand meta-analysis of RCTs has a level I evidence and case control studies have a level III. But what is the level of evidence from meta-analysis of well-conducted case-control studies; is it I or III or otherwise? Many thanks
Relevant answer
Answer
Thanks Suraj Kapoor. I do not see this on the table. If individual case-control studies are level VI, ot would not make sense for systematic review and meta-analysis of a group of these to be level V
  • asked a question related to Randomized Control Trials
Question
8 answers
I am having trouble with interpretation of a prospective superiority randomised controlled trial. Study characteristics:
Study design:
  • Alpha value = 0.05
  • 1-beta = 0.8
  • Predicted effect size =0.15 [P1 = 0.65, P2 = 0.5]
  • Numbers needed = 492
Observed results:
  • Effect size = 0.09 [P1 = 0.50, P2 = 0.41]
  • Numbers enrolled = 502
  • statistically significant P value -> P = 0.03
  • null hypothesis is rejected.
My impression is that the trial underestimated the effect size. If I use the observed results to calculate a sample size that would take into the effect size (same alpha and beta), I get a numbers needed of 900+ patients. This would suggest that the trial is actually under powered. And potentially, the result we see is a false positive result.
Relevant answer
Answer
How are the sample sizes calculated? I assumed the test is about equal proportions and I get quite different numbers.
But independent of that: the effect size (or proportions) needed are estimates. They are never correct, and they can also be set to some (clinically, economically) relevant values that does not need to correspond to some actual (possibly estimated) effect size.
And above all there is chance involved. Having 0.8 power means that - even if the guessed effect sizes are absolutely correct! - you still have a 20% chance to end up with p>alpha, including all values up to 1. And the 80% chance of having p<alpha includes all values down to 0.
  • asked a question related to Randomized Control Trials
Question
3 answers
What is the relevance of superiority margin/ clinically meaningful margin / least relevant margin in randomized control trial for sample size calculation?
Relevant answer
Answer
Suraj Kapoor A superiority study seeks to distinguish between treatments. The first stage in the analysis is generally a statistical significance test to see whether the trial findings are compatible with the premise that there is no difference in the clinical impact of the two therapies.
While Cohen defined big, medium, and small effects as d=0.8, 0.5, and 0.2, respectively, an FDA rule of thumb states that an impact is large if it is greater than 0.8, small if it is less than 0.5, and moderate if it lies between those values.
  • asked a question related to Randomized Control Trials
Question
8 answers
I am doing my first systematic review. I have included two different study types; cohort studies and RCTs.
Regarding the quality assessment. What is the best thing to do?
Should i assess the cohorts using e.g. the Newcastle-Ottawa scale and the RCTs using e.g. Cochrane risk-of-bias?
Or is there a combined quality assessment tool for use in these cases where different study types are included?
Is there any gold standard for this?
Thanks in advance
Relevant answer
  • asked a question related to Randomized Control Trials
Question
3 answers
Falls prevention in ver 2 years: a randomized controlled trial in 80 years older women
has what kind of sampling?
purposive?
Relevant answer
Answer
For sampling you can use Stratified Sampling Method, if it will convinent to you
(I think it'll better choice for your study)
For random allocation you can use computer-generated table with randomly permuted blocks or using any others app for randomization
  • asked a question related to Randomized Control Trials
Question
6 answers
Blinding is done in RCT to minimize biases. Is it mandatory to do Blinding in RCT?
Relevant answer
Answer
I agree with Michael Dan Arvig on this. Although it´s desirable, blinding of patients and/or health care providers/investigators is not always feasible. Check this non-blinded RCT conducted by our group and published in JAMA
But always have in mind that allocation concealment, sometimes also known as "blinding of randomization" is mandatory in RCT.
Best regards,
Erica
  • asked a question related to Randomized Control Trials
Question
2 answers
I am conducting a systematic review about a newly developed psychological intervention for children, but all studies available are of exploratory/feasibility nature and I am not sure how to evaluate/analyse their results.
I am wondering if anyone is aware of guidelines or articles regarding the decision process of going from a pilot/feasibility stage to full RCT?
I am sure this is a relevant question in many departments, deciding if the results from exploratory studies justify going further with expensive RCTs.
I have been looking in to Cochrane Library etc. to see how they evaluate evidence, but what I can find is mostly regarding making clinical guidelines based on evidence availible (RCTs and other sources of evidence).
If anyone knows anything about this process, guidelines for going from exploratory to RCTs, I would be immensely grateful since I feel a bit stuck on this question (how to analyse the results from feasibility/exploratory intervention studies, what criteria can be used for deciding to go ahead or not with RCTs etc..)
Relevant answer
Answer
Unfortunately, there is (as is so often the case in scientific research) no pre-determined answer for your question. You will have to consider, decide and substantiate the criteria for inclusion in your study. You might find this specific part of the Cochrane handbook you already consulted specifically helpful in the process:
  • asked a question related to Randomized Control Trials
Question
9 answers
I'm currently working on a systematic review including more than 150 studies. However, a large amount of experimental studies haven't reported sufficiently means and SDs for the outcome measures. Is there any alternative way to include them in the meta-analysis?
Relevant answer
  • asked a question related to Randomized Control Trials
Question
3 answers
In one study, a person in each treatment group converted from MCI to dementia. The study eliminated them from the analysis, using them as statistical outliers. I know that the two analysis for RCTs are per protocol and intention to treat. It did not use an ITT analysis. It was a pilot study with a control group. Subjects were randomized. Are there other analyses suitable for this case?
Relevant answer
Answer
They did not include all the participants from baseline.
  • asked a question related to Randomized Control Trials
Question
2 answers
Hi,
Can we conduct network meta-analysis for cohorts and case-control studies or only for RCTs? is there a limit to the number of interventions? and what is the best software that does not need R code?
Thank you
Relevant answer
Answer
The researcher and research question dictate the study eligibility criteria used in the systematic review for a network meta-analysis as well as for a traditional pairwise meta-analysis. The researcher must be able to defend those criteria (e.g., inclusion of studies with a given study research design).
  • asked a question related to Randomized Control Trials
Question
3 answers
Sometimes, due to limited/very few number of RCTs, researcher needs to combine data with Non-RCTs study for meta-analysis. But, combining dat from both types of studies may increase the heterogeneity and mean difference.
How one can plan data analysis method to combine RCTs and observational studies in the meta-analysis? What model will be used to combine data of RCTs and observational study?
Relevant answer
It depends on the outcome you are evaluating; for example if you are interested on performing a metaanalysis for prevalence, and you include RCT and observational studies to evaluate the prevalence of certain disease among a specific population, the fact of the different study methodology may not affect the overall effect. What I recommend you is to consider the outcome you are evaluating and assess if the result may change between RCT and observational studies due to bias. Also I recommend you to use I2 and Cochrane Q to evaluate heterogeneity to select the appropriate model.
  • asked a question related to Randomized Control Trials
Question
2 answers
I want to understand the benefits of conducting a quasi-experimental study as opposed to a randomized control trial (which is considered the gold standard) in a population-based/ epidemiological/ or health systems-based research project.
Relevant answer
Answer
Hi Zauraiz Lone
Thanks for the links, I will check them out.
