NCIP / clinical-trials-search

Other
19 stars 7 forks source link

Interventions not being populated for some studies even though the XML api and CT.gov record displays them #7

Open gvrajan opened 7 years ago

gvrajan commented 7 years ago

The JSON document returned for certain studies seems to blank out the "arms"."interventions" element for certain studies even though that study record on clinicaltrials.gov and it's downloaded XML record from the old clinicaltrials api clearly shows the existence of interventions. It does not happen for all studies . To reproduce this issue use the following studies' NCT numbers as examples - 1. NCT02381314 - download and compare the documents extracted from https://clinicaltrialsapi.cancer.gov/v1/clinical-trial/NCT02381314 and https://clinicaltrials.gov/show/NCT02381314?displayxml=true 2. NCT02573324 - download and compare the documents extracted from https://clinicaltrialsapi.cancer.gov/v1/clinical-trial/NCT02573324 and https://clinicaltrials.gov/show/NCT02573324?displayxml=true

You will notice that the "arms"."interventions" in the JSON document version is empty but the XML version is populated. I saw a few other study ids showing this problem and supply if needed.

NCT02542202 is an example of a study where both JSON and XML versions are correctly populated. We use these to enhance the tags on our internal study database and would appreciate a fix or some workaround advice. Thanks

loosed commented 7 years ago

Hello,

Thank you for identifying this issue. You may be looking for a distinction between NCI CTRP data exposed through the NCI's Clinical Trials API and the National Library of Medicine's (NLM) ClinicalTrials.gov system. The NCI's Clinical Trials API uses data from the Clinical Trials Reporting Program (CTRP) database, a database of NCI-supported cancer clinical trials. On most trials, all information is abstracted and available via the API. However, CTRP contains data on industrial trials (i.e. pharmaceutical company trials) which is imported from ClinicalTrials.gov and NCI-designated Cancer Center participating sites are added. When trial records are imported from ClinicalTrials.gov to CTRP, interventions are not imported because CTRP uses NCI Thesaurus for an intervention ontology. We've noted this as a feature request, we will look into abstracting interventions by arm on industrial trials imported from ClinicalTrials.gov.