WorldFAIR (D7.2) Population health resource library and training package
Bibliographic record
Abstract
This project, WorldFAIR – Global Cooperation on FAIR Data Policy and Practice, is funded by the European Commission's WIDERA coordination and support programme under the Grant Agreement no. 101058393. The project consists of 14 work packages, of which work package 7 (WP07) focusses on Population Health. WP07 is led by London School of Hygiene and Tropical Medicine working under the INSPIRE network. The work builds on the delivery of the Observational Medical Outcomes Partnership (OMOP) common data model (CDM) which includes funding by Wellcome (formerly Wellcome Trust) and IDRC Canada. The objective of WP07 is to develop a suite of methods and standards to provide the framework for the Go-FAIR principles for population health data. These standards form the basis of an AI-Ready description of data suitable for use by population health scientists, and understandable across domain and institutional boundaries. The first deliverable (D7.1) identified the Implementation Guide that could be used for population health data, and how it can be developed. This deliverable (D7.2) provides a step-by-step guide as to how to achieve the standards. The deliverable is aimed at population health scientists in low-resource settings, who know their own data and want to make those data FAIR. Population health uses many different tools to collect and manage data. One set of tools includes the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) and an OHDSI data analysis workbench that runs on top of it. The OMOP common data model has been used to harmonise and share data, and previous work has shown the tools needed to make OMOP data FAIR. Beyond the data themselves, the results from the analyses conducted on OMOP data can be used as indicators for the success of development goals, including the United Nations Sustainable Development Goals (SDGs). At each stage the tools, data, models and activities need to be described in a way that can be understood by other scientists and by computer search algorithms. This deliverable provides an introduction to the processes involved in making population health data FAIR in a pipeline that spans data collection through data analysis into an SDMX indicators database, and gives seven tutorials on what is needed at each step in this pipeline. It outlines the need to describe the study and the study context, how to use DDI Codebook and DDI Lifecycle with study data and how to use repositories like GitHub to make the metadata available. The next tutorials describe the extract-transform-load (ETL) process for putting the data into an OMOP CDM and the role of JSON-LD in preparing the data for machine searching in Schema.org in line with DDI-CDI. Together these tutorials give an overview of the steps in the OMOP processes which are a pipeline for the data, and how these steps can be performed and documented. Finally the tutorials show how predictive and causal analysis can be conducted and documented using the OMOP CDM and the OHDSI data analysis workbench and how the results can be integrated into an SDMX data cube, which would align with UN standards for SDG indicators. The deliverable does not provide detailed training for each step, but rather introduces the topic and clarifies the practical knowledge and skills that are needed to make this type of health data more FAIR. Note: Six of the tutorials are hosted on the WorldFAIR Vimeo channel, https://vimeo.com/user/91439529/folder/18642763, which provides functionality for playing videos. Alternatively you can download the tutorials and play in your local environment to experience full functionality. The remaining tutorial is accessible via this report. Visit WorldFAIR online at http://worldfair-project.eu. WorldFAIR is funded by the EC HORIZON-WIDERA-2021-ERA-01-41 Coordination and Support Action under Grant Agreement No. 101058393.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.029 | 0.070 |
| Meta-epidemiology (narrow) | 0.001 | 0.002 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.005 | 0.005 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.006 | 0.005 |
| Open science | 0.006 | 0.011 |
| Research integrity | 0.004 | 0.004 |
| Insufficient payload (model declined to judge) | 0.561 | 0.439 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".