Docket Article
ClinicalTrials.gov FDA BMIS Crosswalk: CRO, Sponsor or Site?
A 2026 ClinicalTrials.gov FDA BMIS crosswalk for PE research: distinguish CROs, sponsors and sites, reconcile submission observations with legal companies, and retain role, date and ownership gaps.

Summary
- 01Treat study and submission records as observations requiring review. People, facilities, legal companies and corporate groups are distinct objects, and repeated observations do not automatically increase company counts.
- 02Preserve roles in context: a collaborator listing establishes support, a facility identifies a location, and BMIS field meanings depend on TYPE. None alone establishes a current, independent service company.
- 03Qualify legal identity, current services, facility operators and ownership separately. Allow multiple supported roles and retain unresolved alternatives rather than forcing a company classification.
- 04Keep source event, public update, snapshot, company evidence, registry and review dates separate. A source's download cadence does not establish when ownership or contractor relationships were verified.
- 05Publish counts with their extraction boundaries and variable-specific denominators. The historical trial study and the article's hypothetical review batch do not estimate the present CRO universe, revenue or market coverage.
Inside this article
- 01Executive Summary
- 02Introduction and Background
- 03Key Changes
- 04Source Selection and Role Crosswalk
- 05Implementation Considerations and Process Changes
- 06From Observations to Current Companies
- 07Data Analysis and Evidence
- 08Case Studies and Real-World Examples
- 09Implications and Future Directions
- 10Frequently Asked Questions (FAQs)
- 11Conclusion
Executive Summary
A ClinicalTrials.gov FDA BMIS crosswalk should connect documented research roles to reviewed company identities. It should never turn every name into an acquisition target. ClinicalTrials.gov contains sponsor- or investigator-supplied study information; the US Food and Drug Administration's Bioresearch Monitoring Information System (BMIS) contains submission-referenced clinical investigators (CIs), contract research organizations and institutional review boards. FDA's non-comprehensive scope covers roles mentioned on Forms FDA 1571/1572 submitted to its Center for Drug Evaluation and Research (CDER) since October 1, 2008, for specified drug and therapeutic-biologic Investigational New Drug studies. [1] [2]
For private equity (PE) origination, the central distinction is between an observation, a person, a facility, a legal company and a corporate group. Under US drug Investigational New Drug (IND) rules, a contract research organization (CRO) assumes sponsor obligations as an independent contractor; an investigator is an individual. Written transfers determine the obligations assumed. These definitions describe responsibilities in context, not permanent company categories. Preserve multiple supported roles for the same company across studies and time. [3] [4]
The Aggregate Analysis of ClinicalTrials.gov (AACT) database offers relational registry data and states that it downloads the registry daily. Extraction frequency does not establish when a contractor relationship or corporate owner was verified. [5] As checked for this report on November 15, 2026, the BMIS search page displayed an introductory update date of May 10, 2022, and a footer date of December 01,2025. [6] Neither label verifies the downloadable extract's current contents. A published 2023 methods article similarly distinguishes its monthly AACT download from the source's daily cadence. [7]
The recommended output is a role ledger with retained source references, raw names, dates, legal identifiers, match evidence and unresolved questions. Research Organization Registry (ROR) identifiers need not represent legal entities; Legal Entity Identifier (LEI) parent relationships concern accounting consolidation and permit reporting exceptions. Neither identifier alone establishes an independent acquisition candidate. [8] [9] [10] The practical decision is pass, exclude, hold or escalate, supported separately for role, identity, current services and ownership. Trial totals and submission totals are evidence about their defined extracts, not revenue, backlog, market coverage or sale intent.
Introduction and Background
PE teams researching clinical services need a usable company universe before they can evaluate potential investments. Public trial information is useful for discovering names and relationships, but the unit being counted must remain explicit. A study identifier connects information about a study. It does not identify the legal company that operates a facility, employs an investigator or provides contracted services. ClinicalTrials.gov assigns an NCT number to each study record. [11]
The mandate should therefore specify the intended service business before extraction. A CRO search asks who supplies contracted research activities. A trial-site search asks who operates research locations. A site-management organization (SMO) mandate asks who manages sites, with the qualifying services defined in the investment criteria. A laboratory mandate asks a different question again. Form FDA 1572 information distinguishes investigators, research facilities, clinical laboratories and institutional review boards (IRBs), illustrating why adjacent clinical functions should not be collapsed. [12]
This report proposes an observation-to-company method for clinical research private equity market mapping. It is a source-reconciliation protocol, not an acquisition-target list. Docket's position is that of a target-research provider; the proposed evidence requirements apply equally to an internal origination team and an external research supplier.
The jurisdiction also matters. The US IND role definitions discussed here do not define every device study, observational project or research arrangement worldwide. The European Medicines Agency describes the Clinical Trials Information System (CTIS) in relation to European Union and European Economic Area authorities. Its records belong in a separately scoped research layer, rather than being treated as interchangeable with US IND observations. (Source: www.ema.europa.eu)
The crosswalk below is a proposed analytical model. This session did not establish a public NCT-to-BMIS submission key. Names and addresses therefore generate candidate matches for review, rather than a verified deterministic join between the systems.
The decision sought is modest but useful: does this observation justify further research into a service company, and what evidence must be obtained before that company can enter a current, independent-company shortlist?
Key Changes
The changes below are changes to the research method proposed in this report, rather than a claim that either source system has recently changed its regulatory purpose.
Replace company labels with record-specific roles
Start with the role actually supported by the source. The US IND sponsor definition concerns initiation and responsibility; the CRO definition concerns assumed obligations under an independent-contractor arrangement. A business description such as “clinical research” does not settle which role applies to a particular observation. [3]
Use separate fields for observed role and mandate classification. The former describes the source evidence. The latter records the team's reviewed conclusion about whether the business fits the investment thesis. A company may warrant multiple classifications when the evidence supports them. Do not encode a rule that a sponsor name can never also refer to a service provider in another context.
- Sponsor evidence: Retain the study-specific responsibility.
- Contractor evidence: Identify the contracted activity and its source.
- Site evidence: Identify the operator separately from the location.
- Laboratory evidence: Preserve the laboratory relationship separately.
- Ambiguous evidence: Leave the role unresolved rather than forcing a category.
This approach also avoids equating funding with sponsorship. The World Health Organization (WHO) states that a primary sponsor may differ from the primary funder. (Source: www.who.int)
Replace implied contracts with documented relationships
Section 312.52 of the US IND rules requires transfers to be described in writing; partial transfers must identify the obligations assumed. Publicly co-occurring names do not reproduce that agreement. [4] Consequently, a sponsor and a collaborator appearing on the same study should generate a relationship to review, not an asserted sponsor-to-CRO contract.
The International Council for Harmonisation (ICH) E6(R3) guidance also addresses agreements documenting activities transferred to service providers. Its framework supports retaining the activity and agreement context rather than treating every service provider as the same operating model. [13]
Replace activity counts with defined denominators
A study can involve many sites, and a company can operate multiple establishments. The Census Bureau's archived establishment definition explicitly distinguishes an establishment from a company or enterprise. [14] Accordingly, the research output should show which objects were counted, which were merged after review, and which remain unresolved. Increasing the number of observed locations should not automatically increase the number of companies.
Source Selection and Role Crosswalk
Table 1 summarizes the evidence layers. Its publisher row identifies a downstream research position, not a regulatory data source.
| Evidence layer | Observation and scope | Dates and access | Appropriate research use |
|---|---|---|---|
| ClinicalTrials.gov | Worldwide study coverage; sponsor-supplied or investigator-supplied information; incomplete coverage. [1] | Public study information and application programming interface (API) metadata. [15] | Discover study-linked names and locations; review role before company classification. |
| AACT | Relational representation of ClinicalTrials.gov protocol and results data. [5] | Published daily download cadence; cloud queries and database downloads. [5] | Reproducible extracts after checking the actual schema and snapshot. |
| FDA BMIS | CI/CRO/IRB observations from specified IND submissions; repeated entries. [2] | Published quarterly cadence; search and ZIP download routes. [2] | Role-aware discovery, with submission scope and file freshness recorded. |
| Company and identifier records | Legal identity, organizational identity or person affiliations, depending on the system. [16] [17] | Record-specific histories and collection dates. | Resolve identity using the appropriate jurisdiction and identifier type. |
| Docket, publisher research layer | Sourced target research positioned above the public-record layers. | Apply the same evidence requirements to the research handoff. | Require supported company conclusions and visible unresolved questions. |
The layers answer different questions. AACT can make registry analysis more convenient, but it does not add independent ownership verification. A legal registry can support identity without showing that an entity currently sells the required clinical service. Keep those evidentiary jobs separate.
Read BMIS TYPE before SITE
BMIS TYPE distinguishes CI, CRO and IRB; older CI coding includes DEM. For a CI, SITE means the study site; for a CRO or IRB, it means the organization. RECEIPT DATE describes FDA receipt of Form 1572. Preserve NG, ?? and 01-JAN-1900 as raw values while applying field-aware missingness rules; FDA treats the last value as invalid. NG in STATE can indicate a foreign address. [2]
The operational consequence is a branching parser, not a universal “SITE equals company” rule. Route person observations to investigator review, location observations to operator review, and CRO observations to legal-entity review. Negative BMIS searches cannot establish that a business is not a CRO, given FDA's stated incomplete scope. [2]
Separate ClinicalTrials.gov relationships
A collaborator supports a study; support alone does not establish a CRO contract. A facility identifies where a study occurs, and the responsible party concerns submission and updating. These concepts should occupy separate columns. [11]
The current API metadata verifies lead-sponsor name and class fields and collaborator name and class fields within the sponsor/collaborator module. Those fields provide source roles, not ownership keys. [15] A name appearing in a role field is therefore a candidate identity observation, subject to further review.
- Study information is supplied by sponsors or investigators.
- A collaborator supports a study; the listing alone does not establish a CRO contract.
- A facility identifies the study location; the responsible party concerns submission and updating.
- TYPE distinguishes clinical investigators, CROs and IRBs; older investigator coding includes DEM.
- SITE means a study site for an investigator, but the organization for a CRO or IRB.
- Observations cover specified IND submissions and include repeated entries.
Names and addresses generate candidate matches for review. The report did not establish a public NCT-to-BMIS submission key.
The key design choice is reversibility. A corrected match should update the reviewed identity link while leaving the original observation intact.
Implementation Considerations and Process Changes
A reproducible extract needs a source contract: which system was used, which fields were available, how the extract was selected and what the team has actually tested.
Establish the extraction contract
Use a documented mandate and a fixed extraction boundary. This report recommends recording:
- Service scope: The contracted activities or site operations that qualify.
- Geographic scope: Operating locations and legal jurisdictions separately.
- Study scope: Included study types and regulatory contexts.
- Source version: The retrieved page, file or API metadata version.
- Extraction time: When data were actually collected.
- Selection rule: Included roles, dates and geography filters.
- Access result: Whether a page, file or query was successfully read.
- Known exclusions: Missing source fields and untested access routes.
The official API metadata is useful for checking field meanings before implementation. It distinguishes study submission dates from subsequent updates. An extract should preserve those distinctions rather than renaming every timestamp “last active.” [15]
AACT offers current and past database versions, supporting snapshot-based comparison. Select and record the version actually used; do not describe a historical local copy as today's registry merely because the service advertises daily downloads. [5] The published methods example demonstrates that a downstream pipeline can deliberately use a different collection cadence. [7]
Preserve separate clocks
Maintain independent timestamps for the observation, publication, extraction and corporate conclusion. ClinicalTrials.gov's last update posted concerns public availability, whereas last verified concerns confirmation of the study information. Neither is an ownership-verification date. [11]
For each candidate, retain:
- Source event date: The receipt or reported study event.
- Public update date: When the revised information became available.
- Snapshot date: The copy actually analyzed.
- Company evidence date: When service evidence was collected.
- Registry date: The corporate record's relevant filing or status date.
- Review date: When the analyst accepted or withheld a conclusion.
If current company evidence conflicts with a historical trial location, preserve both with their dates. An apparent contradiction may reflect a rename, acquisition, relocation or closure; identifying which explanation applies requires further evidence. GLEIF's index contains historical as well as current LEI records, a useful reminder that retained identity history and a current conclusion are separate objects. [9]
State this report's access limitations
For this session, the BMIS search interface and field notice were readable, but no search-result submission was tested. The download could not be decoded through the permitted fetch route, so no actual file layout, row count or extract date is represented as verified. The search page's inconsistent displayed dates remain unresolved. [6]
ClinicalTrials.gov API metadata was retrieved; a study-result extraction was not tested. AACT's landing page was readable, but its attempted data-dictionary route returned an error and no authenticated database was queried. Therefore, this report supplies a logical ledger rather than purportedly executable AACT SQL. The advertised cloud and download routes remain distinct from access successfully exercised in this research. [5]
From Observations to Current Companies
Entity resolution should produce a reviewable connection between a source observation and a company, with explicit alternatives when the evidence is insufficient.
Table 2 gives a compact field contract for a copyable role ledger. It is a proposed research schema, not a claim that either source publishes all these columns.
| Field group | Fields to retain | Interpretation rule |
|---|---|---|
| Provenance | observation_id; source_system; source_snapshot; source_record_reference; study_id | Keep the native reference and local observation key separately. The NCT identifier belongs to the study. [11] |
| Observed role and identity | role_type_raw; name_raw; address_raw; receipt_date; reported_date; update_date; missing_value_flag | Preserve source values before normalization; a role is attached to an observation. |
| Reviewed identity | candidate_entity_id; legal_name; jurisdiction; registry_identifier; identifier_type | Identify whether the key denotes a legal entity, research organization or person. ROR need not identify a legal entity. [8] |
| Operating relationships | current_service_evidence; facility_operator; company_site_group_relationship; effective_from; effective_to | Separate affiliation and ownership; ORCID affiliations describe person-to-organization relationships. [17] |
| Review and disposition | match_rule; supporting_urls; reviewer; review_date; unresolved_issue; disposition | Preserve the accepted match, competing explanation and next evidence required. |
The key design choice is reversibility. A corrected match should update the reviewed identity link while leaving the original observation intact. That allows historical counts to be reconstructed without rebuilding the extract.
A CSV header can be copied directly into a worksheet:
observation_id,source_system,source_snapshot,source_record_reference,study_id,role_type_raw,name_raw,address_raw,receipt_date,reported_date,update_date,missing_value_flag,candidate_entity_id,legal_name,jurisdiction,registry_identifier,identifier_type,current_service_evidence,facility_operator,company_site_group_relationship,effective_from,effective_to,match_rule,supporting_urls,reviewer,review_date,unresolved_issue,disposition
Resolve identity without erasing evidence
Normalize spacing, punctuation and address components into additional fields. Retain the original strings. The SEC's Central Index Key (CIK) guidance notes that filer names can appear differently than expected, supporting explicit alias review rather than a silent string replacement. [18]
A defensible matching sequence is:
- Generate candidates: Use names, locations and known aliases.
- Check object type: Separate people, organizations and facilities.
- Qualify the jurisdiction: Identify the relevant legal register.
- Verify the legal record: Retain the identifier and evidence.
- Verify the service: Find current evidence of the required activity.
- Verify the operator: Link each location to its supported company.
- Review group membership: Assess dated parent and transaction evidence.
- Record ambiguity: Preserve plausible alternatives and hold the merge.
Companies House makes registered-address and incorporation information available, while its register guidance includes company status and filing histories. Those are useful identity inputs in that jurisdiction, not a universal substitute for operating evidence. [16] [19]
Keep identifier semantics in the graph
Use separate nodes for company, person, facility, study and submission observation. Add typed, dated edges such as “listed as sponsor,” “investigator at,” “operated by” and “accounting parent.” Do not manufacture an IND identifier or a submission identifier if the available public reference does not supply one.
GLEIF's parent relationships concern direct and ultimate accounting consolidating parents. Reporting exceptions mean an absent parent record cannot by itself establish independence. [9] [10] ROR offers aliases and organizational relationships, but those relationships must retain their stated organizational semantics. [8]
Similarly, one investigator's overlapping affiliations should not merge the affiliated organizations. [17] A CIK also needs an object-type check because SEC uses it for both corporations and individuals. [18] The graph's purpose is to preserve these distinctions, not to make uncertain connections appear authoritative.
- 01Define the mandate
Specify qualifying activities for the CRO, site-management, independent-site or laboratory search.
- 02Preserve the extraction
Retain raw source values and dates before normalization or reviewed identity aggregation.
- 03Review legal identity
Establish which company, if any, the observed name refers to, retaining ambiguity when evidence is insufficient.
- 04Verify current services
Establish what the company currently does and link each location to its supported operator.
- 05Review ownership
Assess whether independence is supported; retain dated parent and transaction evidence.
- 06Record the disposition
Pass supported candidates, exclude those outside the mandate, hold material gaps and escalate conflicting records.
Data Analysis and Evidence
The useful quantitative output is an analysis of a defined extract and its review outcomes. Public-study activity alone supplies no denominator for the population of independently owned service companies.
