Nested Knowledge is a tool that brings structure and AI features to systematic literature reviews. Traditionally, systematic reviews required careful record-keeping, long Excel sheets, and challenging collaboration. Nested Knowledge brings an innovative tagging structure and AI features, does away with Excel sheets and helps you screen giant databases like PubMed in seconds. It is a one-stop solution to organize your systematic literature review collaboratively. Get an exclusive 50% discount below.
Learn which tools can speed up your research: This free 21-day email series introduces you to a range of academic tools and how they work together to speed up the time needed for a literature review. Most of these tools have been around for less than a few years, some use AI. Knowing them gives you an almost unfair productivity advantage for your research.
Systematic Literature Reviews are reviews that follow a system. For example, it might define a few search terms and review all studies a database like PubMed finds for them. Then, these studies are filtered by well-defined criteria (for example, in medicine, the trials have to be double-blind and placebo-controlled). These studies (usually in the 100s) would have to be read and their results processed statistically. A question for a systematic review could be, “Does exercise increase testosterone in older men?”. As you can guess, reading 100s of studies is extremely laborious and can take months. This is the problem Nested Knowledge aims to address.
Nested Knowledge is a tool designed for teams of researchers or pharmaceutical companies that regularly need to go through these systematic reviews. In this post, we are going to dive into the basic steps of the systematic review and look into how Nested Knowledge makes it faster, easier and less error-prone. We will explore what part of it uses AI and which one is done manually as well as share some success stories.
Here is our workflow:
- Give Nested Knowledge a few search terms and use its AI capabilities to find more related keywords we might not have been aware of.
- Use Nested Knowledge to screen the studies for e.g. randomized controlled trials and include or exclude them. This process can be collaborative.
- Extract key concepts or tags from all of these studies to help us build the systematic literature review.
- Use Nested Knowledge to tell us where different concepts are mentioned in the abstracts of the selected papers so that we can screen through them more quickly.
- Visualize some of the results and look at how Nested Knowledge can assist in synthesizing these findings.
The result is not yet a finished systematic literature review, but the first step is to obtain the necessary studies and get a feel for the outcome and which topics need to be addressed. A scientist still needs to read the studies. Nested Knowledge will speed up a lot of tedium, like downloading studies or pre-screening them and, more importantly, replacing the numerous Excel tables with a powerful interface.
1. Set up a nest
Every systematic review consists of a number of search terms that produce studies and your decisions and insights surrounding these studies (i.e. include or exclude the study or its outcome). These bundles Nested Knowledge calls a Nest. To use the tool you will need to create a nest. Look at the buttons on the top right. Nothing but a name is required.
Notice that there is a “Demo Nest” option. This option loads a demo project containing a few dozen already screened and processed studies. They are the result of your work with Nested Knowledge. For these demos to make sense, I suggest understanding first how the tool works. Our nest will be called “TestNest”.
Setting up and Refining your Search Query
Every systematic literature review starts by defining search terms. Which terms you choose defines how many studies will be retrieved. Choose too few and you don’t have enough statistical power to make a strong statement, choose too many and you end up with possibly 1000s of studies to go through. This is where Nested Knowledge can help. First, start with a single term or a number of broad terms surrounding your topic. Click “Search Exploration”, type in a term and hit enter. When you are don, click “Refresh Exploration” to get started.
The number of results the search will retrieve on Pubmed is shown on the very right below the Preview tab. Adjust your search terms until the number of studies seems reasonable. In this example, I want to start an investigation into the role of the G3BP1 protein in the formation of stress granules and include only these two terms. Nested Knowledge will retrieve the first 250 studies and search their abstracts for keywords, topics, and other criteria after you have clicked “Refresh Exploration.”
Nested Knowledge can extract a number of topics from the studies and sort them by frequency:
- PICO: Keywords in the categories Population, Intervention, Outcome. In my example the population are different model organisms (e.g. rats, mice, yeast, etc). This makes sense primarily for medical research.
- Topics: Keywords most frequently mentioned in these studies. Here, fore example I can learn about a keyword “heat stress tolerance” which can give me a clue what to look at in my review.
- Acronyms: This helps you dive into a new field as well as identify top-acronym-keywords. In my case, “LLPS” is a common acronym that might help me broaden my review scope.
You can also look into locations, sizes and types of studies. Not all these analyses make sense for every domain, but some will surely apply to yours. The result of this search should be a search query that gives you a good amount of studies for your systematic literature review. Nested Knowledge provides quantitative estimates for every keyword it displays and this gives you a feel about how important and/or studied certain topics are.
Screening of Studies for Systematic Literature Review
The next step in your systematic literature review is to screen the studies and include or exclude them. In a systematic review a reason for exclusion must be recorded. The first step is therefore to define a set of reasons (think of them simply as tags). To add reasons click on the cog icon next to “Screening”, select the “Exclusion Reasons” tab and add new reasons.
