The clinical-trial substrate.
Where the protocol meets the world. Sites, EDC, CTMS, eTMF, eConsent, ePRO/eCOA, IxRS/RTSM, the data warehouse, and the CRO landscape that pulls them together. The operational layer beneath every regulator's GCP demand.
Trial sites: academic, private, decentralised.
Where the protocol meets the participantA clinical trial is a contract between a sponsor and a site. The site is where the participant turns up, where source data is generated, and where the inspector eventually opens the binder. Three site archetypes coexist in 2026 — academic medical centres, dedicated private investigative sites, and decentralised constellations — with varying mixes of investigator continuity, patient access, and operational throughput.
Academic medical centres.
University hospital networks. Deep specialty depth, faculty PIs, slower start-up, complex contracting, strong publication culture. Where rare-disease, oncology, and translational-science trials disproportionately run.
Dedicated investigative sites.
For-profit Phase I-IV networks. Faster start-up, professional CRC staff, dedicated infrastructure, throughput-optimised. The bulk of US Phase II/III enrolment volume runs through these.
Community hospital + private practice.
Site-level investigator with local patient base, often single-PI. Important for diversity-of-population goals and for cardiovascular, primary-care, dermatology indications. Variable infrastructure; depends on SMO support.
Decentralised · hybrid.
Direct-to-participant via home-health visiting nursing, telemedicine, eConsent, mailed IMP, wearables. Compressed in 2020 by COVID-19, formalised in FDA Decentralised Clinical Trials Guidance Sep 2024 and EMA Recommendation Paper Dec 2022.
FDA DCT Guidance Sep 2024 + EMA Recommendation Paper Dec 2022.
A pure-DCT trial is rare in 2026. Most decentralised studies are hybrid: site-anchored for screening, decentralised for follow-up visits. The investigator of record remains the same regardless of where activities happen. Direct-to-participant IMP shipping is permitted in the US, EU (varies by Member State), UK, and Australia; restricted or prohibited in many other jurisdictions.
Site selection models.
Feasibility · landscape · site qualificationSite selection is where most trial timelines win or lose. A protocol's enrolment assumption is only as good as the feasibility evidence behind it — and feasibility is increasingly evidence-driven, not relationship-driven, with the rise of investigator databases (CitelineCT, IQVIA Site Atlas, TriNetX), claims-data feasibility (HealthVerity, Komodo), and EHR-network proxies for protocol fit.
Landscape assessment.
Indication, geography, competitive trial density. Tools: ClinicalTrials.gov, EU CTIS, jRCT, CTRI, ANZCTR. Citeline, GlobalData, Evaluate. The denominator question.
Country shortlist.
Regulatory clock + ethics clock + medical practice fit + standard-of-care alignment. EU CTR clock (60+45-day Part I, parallel Part II) very different from CDSCO 30/90 clock in India.
Investigator database mining.
Past performance: enrolment rate vs forecast, screen-fail rate, query rate, deviation rate. Citeline Sitetrove, IQVIA Site Atlas, TriNetX, Veeva Site Connect.
Feasibility questionnaire.
Online questionnaire + phone follow-up. Patient population, equipment, staff bandwidth, competing trials. eFeasibility tools shorten cycle from 6-8 weeks to 2-3 weeks.
Pre-Study Visit (PSV) / virtual qualification.
Final qualification visit before site activation. Increasingly virtual since 2020. Confirms staff, equipment, source-document setup, training plan.
Activation & SIV.
Contract execution, IRB/IEC approval, regulatory CTA approval, IMP shipment, Site Initiation Visit (SIV), first-patient-in (FPI) authorisation. The site is now live.
EDC ecosystems.
Electronic data capture · the workhorse layerEDC is the system of record for case-report-form data. The market consolidated through the 2010s and bifurcated by 2020 into three blocks: integrated cloud platforms (Veeva, Medidata), open-source (OpenClinica, REDCap), and specialist mid-market (Castor, Clinical Ink, Crio). Choice of EDC drives downstream choice of data warehouse, statistical reporting environment, and submission package format.
Veeva Vault Clinical.
Clinical Database (CDB) + EDC + CTMS + eTMF + Site Connect on a single Vault platform. Large-pharma adoption since ~2018; J&J, Bayer, Merck KGaA on Vault for major programmes.
Medidata Rave + Clinical Cloud.