  • asked a question related to Randomized Control Trials
Question
3 answers
I need published papers about it to include in a Cochrane sistematic review.
Relevant answer
Answer
Yes, all of them are RCT
  • asked a question related to Randomized Control Trials
Question
7 answers
I noticed a new trend in recent landmark RCTs that the pharmaceutical company (which is the one sponsoring the trial for the benefit hypothesis of its drug), was involved heavily in data collection, entery& interpretation even the final writing of the article e.g VERIFY trial for Vldagliptin& PIONEER project for Semaglutide& maybe so many others.
My question is about is it enough to be written transparently about that to refute the high possibility of conflict of interrest& till which limit we may accept that?
Relevant answer
Answer
It depends. If the pharma company commissioned the research you don't have much choice, as in the case you cite. Most journals now require a 'conflict of interest' statement so it would be clear to readers that the company was involved. On the plus side, having company experts involved could strengthen the methodology and they may have insights into the data. On the negative side, the very real possibility of bias.
  • asked a question related to Randomized Control Trials
Question
3 answers
Hi unfortunetly in my study there are only 2 RCTs published to demonstrate any reliable effect with a surgical procedure on a pathology (osteoarthritis).
The remaining are case series, which measure patient's MEAN outcome scores at 1year, 1-5 years , >5 years. Some don't publish SD. Others publish just median scores. The series all have the same pathology (osteoarthritis), but do not necessarily have the same patient characteristics - typical of case series and units wanting to present their results.
Can I pool these Mean patient outcome scores together at 1 year / 1-5 years / >5 years and calculate the total mean and present them in a forest plot?
Does it statistically make sense or am I comparing apples and oranges? Do I just calculate the average or are there any tests you would recommend?
I understand that is the limitation that this is poor literature and hard to come to a conculsion, but that is the point I am trying to highlight that we can't keep publishing small case series but need large registry data which I will be working on.
Relevant answer
Answer
I assumed included studies met your predetermined inclusion/exclusion criteria based on protocol. If so, in this case , I would contact author requesting some missing information that I need for meta analyais.
  • asked a question related to Randomized Control Trials
Question
4 answers
I can understand the importance of having an outcome assessor independent and blinded of intervention for a Randomised Controlled Trial (RCT), but when conducting a Case Study investigating a new Clinical Health intervention is it also important to NOT allow the principal researcher to carry out baseline and post-intervention outcome scales/measures?
My understanding is that the principal objective of a Case Study is to establish the feasibility of a new intervention rather than effectiveness and more importantly ensure that the participant does not experience any adverse effects. Because of this, the significance of any non-feasibility scales are minimal so for this reason would it be a poor design for the principal researcher to undertake the outcome scales/measures?
Thank you for your advice regarding this question!
Ken
Relevant answer
Answer
To a degree this will be determined by the tools you are using to evaluate outcome. If the tools have high inter-observer reliability then it would be ideal to have an independent observer
  • asked a question related to Randomized Control Trials
Question
8 answers
I'm currently exploring the opportunity of conducting an NMA of observational studies only. very rarely NMA is done using observational studies as the transitivity assumption is unlikely to be fulfilled because these studies will have different protocols, methods of patient selection and settings ..etc.
I'm aware of the methods that are used to incorporate non randomized studies into a network of RCTs and the existence of many approaches for that such as regression adjustment, propensity score matching techniques ..etc. but I'm not sure if those techniques can be implemented in an NMA of observational studies only.
So, do you think that an NMA is applicable in the context of observational studies only? and if so, how can we fulfil the transitivity assumption? can proper statistical tests address that?
Relevant answer
Answer
The assumptions can absolutely hold, in theory. Transitivity is not violated by default - some level of heterogeneity in all of the factors you mention is inevitable even in randomised trials. In fact, even in a network of purely randomised trials, indirect comparisons are de facto observational in nature. It is also important to note that transitivity is the assumption that the distribution of effect modifiers is the same - if studies differ in areas that do not moderate the the outcome, then transitivity still arguably holds despite the differrences. If you can identify the important effect modifiers and demonstrate that they are consistent across the studies, then you can argue that it holds. You can quantitatively evaluate quantitative effect modifiers (eg. age) and qualitatively evaluate others (eg. route of administration of a drug).
You can also evaluate consistency/coherence, the statistical manifestation of transitivity to evaluate your assumptions regarding effect modifiers. You can further reduce the impact of transivity by critically assessing study quality using the ROBINS-I.
Other assumptions (joint randomizability, exchangeability, etc) can be evaluated in the same way as in networks of RCTs and often come down to clinical plausibility, much in the same way as in networks of randomised studies. While there will be a degree of confounding expected regardless, networks of randomised trials are also subject to confounding. Obviously you should critically evaluate your confidence in the findings and interpret your results in light of the limitations of the evidence, however.
  • asked a question related to Randomized Control Trials
Question
4 answers
Hello! I am working on review comparing cardiorespiratory adverse events b/w Propofol vs. alternative drugs in colonoscopy only. I entered the data from the studies that fit criteria (RCTs,Colonoscopies,Propofol vs. Alternative sedative) into RevMan. Any other statistical tools/programs should I be using for this method of research?
Relevant answer
Answer
R
Stata
Prism- Graphpad
  • asked a question related to Randomized Control Trials
Question
3 answers
I'm conducting a Risk of Bias Assessment in some RCTs and my effect of interest is that of adhering to intervention. However, some studies have intention-to-treat analysis only. Should I exclude these studies from the Assessment or not? In particular, how should I manage "missing outcome data" domain?
Relevant answer
Answer
I think you have posed an interesting question. I think it can only be sensibly answered by fully understanding why you want to determine the effect of adhering to an intervention. ITT is generally preferred for assessing 'real world' effectiveness in an unbiased fashion.
However, if you want to assess a fully adherent to treatment effect (or equivalent) you could still regard an intent to treat analysis as providing a potentially biased but still informative estimate. The likely extent and direction of bias can also be assessed - based on treatment dropouts and data provided on compliance. If such data is not available the risk of bias remains unclear. As ITT studies are likely to give a downward bias (assuming an effective treatment) unless non-compliant participants take more effective treatment, your estimate of effect from studies that are unbiased in other respects can provide you with a lower bound point estimate of the 'on-treatment effect'.
Under all circumstances that I can imagine this is better than selecting a small number of studies that report on treatment effects. This sample is likely to be biased in other ways but, in any case, the resulting estimates will be more error-prone because of the smaller samples (participants in studies and studies across the review).
I hope that helps.
Peter
  • asked a question related to Randomized Control Trials
Question
3 answers
Good day, I am interested in understanding a research paper that used the maximum likelihood of a generalized linear mixed model to predict the missing data from participants who did not complete an intervention in a randomized control trial.
I am not very familiar with statistical functions, and would like to understand this concept in simple terms to explain to my peers. Any information is appreciated.
Relevant answer
Answer
Thank you very much Professor Eliana, I shall read into the article.
  • asked a question related to Randomized Control Trials
Question
6 answers
My research team is conducting a review using the clinical trials and RCTs. For this I would like to know whether I can use the PICO framework or not.
