Step by step:

  1. Create a new project. Give the project a name and a type, select, "Grade evidence profile"
  2. Click on the button labelled "Add management question"
  3. Enter your PICO formatted question, in our case that would be: "Should mindfulness meditation vs. non-mindfulness based approaches be used for stress?"; if you wan to use for specific communities, mention that instead of the outcome. But I recommend you use 'stress' or another outcome based word or expression here. You can add up to 12 management questions
  4. Then save the management question by clicking on the save icon that looks like a floppy disk.
  5. Now click on "Add outcome"
Take a break: we will show you how to appraise individiual articles that are individual or results pooled AND abstraction of data from a Cochrane Review

How to add data from single studies

Let's start with the simplest of them all: take a SINGLE study and add data from a single study to the Gradepro table. The idea is not different for single as opposed to pooled studies. In each case you will need to specify the outcomes first, and then, based on your reading of the study, you will identify individual values and plug in those values in the individual cells. So let's get started with a single study and then we will build up the rest of the entries.We will use the vander Zwan article to put in data.
Judith van der Zwan and colleagues (2015) conducted a study on 76 participants and tested the effectiveness of physical activity (PA), mindfulness meditation (MM), or heart rate variability biofeedback (HRV-BF) for reducing stress. They randomly allocated these 76 individuals into the three modalities of treatment and conducted the study for 5 weeks. They obtained data on their stress levels using a questionnaire at the beginning of the study, during the study, and 6 weeks after conclusion of the study. If you want to follow along, open up the following webpage, download the PDF file and follow along
https://link.springer.com/content/pdf/10.1007%2Fs10484-015-9293-x.pdf
Follow the sequence.
First of all, fill in the outcomes information:
  1. Log into gradepro, start a project or open your existing project, and set up the question till the step of "Add outcome"
  2. We will start with adding the outcome. Here, we will include only ONE outcome (however, you are not restricted to one outcome, you can and should add as many outcomes that are relevant to your research question). We will select the 'anxiety' as an outcome. Let's fill in the entries
  3. For the main outcome, we write, "Anxiety"
  4. Short Name: 'Anxiety"
  5. Assessed/measured with: "DASS"
  6. Length of follow up group: We leave the first entry blank as it does not fit the proper unit (as they did a total of five weeks of study), in the middle box, we write '5' as they had five weeks of study, and in the third box, we select 'weeks'
  7. In the 'Type', we select continuous as they used Cohen's "d", which is a measurement of standardised mean difference in the pre-study and post follow up scores; in the second box, we select 'single study'
  8. For "Type of scale": "Ratio/interval measurement scale"
  9. For conventional lower and upper limit of scale: note from the methods section of the paper that the DASS scale had three subscales; each subscale had seven items; each item had the lowest possible score of 0 and highest possible score of 3, therefore the theoretical lowest and highest scores would be 0 and 21. Enter these figures in the empty boxes
  10. Then we click on the 'floppy disc' to save the entry. At this point, if you want to add another outcome, you can do so. For the sake of brevity, we will move on.
Second, fill in the 'Certainty Assessment' section next -- work from left to right
  1. Number of studies: 1, then click 'Apply'
  2. Study design: Randomised trial
  3. Risk of bias: You have three choices ("not serious", "serious", and "very serious"). This is a subjective choice but to make the decision, you must read the methods section of the study. This was a randomised controlled trial, but there were a few things that limited the scope of the study - first, they did not report 'how' exactly they 'randomised' the participants in their study. Second, this being a lifestyle intervention trial where they tested physical activity, meditation, device based assessment, it was impossible for them to 'blind' the participants or themselves to be agnostic of who received what interventions. Because of these limitations, we would rate this section as 'serious' and proceed to add explanations that we have provided here. If you add the explanations, and save it, the programme adds an 'a' to the box
  4. Inconsistency: For a single study that has several outcomes, check the 'size' of the effect estimate and the 'direction' of the effect estimates. Check the results section of the paper. In this paper, study  Table 2 and columns "pre-test-post-test' and 'pre-test-follow-up' and note the columns under 'd' ('d' = Cohen's 'd', or standardised mean difference). You will see that the numbers are similar for stress, anxiety, and deperession for DASS, with very little variablity. Based on this, we will decide that inconsistency is not a serious limitation here. On the other hand, there may be situations where for similar outcomes and using similar tests, the magnitude and direction of the effect estimates might differ at least by 20% or more (candidates for 'serious' inconsistencies). For pooled studies, we would refer to I-squared values to test for heterogeneity in the estimates plus variation around the point estimates. Here, we would come to the conclusion that inconsistency is 'not serious'. Select the drop down box to indicate that.
  5. imprecision: In order to decide imprecision, study the associated p-value and where available 95% confidence interval around the point estimate. The authors in the present study did not report 95% confidence interval around the point estimate, but they did report p-values. On close inspection, you can see that while there were some drop in the scores of anxiety, the between group estimates did not meet statistical or clinical significance. This is the reason, we will rate this as imprecise.
  6. Other considerations: We will consider four other considerations - publication bias (this is not an issue with single studies), large effect size (we did not find large effect size in this study), dose-response effect (the authors did not report dose-response effect and this was not an issue here), and whether plausible confounding could have impacted the effects (here, because we studied randomised trial, this was not an issue). Accordingly, we would have made these choices. Now you see that the 'certainty' box is labelled as 'low'. This indicates that we would have 'low' level of certainty or surety in interpreting the findings of this study
Third, fill in the summary of findings section next
  1. Number of patients on the mindfulness meditation arm: 24
  2. Number of patients on the non-mindfulness approach: 42
  3. Note that the relative effect is now greyed out, so fill in the 'absolute effect' using the Cohen's d but note that you cannot fill in the 95% CI as this was not reported.  The absolute effect is: -0.24 SD lower  as we have decided to study standardised mean difference, and we have chosen to report the difference between pre-intervention and post intervention. If you want to report both pre-intervention and post intervention plus pre-intervention and post follow up, you will need to add another outcome and report it as such.
This completes reporting of one study, a single study. We will move on, and now we will fill in findings from a Cochrane review.