Next, we can start looking through the studies. The abstracts of all studies are pulled in automatically, and for most studies, you can open the PDF with one click as well. Click on “Screening” to get started. Nested Knowledge will keep track of which studies you included or excluded and if you can stop and resume as needed.
The main screen displays the abstract with AI-generated highlights. Nested Knowledge is designed with a medical audience in mind; therefore, by default, it highlights three categories: Population (e.g. 50 healthy adults), intervention (e.g. a drug they took) and outcome (e.g. disease risk decreased). These highlights can help you decide whether to include or exclude the study. A progress counter at the top right shows how far you are.
Sometimes, it is useful to define your own keywords, and Nested Knowledge will automatically look for these in the abstract. This is done analogously to adding exclusion reasons:
Notice how some words in the abstract above were highlighted in bright magenta. These were the custom keywords. A few other features, like adding comments to each study, can be used for collaboration with colleagues.
Using AI for screening studies
After you have scanned at least 50 studies and included 10 or more, you can train a small AI model that can predict the probability of a study being included based on the abstract:
Tagging the studies
The final step of your literature review is to tag the studies you have screened and included based on the criteria of your systematic literature review. Tagging is the “systematic” part of the literature review as you label each study according to what it studied, found and concluded. Of course, every domain will have its own set of tags and labels. Picking the right number of tags is an art in itself. Here is an example from the medical field following the PICO (Patient, Intervention, Outcome) structure:
Tags, in this example, can follow a hierarchy: Patients are broken down by age, sex, BMI and so on, while interventions are broken down into the various drugs the patients were prescribed, and similarly, outcomes can be broken down into various types of reactions patients had using the prescribed drugs.
The key is to have a good tag structure in place before you start tagging, as every time you significantly change your structure you might need to re-tag existing studies. Doing this step well, can make or break your systematic literature review and greatly help you synthesize the information, let’s look at this next.
Literature Review Synthesis with Nested Knowledge
The tag structure and the papers allow us to ask very specific questions like “Does prescribing a placebo drug (an intervention) to elderly patients (age, a patient characteristic) lead to Hypotension (an outcome)?” . The three highlighted words are tags, and Nested Knowledge can pull out these studies automatically. Click on Synthesis in the menu (it will change to a new view) and then select “Qualitative”. Your tags are now displayed as a sundial, allowing you to assess how many studies are assigned to each tag. To answer the question, we must select the tags: Age, Placebo and Hypotension. The result at the bottom right gives five studies we can dive deeper into to find an answer to this question.
Once you get the idea behind Nested Knowledge’s tagging system, you will quickly fall in love with it.
More Features
This article showed only the basic features and the most minimal workflow with Nested Knowledge, as it is a complex tool. For example, if you extract the data from each study and add it to Nested Knowledge, you can use its statistical analysis capabilities:
Some other features you might find useful and worthy of exploring:
- Collaboration with others
- Manuscript writing with others
- Commenting on studies
- Automatic Import of PDFs for screening and full-text screening
- Defining Questions, Objectives, Level of Evidence and many other properties that might be relevant to your systematic literature review
Who uses Nested Knowledge
Nested Knowledge is designed for medical professionals and pharmaceutical companies but also caters to academics. Dr. Yasmin Aziz, for example, used Nested Knowledge to make her stroke research easier and faster. She found that unnoticed bleeding after stroke treatments can worsen outcomes and uncovered why many stroke trials end early. The software made data analysis quick and efficient, helping her share important insights with the medical community.
Nested Knowledge Pricing and Discount Code
You can start using Nested Knowledge for free. If you don’t need collaborative or advanced features, this might almost be sufficient for a small review. For the full product academics pay $ 20 per user monthly, 15x cheaper than businesses.
Use the discount code “ILYA50” for a 50% discount on your first month.
Summary of Nested Knowledge
Nested Knowledge is an all-in-one solution for organizing systematic literature reviews aimed at the medical field. The steps to conduct a literature review are:
- Identify a set of search queries that provide you with any number of studies. Nested Knowledge helps with providing suggestions for search queries, assessing their volume and pulling the studies from Pubmed.
- Screen the resulting studies and note the reasons for excluding them. Nested Knowledge pulls studies automatically and screens their abstracts (with AI) using the PICO framework or user-defined keywords. For large literature reviews (50+ studies), an AI model can be trained to estimate whether a study is to be included or excluded based on its abstract.
- Tag the study contents, e.g. like population, interventions and outcomes (PICO framework) or any other system you want to use. This group studies by their findings/methods/domains etc. Nested Knowledge can visualize how many studies belong to which tag and make it easy to find studies answering a specific question.
Systematic Literature Reviews are traditionally organized with numerous Excel sheets and are difficult to do collaboratively. Nested Knowledge saw this problem and created a solution for it. Academics pay $20 monthly (use the code ILYA50 for a 50% discount) but small projects will fit into the free account. If you have a systematic literature review ahead of you, bring some patience to learn how this tool works and try it.