Relevant answer
Answer
PICO framework is usually used in systematic and meta-analysis reviews.
  • asked a question related to Randomized Control Trials
Question
12 answers
1.Probability sampling or non probability sampling ?
2.Literarure showed both and more over Consecutive and purposive sampling for similar randomised controlled trial as well.
Relevant answer
Answer
Probability sampling. However, the ideal method of sampling depends on the study design. If the inclusion criteria states a wide range of population demographics, then stratified sampling would be appropriate to form different groups according to age, gender, etc. If the inclusion criteria is more restrictive and well defined - for example an intervention to identify the effects on children aged 11-15, then random sampling could be used as an alternative.
  • asked a question related to Randomized Control Trials
Question
4 answers
Hi to those who are familiar with using KMET scoring for methodological quality and publication. How do you go about scoring the KMET when RCTs are not indicated (emerging treatment / ethical issues), particularly items 2, 3, 4, 10? We had some recent reviewer feedback that the scores could not be so strong when they are not RCTs. Just wondering what other's experiences are? Many thanks in advance !
Relevant answer
Answer
Maria Aggestam thank you
  • asked a question related to Randomized Control Trials
Question
3 answers
Is it feasible to include a pooled analysis (for example post-hoc of two RCTs) in meta-analysis ?
Relevant answer
Answer
This is "flipping the coin". On one hand, post hoc subgroup analysis is "hypothesis generation" and "robustness checking". Otherwise, the bad new is "cherry picking" or "data dredging", one of the scientific misconduct. You have your individual choice with justification. However, there is no hard and first rule which restricts post hoc analysis in Systematic Review.
  • asked a question related to Randomized Control Trials
Question
4 answers
I conducted a controlled experiment to test the effect of a training intervention on workers. We were allocated 40 workers at the start by an NGO. This allocation was random as the workers were selected randomly from the pool of workers that the NGO knew. After this, we randomly divided the workers equally into two groups - treatment and control (20 and 20) and kept all other variables and conditions same for both, e.g. the regular activities of the workers. We then provided a month-long training on a topic to the workers in the treatment group. To know whether the training had some effect on the knowledge of the workers in the treatment group, we tested the knowledge of both the groups before and after the intervention. We found the two groups to differ in their knowledge at the baseline; the average percentage of the pre-test scores for the treatment group was 34%, and the control group was 28%. The difference was statistically significant (Wilcoxon sum test, p < 0.05). We use the Wilxocon tests as the data was not parametric. I would like to know what could be the potential factors which led to this difference.
Relevant answer
Answer
Hi,
compare pre and post scores by paired t-test in each group separately and comment on the change in each group. Can be converted to percentages.
knowledge is a random variable and people can have some degree of prior knowledge. That may not be matched. Matching can be on sociodemographic, years of service, etc.,
  • asked a question related to Randomized Control Trials
Question
4 answers
I have 2 studies for the same author, intervention and numbers of participants. The author mentioned in the title for the first study the effect of the intervention after 6 months with a specific outcome, and the second study title mentioned the effectiveness after 1 year with a different outcome from the 1st study.
Both studies are RCTs.
I've contacted the author but provided an invalid email address.
Can I consider these study as one study and taking both outcomes from both studies? Or taking each study independently?
Thanks for your cooperatives.
Relevant answer
Answer
I think it is a longitudinal study, therefore, one should regard the final outcome derived after intervention
  • asked a question related to Randomized Control Trials
Question
4 answers
I am trying to figure out the method to use to calculate the sample size for a matched cross-sectional study with 3 arms. Participants will be matched based on their age, geographic area, and number of children they have (0 vs 1). The primary outcome is experience of family violence (0 vs 1) and therefore it is a binary outcome.
Since there are more than two arms, ANOVA applies here (rather than a paired t-test). Since it's matched samples - repeated measures ANOVA seems to be the most relevant to calculate the sample size.
The catch is that the repeated ANOVA sample size calculations require at least two time-measurements, but this study is cross-sectional.
If anybody has any advice on how to calculate the sample size in this case, that would be much appreciated!
Relevant answer
Answer
Sample Size Rule
Sekaran (2013) wrote:
"Roscoe (1975) proposes the following rules of thumb for determining sample size:
1. Sample sizes larger than 30 and less than 500 are appropriate for most research.
2. Where samples are to be broken into sub-samples;(male/females, juniors/seniors, etc.), a minimum sample size of 30 for each category is necessary.
3. In multivariate research (including multiple regression analyses),the sample size should be several times (preferably 10 times or more) as large as the number of variables in the study.
4. For simple experimental research with tight experimental controls (matched pairs, etc.), successful research is possible with samples as small as 10 to 20 in size."
Reference
Sekaran, U., 2003. Research methods for business: A skill building approach. John Wiley & Sons.
  • asked a question related to Randomized Control Trials
Question
4 answers
Hello everyone, I would like to ask if the way that the sample size of this research was calculated is valid or correct. Is a study to evaluate the effect of gargling with Povidone iodine among COVID-19 patients. The text says “For this pilot study, we looked at Eggers et al. (2015), using Betadine gargle on MERS-CoV which showed a significant reduction of viral titer by a factor of 4.3 log10 TCID50/mL, and we calculated a sample size of 5 per arm or 20 samples in total”. From this data of the reduction of the viral titer in a previous study on MERS-CoV ¿It is valid to calculate the sample size this way for a new study on COVID-19?
Relevant answer
Answer
there are many different ways to estimate the sample size and you can select the suitable one for your research.
  • asked a question related to Randomized Control Trials
Question
5 answers
Hello dear researchers,
I want to know how should we interpret significant results of a meta-analysis while none of the included RCTs showed a significant difference?
is this because of low number of participants in RCTs? (approx. 20 to 30)
how should we discuss on this issue in our papers?
best,
Parsa
Relevant answer
Answer
Risk of bias is an option I think. To exclude high risk studies, sensitivity analysis should be undertaken on low risk studies. After that, pooled estimate of all studies and that of low risk of bias studies can be compared. When there is no variation between the estimates, it should be interpret as robust. If not, you should interpret with caution.
  • asked a question related to Randomized Control Trials
Question
2 answers
Dear Researchers, can anyone guide me on how can I combine split mouth and parallel design RCTs in the same meta-analysis for continuous outcomes in RevMan software? Which method I will choose?
Secondly, random effect model (inverse variance) in RevMan is the DerSimonian and Laird method?
Thank you in advance.
Relevant answer
Answer
I recommend you to use CMA because it supports different types of data entry
RevMan just supports routine data entry (Mean/SD OR Event/non-Event)
  • asked a question related to Randomized Control Trials
Question
3 answers
For a meta-analysis of interventions composed of only 1 RCT, 5 single-arm (SA) and 5 double-arm observational studies (DA), what is the best way to account for the lack of control arms amongst the SAs? More specifically, is it possible to 'donate' control arms from the RCT / DAs to allow for comparison of data within the recipient SAs?
Some literature proposes matching a control arm from an RCT with an SA (Zhang et al - doi: 10.1016/S1499-3872(12)60209-4), however, I'm unsure if this method adjusts for inherent differences in selection bias between the two designs. Please help?
I would be incredibly grateful for any advice, thank you!
