Wikidata:WikiProject Clinical Trials/Curate
Curation is any activity which edits the data to improve it.
General advice
[edit]- If you are new to Wikidata then take a tour of the project starting at Wikidata:Community portal.
- Perhaps the single most important skill an editor can have is the ability to join conversations and ask questions. Be familiar with the forum at Wikidata:Project chat.
- Wikidata:Be bold in making edits.
- Routinely and publicly show edits and ask other community members for feedback
- The data here is complicated and editors will identify many challenges and content gaps. If an issue is important to you then log it in this project's talk page.
Curation for clinical trials
[edit]Disambiguation of people
[edit]Many clinical trials have associated people. Among them the principal investigator (P8329) is usually the individual most responsible for all aspects of the research, and may be useful to identify. Although this person's name should be listed in the clinical trial record, Wikidata only links to identified records of people and will not permit import of ambiguous name strings. Because of this, all people named in Wikidata clinical trial records must be disambiguated.
There is no easy standard process for personal disambiguation in Wikidata, although many editors do this in many contexts. Wikidata's most developed tool for disambiguation of people is the "Author Disambiguator" which has the intended use of uniquely identifying the authors of scholarly publications. In many cases, the lead researcher in clinical trials is also a person who publishes research articles, so disambiguating a person based on their publications may give enough insight to confirm their role in clinical trials.
- Try the tool: https://author-disambiguator.toolforge.org/
- See documentation Wikidata:Tools/Author Disambiguator
Disambiguation of research sites
[edit]Try to identify each research site (Q57951700) with the property research site (P6153). Uniquely identifying research sites is challenging. Research sites often do not have standard names, so different clinical trial records may call the same site by different names. Many research sites should have the property parent organization (P749), with the parent organization pointing to the research site with has subsidiary (P355). However, it may be that neither the name of the research site, nor other information in the original clinical trial record, nor other available information clarify what organizational affiliations a research site has. Try whatever makes sense and bring issues to the talk page.
Labeling medical topics
[edit]Most clinical trial records clearly indicate which medical conditions and study interventions they are examining. Some original records do not. Apply an appropriate determination process for identifying the medical concept and put it into the Wikidata clinical trial record with a property label such as medical condition treated (P2175) or research intervention (P4844). If there is ambiguity, then additionally, use the label object named as (P1932) as a qualifier on the Wikidata property, and import the ambiguous text used as the basis for determining the Wikidata label. ClinWiki (Q96777166) is an example of a community organization independent of Wikidata which organizes its own experts and patient stateholders to develop processes for label clarification; consider checking their methods.
Translation
[edit]Most clinical trial registries are in English language, but clinical trial participants and stakeholders come from every language community. Translating technical and medical terms are not easy, and often words have no good translation. Even so, Wikidata is a multilingual project with a global userbase in need of language translation. When organizing a translation project, try to collaborate with native language speakers and subject matter experts during the translation. After the translation, it may be useful to plan for long term stewardship and management of the terms, as sometimes questions can arise after years.
Method for translating Wikidata terms
[edit]The needs of translation projects vary widely, but for anyone translating terms in Wikidata, here is a process which works for many cases:
- Collect a list of terms in any language
- Match each term with its corresponding Wikidata item
- Put that list of Wikidata items into the Wikidata:TABernacle translation support tool
- Have a translator use the TABernacle interface to translate terms
- Seek Wikimedia community review from relevant stakeholders
Tools and processes for translation
[edit]The Wikimedia platform hosts multiple tools for translation. Here are some of them:
- mw:Content translation - This is the MediaWiki translation suite for managing translations across Wikipedia language versions. It contains various translation tools and options.
- Wikidata:TABernacle - This Wikidata tool which allows users to create a table for translating terms.
- translatewiki.net (Q9376349), an interface for translating the MediaWiki interface itself, and thereby localizing tools
- Translators Without Borders (Q3110900) - This organization has their own tools on their own website for recruiting and managing community review of translation. They have completed multiple cycles of medical translation for Wikipedia.
- Wiki99 - This Wikimedia community organization process models the selection of about 99 topics as a basis for Wikimedia community discussion and translation.
- en:Wikipedia:WikiProject Intertranswiki - This Wikipedia community project coordinates translation of content across participating language versions of Wikipedias.
Wikimedia organizations doing translation
[edit]While many Wikimedia organizations are willing to support translation to and from specific languages, the below organizations have long histories of promoting translation among all all languages for many topics. Consider collaborating with them or contacting them for advice on organizing translation.
Glossary of Common Site Terms, a ClinicalTrials.gov guide
[edit]ClinicalTrials.gov Glossary of Common Site Terms (Q106891230) is a publication of ClinicalTrials.gov. As it is a work of the United States Federal Government, the text is in the public domain. As a translation pilot WikiProject Clinical Trials organized translation of terms in the glossary in the context of the Wiki99 project. The list of terms is at meta:Wiki99/clinical trials, and translations are with each term. Languages which were part of the pilot translation were Hindi, Bangla, and Swahili. Thanks especially to Wikimedia Bangladesh for overseeing translation and discussing best practices.