How to use Cochrane Review and add information to GRADEpro GDT

Now that we have added the findings from a single study, let us add findings from a Cochrane review. We will use the following Cochrane review to add the findings:
Krisanaprakornkit, T, Sriraj, W., Piyavhatkul, N., & Laopaiboon, M. (2006). Meditation therapy for anxiety disorders. Cochrane Database of Systematic Reviews, (1). https://doi.org/10.1002/14651858.CD004998.pub2
As before, download and open the PDF of the Cochrane review and remember to open the 'full text' of the review. As before, we will follow this in three steps:
To do this, as before, we will refer to the methods and results of the paper itself.  To provide a brief overview of this paper, this is a fairly old paper (almost 13 years old at the time of writing this tutorial) and as you can see, this is a Cochrane meta analysis.  We will use the following PDF to abstract information and fill the GDT
Step by step:
  1. Add outcome (assuming that you have already set up your questions). Open up the full text and read the methods section. The outcome measure we will report here is 'clinical state of anxiety at the end of trial (continuous measure)'; so note we write:
  2. Long name: 'Clinical state of anxiety'
  3. Short name: 'Anxiety'
  4. Measured with: ' different scales and inventories' (this is because in this meta-analysis the author mentioned a number of different scales and inventories rather than single instruments: in cases where the author would indicate that they used a single scale to measure anything, you can mention that)
  5. Type: 'continuous'
  6. Type: 'pooled'
  7. Length of follow up: We leave it blank as different studies may have different lengths of follow up so we are not sure
  8. Choose type of scale: 'Ratio/interval measurement scale' (but as different types of scales and instruments were used, we are not sure about the score range and we cannot fill in the rest of the boxes)
At this stage click on the save button and move to the next step.

Step 2: Fill in the certainty assessment

Number of studies: 2: Read the results section of this meta-analysis and you see that they have 2 studies
Study design: Randomised controlled trials
Risk of bias: Serious as the first study had issues with allocation concealment, and the second study did not describe how they randomised and both studies had issues with not being able to blind participants and uneven ratios of male:female. These would go into explanations
Inconsistency: Serious (you could also rate very serious if you want as one of these studies reported some measures but the other did not, so the findings were not consistent across the studies)
Indirectness: Serious (While the studies used direct measurement of anxiety states from participants using scales, none of the findings or measurements 'directly' reported anxiety state changes)
Imprecision: Not Serious (We have data from only one study and the point estimate relevant for anxiety according to the perceived stress scale is -7.57 (95%CI -13.06, -2.08 on 14 participants)
Other considerations: We found that this meta analysis was based on only two studies, one of which was low quality and did not report anything, the other had reported some measures. It is possible that other studies were conducted but the authors missed them, and being based on one study, they did not report or examine formal publication bias. However the effect size was reasonably large and as they had randomisatio, most confounding variables were adjusted for.
We now move to the Summary of findings but the notable thing here is that, for this outcome, the level of quality of evidence is very low. Let's move to the summary of findings:

Step 3: Fill in Summary of Findings

For this meta analysis , we will fill in the summary of findings tables from the following data: