How to use Consensus to do a medical research review (FAST!)

Imagine getting a trustworthy snapshot of the medical literature to answer a specific clinical question - before your next patient even sits down.

That’s what Consensus, an AI research tool, can do. In this video, see how you can best put it to work. 

You’ll learn how to:

• Reduce hours of searching to just a few clicks
• Get answers from the most prestigious and highly cited journals
• Spot research gaps that can influence your diagnosis and treatment plan

If you want to reclaim time while boosting the quality of your clinical decisions, don’t miss this tutorial. 

[00:00]

Hey guys, it's Franz from Medmastery. Today I want to show you a really cool tool that I'm using almost on a daily basis and it's called Consensus.

What does Consensus AI do?

[00:10]

What consensus does is it helps you answer clinical questions. If you don't have time to do lengthy research on PubMed and read the relevant literature or if you don't have time to search for the latest meta analysis on a topic or if there is no metaanalysis on a certain clinical question, consensus comes to the rescue.

How to use Consensus AI to find medical research and review the evidence

[00:34]

So let me show you what it can do. So let's say we're interested in the question if social media use is associated with negative effects on teenagers mental health. You can do a quick fast search with a summary of 10 papers. You can do a pro search which is a smart search analysis of the 20 most relevant papers and then you can do a deep research and I'll show you what that is a little later.

Let's start with the pro search. Hit enter and you see the results pop up really quickly. So this is really for a fast search but it's already very good and you'll see that this result is very consistent with what we'll get in the deep research a little later. So the answer is yes, social media use is associated with negative effects on teenagers mental health especially with problematic or excessive use. Though the strength and nature of this association vary.

Then you have the consensus meter that clearly shows you where the evidence is leaning. You have the studies that say yes. You have the studies that are mixed and then there are no studies here in this pro search that say no. So the evidence is pretty clear leaning towards one side and then you have a summary with a really good overview. you know these are the factors that mental health effects are associated with then you get the limitations and nuance and conclusion. So that's already really good. You can ask follow-up questions like for example appeal we see that negative effects are especially associated with problematic or excessive use. So let's ask what counts as excessive use in these papers, right? And then you get the answer.

Several studies define excessive use quantitatively more than 5 hours per day or more than 20 hours per week. But what's really cool is you have this table down here where you get the studies and then you get the answer for each of these studies. So for example down here you see more than 5 hours per day. So going back up here this already gives me a really good overview.

How to do a literature review using AI

[2:58]

Now the really cool thing is that you can actually get a literature review which is almost like a customized meta analysis for your search. So let's try that out. Now it will start the search. You can even watch it think here.

It conducts an initial survey of the research based on my query. It basically defines the methodology of the meta analysis. Now it will execute all of these steps down here. Already it has identified 148 studies and that's the number screened. These are the eligible papers. And then at the end it will come up with the count of how many studies were included. So we're not going to wait for that because I've done that exact same search already before. So here we are. This is the finished literature review.

And you see it identified 1,026 studies. It screened 439 of them. It classified 393 as eligible and it included 50 papers in this assessment.

So again, we see social media is associated with negative effects on teenagers mental health, especially when use is excessive or problematic. This is pretty similar to what we got above in the pro search.

The consensus meter points into a very similar direction with 23 studies saying yes, three saying possibly, 12 are mixed, and two say no. It's very comparable to what we got above.

You see the methods.

You see a little graph down here of the number of studies that were identified, screened, eligible, included.

And then finally the results.

And then also down here by paper you can screen the key papers included in this assessment.

What did they look at?

What was the methodology, the sample size, and the key results.

You can also see the top contributors to this topic.

So this O'Reilly person contributed three papers. And then if you hover over this paper, you already see this is the paper that's included social media and adolescent mental health, the good, the bad, and the ugly. And then if you click on details, the sidebar pops up and you can look at several things. If the PDF is available, then you can scroll through the PDF here. You can check out the abstract of that paper and you can also get a study snapshot which I find particularly interesting because you can see at a glance what the study was about, what were the study participants, the methodology and so forth. And then you get a discussion and then the claims and evidence table.

So for example, problematic excessive use is linked to depression and anxiety and the evidence for that is strong. Then you see all the papers that support that. The green ones are the supportive ones. The orange ones are the studies that were mixed regarding that claim. And then the yellow ones say there is a possible association. And the red ones are the ones that say no, there is no association. Then you get a little matrix down here. And you can see that 18 studies linked general use to depression and anxiety. 14 linked problematic use to depression and anxiety, passive use, and so forth. And this is also a great way to spot if there are holes in our knowledge somewhere here, right? If there's no study in this section, then that could be an opportunity for you to tackle.

Filtering results to include only the most prestigious and highly cited journals

[6:40]

Now, another really cool feature is the fact that you can actually filter this review for various parameters. You can for example say exclude preprints please. You can say only include metaanalyses, randomized clinical trials and observational studies. Right? And then you could for example only include Q1 journals. Those are the most prestigious journals, the ones that really are the most highly cited and it's hardest to publish in those. Then we apply this filter.  So it basically rewrote the entire assessment based on these studies. We get a new consensus meter which looks very similar to the one we already had and the answer is still the same.

Exporting citations from Consensus AI

[7:34]

And what is also very cool, you can export the citations into whatever tool you're using: Endnotes or just export it as a CSV file.

Final thoughts and recommendations

[7:45]

So, I hope you can see that this tool can be really, really useful in your daily clinical practice. Also, full transparency, I have no affiliation with Consensus. There are no financial ties with the company. I'm really just recommending the tool because I'm using it a lot and I find it very helpful and I hope that you will get value out of it as well. And before you leave, please make sure to subscribe to our YouTube channel because that means a lot. I hope to see you again soon and check out our courses. Register for a free trial account at Metapmastery and learn the most important clinical skills. Stay tuned and be healthy. See you next time.