Use research denominators carefully
A 2022 PLOS ONE study provides a concrete historical illustration of field availability. Its AACT extraction was dated January 25, 2021, for completed industry-sponsored phase III trials during 2008 to 2019. [20] An initial 5,672 trials included 3,652 drug-intervention trials. [20] Of those, 1,672 had valid site-count data, 606 had recruitment-length data and 1,117 had geographic trial-site data. The authors reported no specific funding. These are that paper's denominators, not a present CRO-universe estimate. [20]
The lesson is to publish a separate denominator for each variable. A count of studies with usable location data differs from a count of all selected studies, and neither is a count of company operators. The Census establishment distinction reinforces why location and company totals should be reported separately. [14]
Define extract-level quality metrics
The following formulas are proposed worksheet metrics, not regulator-issued standards:
- Repeated-observation ratio: Repeated source observations divided by all observations in the defined extract.
- Resolved-role rate: Observations with both a supported role and legal-entity mapping divided by relevant-role observations reviewed.
- Company yield: Distinct reviewed qualifying companies divided by candidate names reviewed.
- Open-question count: Unresolved identity, operator, service or ownership questions.
Define a repeated observation as an additional observation beyond the first for the same reviewed source subject and role, within the stated scope. Keep exact technical duplicate rows as a separate ingestion issue. Neither category should be deleted merely to improve a metric.
Table 3 shows a Hypothetical Example, with invented inputs and arithmetic. It is not an extracted BMIS dataset or a measured vendor benchmark.
| Measure | Invented input or calculation | Interpretation |
|---|---|---|
| Source observations | 12 total | Rows retained before reviewed identity aggregation. |
| Reviewed source subjects | 5 subject-and-role groups; 7 additional observations | Repeated-observation ratio = 7 / 12 = 58.3%. |
| Relevant-role review | 6 observations reviewed; 4 have supported role and legal mapping | Resolved-role rate = 4 / 6 = 66.7%. |
| Qualifying companies | 2 distinct reviewed companies | Report separately; do not derive the count from the ratios. |
| Ownership conclusion | 1 supported independent; 1 unresolved | Qualification for service scope does not settle independence. |
The ratios summarize only the invented review batch. In a real extract, repeated submissions can increase row counts while leaving the reviewed-company count unchanged. The unresolved company should remain visible in a separate ownership bucket, rather than being silently counted as independent.
Separate trial linkage from company linkage
WHO's International Clinical Trials Registry Platform (ICTRP) links related records using main and secondary trial identifiers. Multinational trials can appear in more than one registry. (Source: www.who.int) This is a trial-identity problem. Solving it does not identify a site's corporate operator or prove that two named companies share an owner.
Publish source-level totals, reviewed-company totals and unresolved totals together. Do not estimate revenue per trial, contractor accuracy or whole-market coverage from these reconciliation metrics. Those would require additional evidence and a justified population denominator.
Case Studies and Real-World Examples
Sponsor, CRO and three sites (Hypothetical Example)
Consider invented Sponsor Alpha, Service Beta and Sites North, Central and South. A fictional study record names Alpha as sponsor, Beta as collaborator and the locations as facilities. These are invented names and observations, with no relationship to actual acquisition candidates.
The initial record yields five named observations. It does not yield five service companies. The collaborator definition establishes support, leaving the nature of Beta's services open. The facility definition establishes a study location, leaving its operator open. [11]
Suppose subsequent fictional evidence establishes that Beta is a legally identified CRO offering the required services, and that North and Central are operated by one site company. South is a university research location outside the chosen commercial-site mandate. Alpha remains sponsor-only within the evidence assembled for this case.
The reviewed result is two qualifying service companies, subject to separate ownership review: Beta and the operator of North and Central. Alpha is excluded under this fictional services-only mandate; South is excluded from the commercial-site screen. None of these dispositions asserts that an organization could never perform a different role elsewhere.
If the operator of Central remains ambiguous, the proper output changes: one site-company match is held, with its unresolved issue retained. A shared investigator or address would be insufficient to settle that relationship; overlapping person affiliations can legitimately exist. [17]
Repeated BMIS references (Hypothetical Example)
Suppose a fictional extract contains repeated references to the same named CRO across submissions. Retain each observation and its available source reference, then link the supported identities to a reviewed company node. This mirrors BMIS's submission-observation structure without asserting a real row count. [2]
If a historical name is later connected to a current legal entity through dated records, retain the historical alias and document the continuity decision. If the evidence instead suggests a different company, split the candidate and preserve the competing explanation. Historical and current identifier records can assist this review without making continuity automatic. [9]
The result is a count bridge: source observations, reviewed identities, qualifying service companies, and companies whose independence is supported. Each step requires its own evidence.