Relevant answer
Answer
I would include the RCT and the 5 double-arm observational studies and exclude all single-arm studies.
  • asked a question related to Randomized Control Trials
Question
2 answers
Hello Everyone
Wanted to know about the most common Randomized Controlled Trails used in Occupational Settings. What is the rationale for preference of such RCTs.
Thank you
Relevant answer
Answer
Hi,
Before starting RCT kindly get the following into consideration.
Population Intervention Comparison Outcome Time (PICOT).
next, it could be
Open labelled or randomized
Blinding yes/No
parallel or cross over and number of periods
Type of intervention
Loss to follow up.
More rigorous the better identification of variables.
Follow EQUATOR guidelines
  • asked a question related to Randomized Control Trials
Question
1 answer
Hi
I am performing a systematic review. However, my topic has a limited number of RCTs. Thus, I was advised to add animal studies as well as other types of study designs (non-randomised, uncontrolled single-arm studies). Would that be an appropriate approach?
Please share your thoughts as well.
Many thanks
Relevant answer
Answer
depend on your research question, objectives, and Inclusion criteria.
  • asked a question related to Randomized Control Trials
Question
3 answers
I will have a 2X3 (pretest-posttest-follow up X experimental-control groups) split plot experimental design study, which is an RCT trying to test the effect of a group intervention on various variables. I want to see whether potential mediator variables say, a and b have mediated the effect of the intervention on the variable c. I will measure the potential mediators with the outcome for the three assessments. Does this design make sense and what kind of statistical procedures would you recommend? Advice on any introductory document, software as well as your answers would be very much appreciated.
Edit: I have previous experience on PROCESS macro for testing mediation on cross-sectional data, and am curious whether I can use it for this kind of experimental data.
Relevant answer
Answer
By type of experimental design
  • asked a question related to Randomized Control Trials
Question
4 answers
I would be most grateful for advice on interesting clinical cases where interventions have been approved based on hypothesis test results from high-quality RCTs, but where it has subsequently been discovered that the hypothesis test results corresponded to false positives. I am particularly interested in cases where, despite the positive RCT finding, the scientific rationale behind the hypothesis was later discredited.
Many thanks in advance!
Relevant answer
Answer
You would have to go very far back, I think, to the era before systematic reviews, and even there the problem was more generalisation of RCT findings to groups that hadn't been included in the trial, such as the use of beta-blockers for hypertension in older patients.
The only case I can think of where there were positive results, but the underlying rationale was later discredited was the Paris streetscape. After the cholera epidemic in (?) 1832, they demolished slums, widened streets and installed sewers to reduce exposure to 'miasma' (foul air, believed to cause cholera). It worked, but not for the reasons they thought.
  • asked a question related to Randomized Control Trials
Question
6 answers
In conducting meta-analysis of binary outcomes (in RCTs), in case the risk ratio has not been reported, what should we do with the studies that the number of outcomes occurrence is zero in both intervention and control groups? (for example, if we wanted to assess the effect of a drug on mortality, in some of the studies the number of deceased patients in both groups is zero), Is there any solution to not exclude these studies?
Relevant answer
Answer
you could do the meta analysis on risk differences (rather than ratios) if data available
see
  • asked a question related to Randomized Control Trials
Question
2 answers
CRIS Guidelines (Checklist for Reporting In-vitro Studies) are used, as the name implies, for laboratory studies in a similar manner to the CONSORT 2010 statement developed to improve the reporting of RCTs. For the CONSORT 2010 statement one finds a downloadable checklist (table format in a word doc). Has anyone come across such a table for the CRIS guidelines?
Relevant answer
Answer
dear Dr Ola,
there is the modified Consort guidelines for in vitro studies. Send me your email and I can send it.
  • asked a question related to Randomized Control Trials
Question
7 answers
Hi all,
I'm conducting a Systematic Review and the included studies have varying study designs including cohort, case control and cross-sectional (no RCTs found for my research question).
I cannot seem to find one Quality Assessment tool that assesses all these study designs in one, however the Newcastle Ottawa Scale has 3 different versions covering each of these study designs. Can I use the three different versions to assess my studies, or should I only be using one tool for all (which may require customizing it)?
Thanks!
Relevant answer
Answer
You can't use one version of the Newcastle Ottawa Scale for all types of studies. I recommend using the specific version for each type.
  • asked a question related to Randomized Control Trials
Question
3 answers
I have a question if anyone can answer it...
For a randomized control trial, when I come to publish the results to a journal...can we add more authors than those originally registered as co investigators? Or is it a rule that only those listed during registration, are the only authors that I can list when I come to publish to a journal?
Thanks in advance
Relevant answer
Answer
If I'm understanding your question correctly, you want to know when submitting an article to a journal who can be listed as an author?. Also, is that limited to co-investigators? Some journals do specify who should be listed as authors but I have not seen that limited to co-investigators. For example, someone who reviews your article before submission who is not a co-investigator but makes a contribution to the article can be listed. Here are some general guidelines for appropriate author listings;
Someone who
  • Contributed to some parts of the article and participated in interpreting the results.
  • Was involved in writing a draft of the article or revising it for intellectual content
  • Reviewed and approved the final version before submission. 
  • asked a question related to Randomized Control Trials
Question
3 answers
It's a debatable issue that RCTs are conducted in very controlled environment, whereas, in clinical settings, the scenerio is quite different. Further, it's not suitable research design for all systems of medicine. Therefore, is it correct to say RCTs as gold standard evidence or we have to explore more N of 1 type of designs.
Relevant answer
Answer
I would treat this an issue of internal versus external validity. If you are trying to rule out other factors that may have produced an experimental result (i.e., internal validity), then RCTs have very notable strengths. On the other hand, if you want to determine whether the results from experiment can apply to other settings (i.e., external validity), then RCTs have equally notable weaknesses.
In addition, I have recommend the concept of "transferability" which is more commonly considered in qualitative research. In that case, you have to consider the extent to which any set of research results can be applied in some other setting. The more the two settings differ, the less like it is that the results will be transferable.
  • asked a question related to Randomized Control Trials
Question
4 answers
Is there any organization can support the research proposals of clinical trials especially in countries where the funding and support of these trials is so weak or not finding?
Relevant answer
Answer
Some international organizations, such as AFRIMS, had received funding from Global Funds to conduct RCTs on anti-malarial drugs in Thailand and Cambodia. Funding can also be sought from pharmaceutical companies.
  • asked a question related to Randomized Control Trials
Question
4 answers
As you know the nature of split mouth design is paired, however when we have both split mouth RCTs and Parallel RCTs:
1) Can we summarize and merge both study designs in one analysis?
2) should we find External Correlation amount for split mouth design?
3) generally, How to do meta-analysis with split mouth designs?
Relevant answer
In some cases, it is useful to extract statistics from tables or data- if provided- and analyze each mouth part or quadrant along with side similar study groups from other included studies sharing same intervention.
  • asked a question related to Randomized Control Trials
Question
3 answers
I am designing a RCTs and I performed power analysis for a set of outcome variables. Some have ok power but very low in others. Do you have any idea how to improve power of a RCTs? We have a very limited control to recruit more study participants due to our budget constraint.
Thank you in advance.
Relevant answer
Answer
Cross-over designs are problematic because any treatment effect carry-over from one period to the next will invalidate them.
Sample size and power depend, among other things, on the nature of the primary outcome measure. Sample sizes for continuous variables are much smaller than for binary variables,
  • asked a question related to Randomized Control Trials
Question
3 answers
I‘m doing my dissertation and the results data are all positive in the reduction of tumor size with a drug I Would like to know what method should I use or any suggestion would be really appreciated
Relevant answer
Answer
Unless you have fewer than 3 well-designed RCTS, I suggest you include both RCTS+Cases in your qualitative analysis and only the RCTs in your quantitative analysis to conserve the quality of the evidence. Or you can include it if there is a methodological explanation, then you can perform sensitivity analyses and subgrouping. In that case, you will need to employ the random effect IVhet model and pool the continuous data (mean/SD) as a weighted mean difference, and the binary data (event/total) as a relative risk.
  • asked a question related to Randomized Control Trials
Question
4 answers
Hello.
This study had problems in randomization process. Yet, it has added in several MA of RCTs.
Discussing with several researchers, I always heard different opinions about its inclusion or not.
In case it wouldn't affect I2 or the effect size, would you include it in a MA of RCTs?
" At first, we intended to divide participants randomly into 3 groups, each with 20 men and 62 women: no walking training, moderate-intensity continuous walking training, and high-intensity interval walking training. However, a few of the participants were married couples and wanted to join the same group, and others, who lived a distance from an administrative center, wished to be assigned to the interval walking group so that they could visit a local community office nearer their homes."
Effects of High-Intensity Interval Walking Training on Physical Fitness
and Blood Pressure in Middle-Aged and Older People
KEN-ICHI NEMOTO, MS; HIROKAZU GEN-NO, PHD; SHIZUE MASUKI, PHD; KAZUNOBU OKAZAKI, PHD; AND HIROSHI NOSE, MD, PHD
Thanks,
Jorge
Relevant answer
Answer
My suggestion is pool im as many studies as possible.when we include studies all that we have to see is that the experimental variable has to he presnet in the study.For Cohen Ds calcultaion all we need is the SD mean and sample size of the two group of sMples one control and one experimental be taken.Inclusion in systematic review could be with unpublished work,dissertations,all sorts of journals.This reduces bias in selection of papers
  • asked a question related to Randomized Control Trials
Question
4 answers
I am planning to prepare Synopsis for Randomised Controlled Trial with Clinical Superiority Trial. I want to compare the Periapical Healing Outcome of Sealapex vs Endosequence BC Sealers. Primary Outcome is Post Endodontic Treatment Decrease in periapical radiolucency. There are currently no Human Studies available, but Limited Animal Studies are available in literature. I'm having difficulty in calculating the Sample Size for my Clinical Trial among Humans.
Please Guide.
Relevant answer
Answer
Here is a site that has start-up information, as well as a link to the full program: https://stats.idre.ucla.edu/other/gpower/
  • asked a question related to Randomized Control Trials
Question
5 answers
I have seen for the first time a systematic literature review and meta-analysis that has prospectively registered their review in PROSPERO and I am wondering what is the main reason for creating this database and will registration become essential to publish systematic literature reviews and meta analses in the future.
I can understand the reason for registering Clinical Trials in clinicaltrials.gov so that we have a record of all RCTs and can follow up with unfinished trials when carrying out a Cochrane Review etc but am not able to understand the need for systematic reviews.
Is it simply a database where people can learn the technique of carrying out good quality systematic reviews etc?
Ken
Relevant answer
Answer
Hi Ken,
I have registered systematic literature review protocols on PROSPERO previously. I find this a very useful database for the following reasons:
1. In planning to undertake a systematic literature review, the main motivation would be to compile and critique evidence on a specific research question. Usually a pragmatic search would highlight whether or not a systematic literature review in this area has been undertaken. However, given the upsurge in research teams conducting systematic literature reviews and the time commitment to undertake these robustly, there was always the possibility that some other team were conducting the same review and may have it published before your team completed it. This may then make all your hard work obsolete. More and more research teams now check PROSPERO before commencing a review in case another team is working on the same question.
2. The fields PROSPERO require when you register your protocol can be helpful to adopt as a framework for developing your review strategy. You can also learn from other researchers on how they approached research questions in similar clinical areas.
3. It is also possible to develop research collaborations with active researchers who are interested in the same research area.
I am not able to advise whether this will become a requirement by journals to publish the actual review when completed. However, in my experience some reviews don't make it to the publication stage due to limitations on the evidence, value to the academic community or other technical issues. However, a body of evidence was generated in the process of the review which can be helpful to other researchers and PROSPERO affords researchers an extra outlet to identify this work.
I hope this is helpful.
  • asked a question related to Randomized Control Trials
Question
3 answers
Dear all,
I want to conduct a network meta-analysis to find out the efficacy of different surgical interventions. I will use R as the main software. However, this is the first time I do a network meta-analysis, would you please give me some suggestions?
1) Frequentist vs. Bayesian methods, which one that I should use?
2) The network needs only RCTs, or should I include high-quality protective cohort studies, make a comparison between them?
Thank you very much.
Relevant answer
Answer
Have a look at the Generalised Pairwise Modelling framework available through MetaXL (www.epigear.com)
The User Guide is on the site and has useful information
  • asked a question related to Randomized Control Trials
Question
5 answers
Hello RG researchers
What is the wrong that need major correction?
Data integrity of 35 randomised controlled trials in women’ health
Thanks for your kindly help
Relevant answer
Answer
I'm not sure if I understood correctly your question. Anyways, interesting paper!
The researchers identified that in 26 trials published by two authors (from the same Egyptian institution, in the same disease, in a short period of time) the baseline characteristics and outcome data were similar, and the statistical tests showed that the data do not behave with the variability expected by chance. Thus, they suggested that there is an issue with data integrity or fraud in such trials. If I were the author of these 26 trials and if the data published in the studies are legitimate, I would respond with a letter to the scientific community and make available the databases and informed consent signed by patients as proof of integrity. Otherwise, it is worth retracting the trials at the journals where they were published. Best Regards,
Erica
  • asked a question related to Randomized Control Trials
Question
6 answers
Hi,
I should performe a network meta-analysis and I would like to know if in the protocol of systematic review for NMA I can exclude the RCTs that assess the same drug with different dosage or instead is better include that kind of RCT in the systematic review and exclude them in the NMA directly.
Many thanks,
Giuseppe
Relevant answer
Answer
I would advise you to include all studies on all doses and highlight whether increasing the dose was associated with more improvement? or safety issues? This can be further examined by a Meta-regression analysis.
  • asked a question related to Randomized Control Trials
Question
3 answers
Enrichment design in a clinical trial to evaluate efficacy and safety of a new investigational product
Relevant answer
Answer
I would recommend the answer of Andrew Sutton
  • asked a question related to Randomized Control Trials
Question
6 answers
Any comment on this English study ( 2 parts) ? Is this procedure can be used for protection?
