Dimensions has integrated the CSO classification from the International Cancer Research Partnership (ICRP)

The integration of a classification system for cancer research into Dimensions Analytics offers new ways to track and understand cancer-related activity across publications, grants and more.

Digital Science has been working closely with the International Cancer Research Partnership, ICRP, which partners with governmental, public, and non-profit cancer research funding organizations around the globe. ICRP itself maintains the only worldwide public source of current and past cancer research grants (more than $50 billion has been awarded since 2000 alone).

Finding a common language

For many years, the stumbling block for funders like ICRP, researchers and institutions was that there was no clear way to categorize the cancer-related grants awarded or the findings they funded.This made conducting any kind of meaningful data analysis a challenge.

In response, the ICRP launched the CSO – Common Scientific Outline – coding system; it contains 34 codes organized into six broad areas of scientific interest in cancer research:

  • Biology
  • Etiology
  • Prevention
  • Early Detection, Diagnosis & Prognosis
  • Treatment
  • Cancer Control, Survivorship & Outcomes

The CSO is complemented by a series of codes describing the cancer type. The codes for both can be applied to many forms of data – in the case of publications, for example, they are determined by looking at the text the researcher has used to describe their research, e.g. the award title, lay and/or scientific abstract and keywords (where available).

The big advantage of the ICRP CSO system is that it provides a common language for comparing and contrasting the research portfolios of agencies, whether they are public, philanthropic, or government-run.

CSO and Dimensions

Some years ago, Digital Science and ICRP began working together to integrate the CSO classifications into Dimensions. Initially, adding the coding to data was a manual process.

We then drew on machine learning technology to automate classification; training and refining the algorithm with pre-coded documents and the input of ICRP subject matter experts.

Unlike other databases, Dimensions doesn’t use a journal classification approach, which has allowed us to include non-journal related records, such as grants.

As a result, we’ve now taken the next logical step and applied the ICRP CSO coding to all cancer research-related records among the 100 million publications, 2.4 million grants, clinical trials, policy documents, and patents that Dimensions is home to.

A new phase for Dimensions Analytics users

Users of Dimensions Analytics are now able to explore and understand cancer research-related funding and publications in a much deeper way. For example, they can compare the activities of funders, universities, countries, and researchers (see figures 1 and 2).

Figure 1: A Dimensions screenshot contrasting CSO-categorized publications by researchers at the University of Oxford and University of Cambridge in the UK. Oxford is more active in the area of ‘Systemic Therapies / Clinical application’, while Cambridge is more dominant in the early stages and genetic causes of Leukemia development, e.g. ‘Cancer Initiation: Alterations in Chromosomes’.
Figure 2: Select UK universities and their outputs in the CSO categories relevant to clinical treatment of Leukemia – important information for patients.

How to filter using the new ICRP categories

Whether searching publications, grants, patents, clinical trials, or policy documents, Dimensions Analytics users will find a filter called ‘Research categories’ in the left hand column. Clicking on it opens out a list of six options (see figure 3). These include the two new ICRP-related filters:

  • ICRP Cancer Types: Lists more than 60 cancer types determined by a coding system linked to the World Health Organization’s International Classification of Diseases.
  • ICRP Common Scientific Outline (CSO): The 34 CSO codes organized into six broad categories of scientific interest in cancer research.

Users can filter and browse on these ICRP categories and their sub-categories, or use them to refine search results.

Figure 3: The two new ICRP classifications appear under ‘Research categories’ in the left hand column.

The ICRP classifications are not the first field-specific categories integrated into Dimensions Analytics – previous additions include the NIH’s RCDC system and the HRCS system from the UK.

For more information on Dimensions Analytics please contact us.