Implications and Future Directions
A public-record research program should make uncertainty actionable. Each open issue needs a next source or review step rather than a generic “low confidence” label.
- Pass: Role, legal identity and current service evidence meet the defined mandate; state ownership separately.
- Exclude: Supported evidence places the candidate outside the mandate.
- Hold: A material identity, operator or service question lacks evidence.
- Escalate: Conflicting records or complex corporate history require specialist review.
- Refresh: A previously accepted conclusion needs newer supporting evidence.
- Preserve: Historical observations remain accessible after a disposition changes.
IRB observations need particular care. The US Department of Health and Human Services (HHS) says IRB registration is not accreditation or certification. Treat it as evidence of the stated registration, not as proof of operating-company quality or fit for a CRO mandate. [21]
Site-focused teams should likewise distinguish a location from responsibility for the activities conducted there. ICH's investigator-site and delegation concepts support retaining site, investigator and service-provider relationships independently. [13] An acquisition thesis may encompass several of those businesses, but the source model should not make that decision implicitly.
For geographic expansion, add jurisdiction-specific source contracts. EMA's CTIS scope provides a reason to keep EU/EEA regulatory observations distinct from US IND observations, then reconcile trial and company identities deliberately. (Source: www.ema.europa.eu) WHO's cross-registry linking method can assist trial reconciliation without replacing company review. (Source: www.who.int)
The next practical development is a versioned ledger with dated role edges and reversible identity decisions. Improvements should be judged by better supported mappings and fewer unresolved questions in a defined batch. A faster refresh is useful only when the team knows which conclusions it actually refreshes.
Frequently Asked Questions (FAQs)
Can ClinicalTrials.gov identify CROs?
It can supply candidate names and study relationships. A collaborator listing alone identifies support, not a verified CRO contract. Combine the observation with current service and legal-identity evidence before classifying the company. [15]
Is a facility name an acquisition target?
It is a location observation requiring operator research. Keep the facility and operator as separate records, consistent with the distinction between an establishment and a company. [14]
Can a company be both sponsor and contractor?
The ledger should allow it when separate evidence supports each role. Record the study, activity and effective dates. The US IND rules address sponsor responsibilities and transfers to CROs; they do not justify a permanent, mutually exclusive company taxonomy. [3] [4]
Does an identifier prove current independence?
No. Check what the identifier represents and what the relationship data cover. GLEIF allows parent-reporting exceptions, and ROR identifiers need not correspond to legal entities. Missing or differently scoped relationships leave an ownership question open. [10] [8]
Conclusion
A useful ClinicalTrials.gov FDA BMIS crosswalk is an evidence model for candidate research. It connects source observations to people, facilities, studies and reviewed legal companies while preserving the role and time attached to every relationship.
The workflow begins with the mandate. CRO, site-management, independent-site and laboratory searches should have explicit qualifying activities. Extraction then preserves raw source values and dates. Identity review establishes which company, if any, a name refers to; operating review establishes what it currently does; ownership review determines whether independence is supported. A favorable answer at one stage does not supply the missing answers at the others.
The handoff should include the source-selection record, role ledger, identity decisions, current service evidence and open issues. Counts should distinguish observations, subject-and-role groups, facilities, companies and corporate groups. Any ratio should retain its numerator, denominator and extraction boundary.
For origination teams, the benefit is a candidate universe whose construction can be reviewed and corrected. A historical study relationship can justify further research without proving current revenue or willingness to transact. A repeated submission can strengthen the record of an observed role without increasing company count. A source gap can remain visible without being filled by an unsupported assumption.
That discipline makes public clinical-research records usable for services-market mapping: they become documented leads that can be qualified, held or excluded, with a clear account of what remains to be established.
External Sources (21)
Published by
Docket Research
Docket provides deal-origination research software and managed research for private equity teams, screening companies against a mandate and showing the evidence behind each finding.
Editorial note
Researched and written by Docket Research with AI-assisted research tools; sources are cited inline and checked before publication. Report an error: https://docket.capital/evidence-standards#corrections.