-Hypertonic saline nasal irrigation and gargling should be considered as a treatment option for COVID-19(Jogh 2020)
-A pilot, open labelled, randomised controlled trial of hypertonic saline nasal irrigation and gargling for the common cold (nature 2019)
Relevant answer
Answer
Hypertonic saline nasal irrigation and gargling is often used for common cold and other throat infections in India since many years.
  • asked a question related to Randomized Control Trials
Question
6 answers
I'm conducting a systematic review and meta-analysis, and have a list of publications that employed 2-group (mobile app vs control) or 3-group (mobile app, on-site, control) RCTs. Most of the publications reported 2-groups RCTs. All studies administered similar interventions (use of mobile app for health services), but 3-group RCTs added an additional intervention that involved a non-mobile app intervention.
Would it be alright to combine 2-group RCTs with 3-group RCTs in a meta-analysis? I'm guessing no as it would be very inconsistency to do so. If anyone else could advice and or have better solutions (and references!), that'll be really helpful!
Thanks!
Relevant answer
Answer
I see... Yes that also works, thanks very much for your explanation Robert Boer !
  • asked a question related to Randomized Control Trials
Question
6 answers
Although originally used in meta-analysis, Forest plot is a popular graphical approach for displaying the results of subgroup analysis in randomized controlled trials. Can SPSS create a forest plot like that? 
Relevant answer
Answer
Stata, Prism Graph will plot it. We combined a table and forest plot together and generate this figure.
  • asked a question related to Randomized Control Trials
Question
2 answers
We are analysing data on a cluster randomised control trial. We have treated the clusters (classes, N = 8) as fixed effects, thus dummy coding to create 7 different variables. This is because we had non-convergance issues when running a MLM.
However, when we run the analysis (in SPSS or STATA), the analysis removes one of the classes (not always the same one, depending on the outcome), and the treatment variable. We are perplexed why this is happening, so if anyone has any suggestions and workarounds, they would be gratefully received.
Relevant answer
Answer
you need k-1 levels so that estimates for each level are relative to the left-out level.
For example, if you're dealing with sex as a variable in a multiple regression, you'd code with something like (male=0, female=1), and a regression formula may look like this:
Y = B0 + B1X1 + B2Sex
Which means that the predicted value for a male is
Y = B0 + B1X1 + B2*0
= B0 + B1X1
but if it's female then the predicted value is
Y = B0 + B1X1 + B2*1
If you have three groups then the model might look like:
Y = B0 + B1X1 + B2Grp2 + B3Grp3
so the predicted value for group#1 is
Y = B0 + B1X1 + B2*0 + B3*0
for group#2 the value is
Y = B0 + B1X1 + B2*1 + B3*0
and for group#3
Y = B0 + B1X1 + B2*0 + B3*1
  • asked a question related to Randomized Control Trials
Question
3 answers
I am going to do a meta-analysis of RCTs for a drug. Is is possible to include that drug derivatives RCTs into meta-analysis?
Relevant answer
Answer
It is up to you. You determine the research question and the study inclusion/exclusion criteria for the systematic review. Remember, for whatever you decide, you must be able to justify it. Think "would an expert agree with your decision?".
  • asked a question related to Randomized Control Trials
Question
5 answers
Which analysis can I conduct to determine effect of a health education intervention in a cluster randomised control with with both baseline and end term data for both control and intervention group.
Relevant answer
Answer
I found this a generally useful book on the design and analysis of such designs
Richard J. Hayes Lawrence H. Moulton Cluster Randomised Trials ;Chapman & Hall/CRC Interdisciplinary Statistics Series
The following is short and to the point
and fthis makes the argument that you need appropriate analytical methods
Campbell MK, Mollison J, Steen N, Grimshaw JM and Eccles M. Analysis of cluster randomized trials in primary care: a practical approach. Family Practice 2000; 17: 192–196.
If you are examining the effect of a group-level intervention, then a multilevel model seems a natural way to analyses it- an example on Research Gate is
Reducing child conduct problems and promoting social skills in a middle-income country: cluster randomised controlled trial.
They are quite widely used in epidemiology often under the name of a community intervention trial. The key is that the intervention is at a different level to the outcome. For example intervention to a class but pre- and post- outcome measures are at the pupil level.
  • asked a question related to Randomized Control Trials
Question
5 answers
Dear All,
Thank you for reading this question. I am conducting a systematic review of literature. The included articles have different study designs, for examples some of them are randomized control trials and some observation studies. I know how to assess the quality of each kind of these articles and I think I should use a checklist for RCTs and another checklist for observational studies. As a result, I should report their quality separately. Is it right?
Thanks,
Relevant answer
Answer
Hi Behzad,
You refer to assessing "quality" and "checklists" for the different study designs. I am not sure if you are alluding to risk of bias assessment (which typically uses checklists such as RoB2, ROBINS-I, NOS, among others) or the GRADE assessment of quality of evidence (of which risk of bias assessment is only one of the criteria for rating down the quality of evidence). Since you explicitly mentioned checklists, I shall assume it is the latter.
As you pointed out, there are different checklists for assessing risk of bias for different study designs, i.e. there are different criteria used for each study design. It therefore makes sense and is common practice to summarize risk of bias assessments according to study design. The advantage of doing so is particularly apparent when you are summarizing risk of bias using a summary figure (for which the checklist criteria comprise the column headings): https://handbook-5-1.cochrane.org/chapter_8/figure_8_6_c_example_of_a_risk_of_bias_summary_figure.htm
Best Regards,
Elenore
  • asked a question related to Randomized Control Trials
Question
3 answers
I am planning to conduct a single centre, participant blinded RCT to assess the effectiveness of a pharmacist-led educational intervention on preventing hypoglycaemia in elderly patients with type 2 DM. I am also planning to conduct a nested qualitative study as a process evaluation for the intervention itself to see what worked and what didn't? and why? the contextual factors that may affect the delivery? and how the intervention can be scaled up in the future?
Each participant will be followed up after 3 months from recruitment and I am planning to conduct the qualitative interview at the follow-up visit. I am planning to recruit 200 patients in each arm.
If anyone can help with the best sampling strategy and sample size for this RCT? If you also have any suggections about the questions to be asked in the interview?
Thanks in advance
Relevant answer
Answer
RCT involves randomly allocating research participants in two or more groups. As shown in several nested qualitative studies of RCT, either convenience or critical case sampling strategy can be suitable sampling strategies for such research.
  • asked a question related to Randomized Control Trials
Question
8 answers
I want to analyze the efficacy of an intervention. The intervention is a cluster randomized control trial and we are comparing two waves (pretest vs. posttest). We have also two groups (experimental vs control group). I found that the basic formula is that ICC = between variance/between variance+whitin variance; However, I think that this formula doesn't consider that I have two time points. I found another formula in Kärna et al. (2011)' paper (see link below), where ICC is variance between cluster / variance between waves of measurement + variance between cluster + variance between individuals. I have two questions:
1. What do you think about Kärna et al. (2011)` formula? And if you think that it is ok, then, How I can compute the different variances?
2. If you consider that the formula proposed by Kärna et al. (2011) is not appropriate, How would you compute ICC in my case?
Thank you in advance!!!
