Академический Документы
Профессиональный Документы
Культура Документы
Indexing Guidelines:
Company/Organization Names
in PipelinePlus
Infodesk
Company Confidential
31 October 2013
Document Log
version
1.0
Date
10/31/2013
Comment
Indexing Guidelines
Company/Organization
Names in PipelinePlus
Author
Bonnie Snow
Table of Contents
Introduction: Overall Objectives
Page 3
Historic Names
7-8
Licensing Agreements
10
31 October 2013
The name used to identify a given organization in one source may differ from that used in
another database. Infodesk taxonomy establishes one preferred name for each organization
and maps each variant name found in individual pipeline databases to that preferred term.
Thus, when the user selects a preferred term from the company name list, all variations of
the organizations name are searched. The resulting list of drug records brings together
references to products associated with the same company, regardless of variant names used
to identify it in each source database.
Because Parent companies are not always linked with their subsidiaries or
divisions in pipeline databases, it is difficult to compile data on drugs in development under
the auspices of large corporations with many branches operating under different names.
Accordingly, a second major goal of Infodesk taxonomy is to identify corporate relationships
of this type in a hierarchical index. The View hierarchy option in the PipelinePlus (PP+)
company name index enables display of this portion of the taxonomy. In each Company
Tree, preferred terms for subsidiaries or divisions are indented under Parents, when
relevant. When a user selects a Parent companys preferred term, the resulting query will
automatically search all subsidiaries or divisions indented under it in the Tree. Similarly,
when a specific branch or child term is transferred to the Query Builder, any terms
indented under it will be searched. In effect, the hierarchical company taxonomy in
PipelinePlus represents pre-coordination of associated corporate entities and facilitates
retrieval of multiple, related organization names.
Another underlying goal in the indexing guidelines described in the following pages is increasing the
accessibility of data elements commonly searched by users of pipeline files. An important step
toward accomplishing this is presenting clean lists of company names in the Query Builder, without
the distractions of duplicate entries or synonymous name variations which seem to imply that
multiple terms must be selected to retrieve references to a single corporate entity. Users should be
able to find desired company names quickly and be re-assured, when viewing pertinent hierarchies,
that a thorough search of relevant synonyms and other related terms will be executed behind the
scenes in response to their selection of a PP+ Preferred Term.
31 October 2013
1. Choice of a Preferred Term for a Company When Several Name Variations Are Cited in PP+
Sources
Examples: Chimeracom LLC vs Chimeracom
Medeva LTD, Medeva PLC, Medeva Inc
1.1 As a general rule, the Preferred Term (PT) will be the name without qualifiers such as
Inc, A/S, AB, BV, SA, LLC, LTD, Limited, or GmBH, appended. Thus, in the example
above, Chimeracom would be the PT, with the LLC variation mapped to it. In the 2nd
example, Medeva would be the PT, with the three alternatives mapped to it.
1.1.1 Historically, selection of preferred terms for newly-added companies was
automated according to a source precedence rule (TP>RDI>PJB>RDF). Therefore, a
pre-existing preferred term automatically designated by the system when a company
name was initially cited in source databases is not always the appropriate PT per the
rule described above (1.1). When these discrepancies are discovered during
retrospective audits of the taxonomy, new preferred terms -- per the 1.1 guideline --should be added and other variations mapped to them.
1.1.2 Further note: To avoid inadvertent errors in mapping and introduction of new
Preferred Terms that are subsequently found to be variations/synonyms of existing
terms, it is essential that editors conduct thorough searches of the current taxonomy
before approving additions or revisions. If a name under investigation consists of
multiple words (i.e., a phrase, such as Celltech Therapeutics Ltd), a search in the
EditorialWEB should filter by the core keyword (Celltech) to locate possible name
variations and/or related companies that must be taken into account before
constructing a Tree hierarchy.
Examining the alpha and corresponding hierarchical lists of company names
in the PP+ platform is also strongly recommended. Possible errors requiring further
investigation are often easier to detect in the public platform than in the
EditorialWEB and their high visibility within the application prompts questions from
potential customers about quality control in our indexing.
1.2 Possible Exceptions to the general rule: Names that could be ambiguous without
qualifying terms to identify them as Companies. Example: Histogenics Corporation versus
Histogenics as a stand-alone term. Questions regarding possibly ambiguous names should be
raised for resolution in weekly editorial meetings.
31 October 2013
3. Historic Names
Examples: Theranox superseded by NB Therapeutics (which was formerly Nitric Biotherapeutics
or Nitric Bio).
Adeza Biomedical acquired by Cytyc Prenatal Products, which was a subsidiary of Hologic Inc.
Later, Cytec Prenatal Products was completely merged with Hologic Inc.
3.1 Unlike aliases, historic names cited in source pipeline databases need to remain
accessible in our indexes as Preferred Terms and need to be incorporated into, and
retained in, appropriate company hierarchies.
Following up the 1st example above, the Parent Preferred term would be NB
Therapeutics. Theranox and Nitric Biotherapeutics should each be designated as PTs and
indented under the Parent NB Therapeutics in the company name hierarchy. The
synonym Nitric Bio would be mapped to Nitric Biotherapeutics, but not be adopted as a
PT in its own right. In the 2nd example, both Cytyc Prenatal Products and Adeza
Biomedical would be retained as PTs and each indented as secondary-level nodes under
the parent company Hologic Inc. Use the Comment field in the EditorialWEB to
document the basis of your additions/revisions.
31 October 2013
31 October 2013
31 October 2013
developer, org Y as originator, etc.). To index all of these collaborations accurately would
require detailed review of all records where, for example, Charit is cited.
7.1.2 Examples: RDI drug record: Research programme: anti-prostate cancer therapeutics Stem Cell Therapeutics/The University of York
Another sample record: rhKGF-2 (ophthalmic, corneal injury/keratopathy), Wenzhou Medical
College/Guangdong Jida
Text: Wenzhou Medical College, in collaboration with Guangdong Jida, is
investigating an ophthalmic drop of recombinant human keratinocyte growth factor2 (rhKGF-2) for the potential treatment of corneal injury and keratopathy.
Neither of these collaborations should prompt creation of hyphen-linked entries in our
Company taxonomy, as theyre not recognized as new corporate entities. Since the
collaborators are searchable as individual companies associated with pertinent drug records
where they are cited, the PP+ search interface will enable users to find drug entries where
both names appear.
Note: There is currently no mechanism for ANDing together two or more selections from the
company name list in our Query Builder (multiple selections from the same index are
automatically ORed together). However, more flexible, user-defined logical combinations will
be implemented in future.
8. Licensing Agreements
Example: A.R. Kamm Associates licenses rights to market a drug developed by Recordati SpA
8.1 Should A.R. Kamm Associates be included in Recordati SpAs hierarchy (and vice versa)?
No. Both will be searchable as independent (root) Preferred Terms and would not be added
to the hierarchies (Company Trees) of their respective partners. The agreement did not result
in the formation of a new, legally recognized corporate entity, so Guideline 7.1 (above)
applies.
8.1.1 Rationale: We cannot attempt to associate all licensees with their respective
licensors in our indexes (i.e., pre-coordinate company relationships of this type
there are simply too many!).
Eventually we would like to offer the option to search a company/org name
as a licensee versus an originator or developer. Thus, it would be possible to search
for records where Recordati is the Developer and Kamm is the Licensee.
Currently, Licensees appear in the All Companies list, but only those also
identified in our sources as Developers will lead to results when selected from the
Developer Company list in the Query Builder. Thus, Kamm (a licensee) yields zero
results when searched as a Developer.
31 October 2013
31 October 2013
10