A Data Package which is a database of public and private sources of information on health conditions related to clinical trials.
How will we take info from these multiple resources to create a new, normalized list of health conditions?
Normalized lists of health conditions have been created by the Unified Medical Language System (https://www.nlm.nih.gov/research/umls/quickstart.html). The UMLS maps between various “controlled vocabularies” e.g. MeSH, SNOMED-CT and ICD-10. This is free to use, but come with license terms.
My recommendation as a first pass is to adopt MeSH as the primary database of health condition terms. It is freely downloadable as XML. It is directly used by ClinicalTrials.gov and therefore will cover a large amount of the clinical trials we will have in OpenTrials.
As an example, the CT.gov trial NCT02640105 has the following XML bits:
<condition>Cluster Headache</condition>
...
<condition_browse>
<!-- CAUTION:
The following MeSH terms are assigned with an imperfect algorithm
-->
<mesh_term>Cluster Headache</mesh_term>
<mesh_term>Headache</mesh_term>
</condition_browse>We can look up “Cluster Headache” in the MeSH database: