Sign up

Why You Should Use Tagging Hierarchies?

Peace Olaniran

Peace Olaniran

December 6, 2022

How Nested Knowledge users harmonize data, structure study content, and build expert knowledge into the platform.

When reviewing the medical literature, the general set of steps are usually constructed as:

  1. Search for studies (on PubMed, Embase, and similar),
  2. Screen to include relevant articles (sometimes with multi-step or dual screening methods),
  3. Extract data (for qualitative summary and/or meta-analysis).


At Nested Knowledge, we offer AI-assisted Search strategy and Screening in multiple modes to get you through steps 1 and 2 above. However, we recognize that the success of a project depends on whether the study design is reflected in the concepts extracted from underlying studies. Therefore, for step 3 we have built out a unique feature offering: customized, template-driven Tagging Hierarchies.

What is a Tagging Hierarchy?

Tagging hierarchies are a well-studied method of organizing concepts for collaborative research. The goal of a hierarchy, when compared to ‘flat’ tags, is to show both (1) connections between concepts and (2) structural organization. By layering category tags, such as “Patient Characteristics”, over specific ‘child’ tags, such as “Age (Median)”, users both establish a link between Patient Characteristics and Age, and also identify the former as the category and the latter as the concept of interest.

When this practice is generalized across a study design, the tagging hierarchy becomes the method for categorizing and showing relationships among concepts. For example, in our Tagging Configuration page (shown below for the Basilar Artery Stroke review published by a collaboration led by Stanford neurosurgeons), a typical hierarchy may include general categories such as:

  • Study Characteristics, such as Study Type;
  • Patient Characteristics, such as comorbidities and demographic information;
  • Interventions, reflecting the treatments applied to patients;
  • Outcomes, such as efficacy and safety endpoints in underlying trials.

What are the advantages of Tagging Hierarchies?

Tagging Hierarchies have several advantages over ‘flat’ tags, and Nested Knowledge has built on top of this to provide additional, unique capabilities in the review process:

  1. Hierarchies structure relationships: Without hierarchies, applying tags to underlying records has limited value, since it only helps you figure out ‘is X concept here, yes or no?’ Searching for general concepts, drilling down on related ideas, and displaying the relationships among them is made possible by hierarchies.
  2. Hierarchies make reading full texts easier: By building a hierarchy in advance, you are setting yourself up to easily search and apply tags in underlying studies. Either by structuring a query directly into the hierarchy, or by using our Tagging Recommendations to automatically jump to the concepts of interest in the full text, your concept identification can follow the structure you lay out in study design.
  3. Hierarchies build institutional knowledge: Every time you build a hierarchy in Nested Knowledge, you can turn it into a Hierarchy Template and share it with your organization. That way, you never need to rebuild the same hierarchy, and can customize from the templates to reflect the specific research question(s) of interest.
  4. Hierarchies enable data visualization: Once you have built your hierarchy and applied tags to underlying studies, the very fact that you provided the structure from the start enables Nested Knowledge to automatically create your data visualizations. For example, in the Basilar Artery Stroke study above, the authors published the Qualitative Synthesis visual embedded below. This visual helps your users drill down on concepts of interest simply by filtering your hierarchy; it also allows you to add Insights that guide your researchers to key concepts and studies!
  5. Hierarchies enhance replicability of science: Scientific experts have long recognized that replicability is vital to experimentation. A key methodological step in designing studies is identifying endpoints, and to ensure replicability, many organizations (led by the NIH) have stressed the importance of Common Data Elements (CDEs). CDEs are trial endpoints that have identical definitions, timepoints, and measurement techniques across related studies, thus enabling apples-to-apples comparisons of interventions reporting CDEs. The same Stanford collaborators actually published a review that employed the Nested Knowledge hierarchy to determine the level to which current trials report similar endpoints in chronic subdural hematoma interventions, and noted that before this tool, choosing endpoints in any trial would lack any structured method of examining existing commonly-reported data elements. Thus, with hierarchies, you can build out the set of common concepts and bring your discipline toward replicable reporting of outcomes.

How does the Hierarchy help with quantitative analysis?

On top of the advantages of Tagging Hierarchies for finding concepts, Nested Knowledge offers the additional benefit of employing your Hierarchy for quantitative Extraction. After you have configured and applied tags (thus creating your Qualitative Synthesis diagram to help readers drill down on concepts), we know you may move forward with meta-analysis.

Unlike any other tagging system in the systematic review universe, Nested Knowledge’s Hierarchies are also the jumping-off point for automated biostatistics. Specifically, if you build a Hierarchy that contains your Interventions of interest, and if you identify which tags are Data Elements (and the statistic type, such as “Mean with Standard Deviation”), we will automatically apply Network Meta-analytical (NMA) statistics to your dataset and visualize it in Quantitative Synthesis. Because you created the structure underlying these concepts, you can even run new NMAs on any Data Element, at any level of your Intervention Hierarchy, just by selecting them in the Quantitative Synthesis page.

How to get started?

If you are looking to get started building hierarchies in Nested Knowledge, you can get started by Signing Up on our homepage; you can even use our Demo Nests to practice hierarchy-building in a risk-free environment. If you’re already a user and want to learn more about the capabilities Tagging Hierarchies enable, you can see our Documentation for video guidance. If you run into any issues, don’t hesitate to contact our Support team for live video-conferencing support!

Have a question?

Send us an email and we’ll get back to you as quickly as we can!