I am trying to query these databases and identify trials that passed, and those that didn't. I have gone through and parsed all of ClinicalTrials.gov and matched it to records from chEMBL to build a database.

The fields I have available to be for each drug are:

+ is_fda_regulated_drug: which is null in 99% of cases
+ overall_status: Terminated, Completed, Recruiting, etc. 
+ phase: "Phase 1", "Phase 2", ...
+ primary_completion_date: Date as a string or null
+ primary_outcome: Date as a string or null
+ primary_outcome: mess of text that is too inconsistent to parse
+ start_date: either date as a string or na
+ time_frame: mess of text and often not a timeframe at all
+ CHEMBL_ID: as string
+ FIRST_APPROVAL: a year, but this is first approval for any application, not the study in question
+ MAX_PHASE_FOR_IND: 1,2,3,4

So far the only definitive logic I've been able to define is:

if overall_status = terminated and phase != 4 then the trial failed

Suggestions?

  • What do you mean by "passed"? Successful vs not successful? – Grimaldi Mar 16 '17 at 14:58
  • @Grimaldi yes exactly. Passed clinical trials and was approved for use. – Will Manidis Mar 16 '17 at 15:08
up vote 3 down vote accepted

Drug approval is called "marketing authorization" and is not directly related to clinical trials. A program of clinical trials will usually lead to a marketing authorization in one or more countries.

You probably have to look somewhere else on FDAs website. Try Drugs@FDA or have a look at WhoDD to get from drug substance to marketed drug (and approval, maybe).

  • 1
    Do you know where that data is available? I have not found it in chEMBL or clinicaltrials.gov, nor found a relevant API call from openFDA. – Will Manidis Mar 16 '17 at 15:14

Like you, I am trying to make a list of failed drug candidates. DrugBank marks some drugs as Experimental, Investigational, or Withdrawn. It's harder than I expected.

Alternatively, given a drug name, it is possible to look it up in http://www.chemspider.com/, then check for the presence of a link to DrugBank with the CSID2ExtRefs API.

Chembl has a table MOLECULE_DICTIONARY which includes preferred name, is_therapeutic, first_approval, and a discontinued flag. It is available for download from here: ftp://ftp.ebi.ac.uk/pub/databases/chembl/ChEMBLdb/latest/ A join to COMPOUND_STRUCTURE will provide a SMILES. There is also a table of synonyms. By a couple of joins, you can also find information about patents - candidate drugs have probably been patented, so failed drugs have a patent but not first approval. (However, insecticides will also be in that list.)

Unfortunately there is no column for drug name in ClinicalTrials.gov. It is usually to be found in intervention.name - words in there that have no match in ChemSpider are not the names of chemicals at all. Are you willing to share your parsing code?

  • Can you add links to those? Thanks. – Jan Doggen Aug 21 '17 at 12:47

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.