Relevant answer
Answer
Hi Felix. I calculated ICC in MPLUS. First, I transposed my variables (for example anger time 1 and anger time 2) in cases in SPSS. I created an ID to identify two-time points. I also included my cluster variable (in my case, "school") in my dataset.
With this information, in MPLUS, I employed this syntax:
VARIABLE:
NAMES ARE
id
School
anger;
USEVARIABLE ARE
anger ;
CLUSTER = school id;
ANALYSIS:
TYPE = ThreeLEVEL basic;
I don't know if I resolve your doubt.
Best,
  • asked a question related to Randomized Control Trials
Question
1 answer
Basically I have two locations (Y and Z) and each location had an intervention and control arm. For each location I am trying to check the following: Location A: I want to check the effect of Variable A on Variable C on the intervention group. Variable A and C have a casual relationship. Ideally, the outcome of A will determine the outcome of C. This will be repeated with variable B and C. A and B are independent variables with C a dependent variable. Would a coefficient of correlation be ideal?
On the other end, I want to check the effect of variable A and B on C. Again within both arms and both locations. Again A, B and C have a casual relationship. Would a coefficient of correlation be ideal seeing as 1 variable is dependent on 2 others?
Relevant answer
Answer
If the variables you measure are on continuous scale normally distributed correlation coefficient will do.
  • asked a question related to Randomized Control Trials
Question
6 answers
Most of the journals accept RCT study if it is registered. Where to get it register? NIH? WHO? which way is easy to register it. Kindly help in this regard.
Thanks
Relevant answer
  • asked a question related to Randomized Control Trials
Question
2 answers
For medical researchers, "external validity" seems to show up as a limitation in RCTs. I think medical researchers typically deal with the issue of generalizability -- external validity -- through meta-analysis. Economists can sometimes run several RCTs and complete a meta-analysis (see, for example, the MIT poverty lab), but in the case of large field experiments, the number of times a similar study can be run is limited. So you end up with one or two field experiments on a given topic, and concerns about external validity.
Deaton and Cartwright address this issue in the context of economics. However, it often seems to me that philosophers argue that results of a single trial either "are" or "are not" externally valid, while economists look a bit more closely at the context to which the results are about to be generalized. When they do that, they are (presumably) importing knowledge from other research or from history to determine whether the new application is sufficiently similar to the experimental situation to allow generalizability. Is this a substitute for a meta-analysis?
Relevant answer
Answer
I think the issue of external validity or generalizability should not be restricted to randomized control trials alone, it should be applicable to most research designs, cross-sectional, case-control, cohort and others.
External validity should depend on quality of the research, the setting, methodology, limitations etc.
  • asked a question related to Randomized Control Trials
Question
6 answers
Hello,
I have 3 one-arm pretest-posttest quasi-experiments.
Is it possible to do meta-analysis to their findings?
If so, how can I analyse them using Revman?
For me, I considered pretest data as control group's data, and posttest data as experimental group's data. Is that correct?
Also, I have other 3 two-arm RCTs. These studies have the same outcome in the former three studies. my question is do you recommend me to make two meta-analyses: the first to meta-analyse the former study, and the second to meta-analyse the latter studies? Or Conduct one meta-analysis using data of pre&posttest for the experimental group in all 6 studies.
Thanks in advance.
Relevant answer
Answer
Whether to do or not to do this meta analysis is a difficult one. I agree perfectly well with Peter. You can do what most people would have done: summarize and try to investigate the effect of study design on treatment effect.
  • asked a question related to Randomized Control Trials
Question
5 answers
I am currently conducting a systematic review and meta analysis.
Relevant answer
Answer
I agree with Warren. However do consider submitting as a brief report.
  • asked a question related to Randomized Control Trials
Question
8 answers
Hi All,
I am currently undertaking a systematic review. Two of the studies, completed by nearly the same authors, seem to have used the same set of participants.
The differences mainly lies in the study design:
Study 1 included a double-blind, randomized, sham controlled trial (5 days/ week for 2 weeks) in twenty patients with 3:2 ratio. At baseline, each patient underwent a clinical evaluation. Assessments were then carried out immediately after either shaml or real stimulation (post-stimulation, T1), at one-month (T2) and at three-months follow-up (T3).
Study 2 included a randomized, double-blind, sham-controlled, crossover trial (5 d/wk for 2 weeks) in 20 patients with a 1:1 ratio. Each patient underwent a clinical evaluation before and after real or sham stimulation. A follow-up evaluation was performed at 1 and 3 months with a crossover washout period of 3 months. After the washout, After a washout period of 3 months after the last visit (i.e., T3), each patient received the opposite treatment (crossover phase)
and underwent the same standardized assessment as in the first
phase, at baseline, at 2 weeks, at 1 month, and at 3 months
Both studies made use of the same outcome measures and data analysis involves identical data analysis measures.
What is your view on the below:
1. Should both studies be included in the review?
Since the authors seem to have made use of the same participants (same research authors, same institution, identical sample size, same outcome measurements), could the inclusion of the two studies bias the findings of the review?
Thank you!
Pasquale
Relevant answer
Answer
I would contact the authors and ask if those participants in two papers are same or not. If the participants are same in both papers, I would not use both of them. If the participants are different or the outcome differ (for example different time, different outcome ect) I would include both of papers. However, you must write the decision you have made, or otherwise readers considers with the same questions.
  • asked a question related to Randomized Control Trials
Question
2 answers
Traditionally we know that educational strategies are rarely subjected to rigorous evaluation and so we would ideally perform an RCT approach to test the theory for my study. However, as we all know, in tertiary education this is not feasible, ethical or equitable as students can’t be individually randomised.  So the method we employed was a non Randomised control trial. The trial was 2 arm.
Participants were allocated to an intervention arm on the basis of what course they were enrolled in. Even though there were 140 students in one arm and 77 in the other, we tried to ensure we had the best opportunity for the 2 groups to be equivalent in some ways like first year foundation courses on their first day of uni in first semester with learning outcomes almost the same in all courses.
Group 1 undertook the standard experiential task.
Group 2 undertook the standard experiential task with an additional experiential component.
I need to find guidelines for writing this study up.
I have looked at the CONSORT guidelines.
I have read a little about TREND guidelines and STROBE guidelines but I would dearly love input from academics who may have had experience in this study design.
Relevant answer
Answer
Thanks so much Victor, how generous of your time to write such a thorough response. K
  • asked a question related to Randomized Control Trials
Question
3 answers
One of the main objectives of a meta-analysis is to obtain reliable estimates of treatment effects when RCT results are not sufficient. The optimal information size (OIS) can be defined as the minimum amount of information needed in a meta-analysis to obtain reliable conclusions about an intervention. The calculation of OIS suggests a limit on the conduct of clinical trials in the area, indicating whether the existing information is sufficiently convincing to the extent that new RCTs on that subject are no longer needed.
The OIS assumes that all RCTs studied are part of a single large RCT. However, when we observe data heterogeneity, it is not appropriate to make this kind of assumption. Consequently, heterogeneity should be considered when the OIS is calculated, since its increase requires a larger information size.
In this sense my questions are as follows:
- How to calculate the optimal information size (OIS) adjusted for meta-analysis heterogeneity?
- If there is no possibility of determining if the meta-analysis is conclusive using the tool of the calculation of the OIS, is it possible and advisable to perform a qualitative evaluation of the confidence interval generated by the meta-analysis?
Relevant answer
Answer
Dear Prof. Luís, according to Higgins and Green in the Cochrane Handbook (2011) you can deal with the heterogeneity, so please, take a look at the chapter 9.6 Investigating heterogeneity. About the Network Meta-analysis and the Optimum size of information adjusted for heterogeneity (OSI-H) can be calculated in conjunct with other parameters. So, you can read Georgia Salanti and Sofia Dias in many articles and also videos. There's a book by Sofia Dias - NMA for Decision Making, 2018. You can also see the following summary at the page: https://methods.cochrane.org/sites/methods.cochrane.org.statistics/files/public/uploads/SMG_training_course_cardiff/2010_SMG_training_cardiff_day1_session4_harbord.pdf. You can also find some formulas in a Brazilian Dissertation written by Mathias Azevedo Bastian Bressel about pharmacological interventions and he used the TOI/OIS: https://www.lume.ufrgs.br/bitstream/handle/10183/16365/000695956.pdf?sequence=1
So, I think with this material and the Kenneth Lo's material you will have a good overview about so many important points to reflect and think critically. Don't forget to read the Cochrane Handbook. If you need a special version, I can send you by email, Ok? Have a nice weed ahead! Best regards from Brazil. Ro
  • asked a question related to Randomized Control Trials
Question
2 answers
Basically I`ve done a review of the literature on disengagement from Early Intervention Psychosis services using systematic methods. There was heterogeneity across the studies, no RCTs and I`ve used `vote counting` to make sense of the results
Relevant answer
Answer
According to Cochrane's handbook, there are two main concerns about using vote counting to make sense of results: by considering each study on its own you are not allocating different weight to them, and too much emphasis is given to significant/not-significant, without considering range of effects (e.g., CI) (https://handbook-5-1.cochrane.org/chapter_9/9_4_11_use_of_vote_counting_for_meta_analysis.htm)
  • asked a question related to Randomized Control Trials
Question
3 answers
I know that the Education Endowment Foundation (EEF) and National Center for Educational Evaluation and Regional Assistance (NCEE) publish their RCT data about educational interventions. Do you know of any other organisations/funders that publish RCTs data?
  • asked a question related to Randomized Control Trials
Question
3 answers
I have information regarding a study using following equation
for continuous outcome
Noordzij, M., Tripepi, G., Dekker, F., Zoccali, C., Tanck, M., and Jager, K. (2010). Sample size calculations: basic principles and common pitfalls. Nephrology Dialysis Transplantation25, 1388-1393. doi:10.1093/ndt/gfp732.
Here, U1= log10(6.21), we want to reduce one log of E.coli count by providing intervention so U2 would be= log10 (1.61), SD= log10(1.21), determine n/arm=?, two arm two sided test will use.
After converting all these result and value I got very odd number.
Relevant answer
Answer
Hi, the approach is correct in principle. However you may need to consider whether your study is going to be a superiority (clinical vs. statistics), a non-inferiority, or an equivalence trial. You may want to consider the attrition rate as well. I recommend you the paper by Zhong B., 2009 (How to Calculate Sample Size in Randomized Controlled Trial?). A user-friendly, yet a basic, software for this purpose is IBM Sample Size calculator. It includes prospective studies with dichotomous outcomes, controlled. I think this software is now discontinued and it is part of IBM SPSS Statistics (or its plug-ins) nowadays. Besides that, NCSS is the one I recommend and I used myself. I am confident additional software can handle this computations, including G-Power and R. Good luck with your project. P.S. please invest your time in proper sample computation because this is an essential step towards valid results and clinically-informative effect sizes. Michele.
  • asked a question related to Randomized Control Trials
Question
5 answers
Double blind RCT
Relevant answer
Answer
Hima - 'authenticity' is not just reliant on levels and degrees of blinding. More blinding does not necessarily equate to more rigour. Other factors - such as levels of control, randomisation are key. A well conducted single-blinded trial is better than a poorer double-blinded. If both a single and double-blinded study are of equal quality - I would argue no one is better than the other - as long as the context for the level of blindness is well rationalised and justified in each case.
  • asked a question related to Randomized Control Trials
Question
1 answer
Oncology phase III RCTs usually compare a standard of care treatment with an experimantal one on the basis of a survival primary endpoint (e.g. PFS or OS). Along with the overall results, subgroups analyses (e.g. sex, age, ethnicity) are usually presented. Sometimes, a specific subgroup is identified as benefitting mainly from the experimental treatment; in this case, a new RCT aiming to determine efficacy in this specific subgroup is planned.
I would like to run a Bayesian network meta regression with a dichotomous covariate (say Z) on a validated cutoff for a biomarker (Z=0 if the continuous value of the biomarker is less than the cutoff, Z=1 otherwise). In doing so, I would have to regress results coming from the biomarker subgroups hazard ratio on three RCTs and one coming from a specific RCT pertaining only Z=1.
Is it correct to use these results in a single network meta regression? Or would it be better to run separate network meta-analyses for each subgroup (Z=0 and Z=1)?
I sense there is bias in both approaches, but obviuosly the Bayesian network analysis would only serve as an exploratory tool.
p.s. in the case of the Bayesian network meta regression I think that a prior on the common variance which allows only a small borrowing is more appropriate.
Relevant answer
Answer
Follow.
  • asked a question related to Randomized Control Trials
Question
3 answers
I am comparing two groups in a randomised controlled trial ( pre post design) using Linear Mixed Model in SPSS. How can I plot the group*time interaction?
Relevant answer
Answer
IMHO just one more reason to discard SPSS. Suggest something like R. The link to Prof Davidian's notes may be helpful. Best, D Booth
  • asked a question related to Randomized Control Trials
Question
6 answers
Hi all, I have to present a research proposal for a job interview. I want to see the effects of value-reminder pictures near the healthier food on dietary intake. Dependent variable is the sales in healthy food.
1. group: will see the pictures of family, travel etc. to remember their values, the healthier option will be closer to the entrance as proximity affects the decision.
2. group: will only have the healthier option closer to the entrance of cafeteria, so testing only proximity effect.
3. group: No intervention
Should I use regression or anova? Is there anything specific I need to do because the 1st group's intervention has 2 components? Can I call it randomised controlled trial? Any help will be appreciated. Thank you!
Relevant answer
Answer
The normality assumption of the residuals applies to regression AND the ANOVA approach, both are part of the general linear model, so this is not a distinction to decide which one to use. I would use an ANOVA, since it is typically used for experimental data and directly handles and shows mean differences between the groups. So it is easier to conduct and interpret.
Strictly speaking: you can get the exact same results with the regression approach, but you have to some form of dummy coding to handle the groups, since predictors in regression have to be metric or dichotomous. So it is more work to do and a bit more difficult to get the relevant information if you are not used to multiple regression. But it works, too.
  • asked a question related to Randomized Control Trials
Question
4 answers
Is there a tool that can be used across all study designs?
In my SR there are RCTs, observational studies, etc.
Or do I have to use several?
Many thanks
Georgina
Relevant answer