Rave EDC the EDC market leader by trial count for ~15 years. Acquired by Dassault Systèmes 2019. AI Clinical Co-Pilot launched 2024. Pharma + biotech adoption broad.
OpenClinica.
Open-source EDC widely used in academic, investigator-initiated, and global-health trials. Hosted commercial tier OpenClinica Enterprise. WHO PQT clinical data submissions often via OpenClinica.
Castor.
Specialist EDC + ePRO + eConsent. Strong in academic and biotech mid-market. Self-service trial build; common in investigator-initiated and academic-IRR studies.
Clinical Ink.
Direct Data Capture (DDC) at the site, tablet-based source. eSource at point of care eliminating SDV in pivotal-design trials. Strong in CNS, oncology, rare disease.
Crio (Clinical Research IO).
Site-first EDC + eSource + CTMS. Adoption strong in dedicated investigative-site networks. Velocity Clinical, Headlands, Centricity Research run on Crio for source-of-truth at the site.
21 CFR Part 11 + EU Annex 11 · computerised system validation.
Every EDC platform deployment requires CSV: User Requirement Specification (URS) → Functional Specification → IQ/OQ/PQ. Cloud platforms ship with vendor-side validation evidence; sponsor still responsible for use-specific validation. Validated configuration is a deliverable to the regulator on request.
The clinical operations stack.
CTMS · eTMF · eConsent · ePRO/eCOA · IxRS/RTSMBeneath the EDC layer sits a constellation of operational systems. Each was a separate vendor ten years ago; each is becoming a module within a Veeva or Medidata cloud platform today. The mid-market specialist tier remains lively because protocol-specific needs do not conform to one-size platform configurations.
Data warehousing · SDTM · ADaM · Define-XML.
From CRF to submissionThe path from EDC to submission is paved with CDISC standards. SDTM (Study Data Tabulation Model) standardises raw, observation-level data; ADaM (Analysis Data Model) feeds the analyses in the Clinical Study Report; Define-XML describes the metadata. FDA has required SDTM for new submissions since Dec 2016; PMDA since Apr 2020; EMA accepts SDTM and is mandating it through 2026 for centralised submissions.
The data warehouse is where SDTM and ADaM live, where SAS/R programs run, and where the eCTD Module 5 datasets get assembled. Tableau, Spotfire, and Power BI sit on top of it for ongoing study oversight; the same warehouse later feeds Real-World Evidence post-marketing analyses.
EDC capture.
Veeva Rave / Medidata / Castor / Clinical Ink — CRF entry by site CRC, eCOA via patient device. Data flows nightly to the warehouse via validated ETL.
SDTM mapping.
Raw CRF data → SDTM domains (DM, AE, CM, EX, LB, VS, etc.). CDISC IG version locked at study start. Define-XML 2.1 metadata.
ADaM derivation.
Analysis datasets per the Statistical Analysis Plan: ADSL (subject-level), ADAE (adverse events), ADTTE (time-to-event), etc. Pinnacle 21/Pharma-XML Validation.
TLF production.
Tables, Listings, Figures for the CSR. SAS, R, or hybrid. Output validated by independent QC programmer.
eCTD Module 5 packaging.
Datasets + Define + reviewer's guide + study report. eCTD v4.0 supported by FDA since 2018, EMA since 2024 phase-in.
The CRO landscape.
Large vs niche · sponsor-of-record vs functional-service-providerMost pharma trials are outsourced. The Contract Research Organisation acts either as full-service (sponsor delegates whole trial conduct) or as Functional Service Provider (sponsor retains oversight, vendor delivers a function: monitoring, biostatistics, data management). The market consolidated through 2018-2024 around five global full-service incumbents and a long tail of mid-market and niche specialists.
Top-five global incumbents.
Full-service Phase I-IV across all therapeutic areas, with global site networks, integrated biostatistics, data management, pharmacovigilance, regulatory, medical writing. Increasingly differentiated by AI, RWE, and data-asset offerings.
Large mid-market full-service.
Full-service or near-full-service with specialty depth. Often preferred by small-to-mid pharma wanting senior-team attention. Several from acquisitions of regional CROs.
Specialty · therapeutic-area niche.
Deep niche: oncology, rare disease, ophthalmology, dermatology, CNS, gene therapy. Investigator networks often inherited from a founder PI, hard for tier-1s to replicate.
Functional service providers.
FSP delivers a single function across many sponsor protocols: monitoring (CRA bench), biostatistics, data management, pharmacovigilance, medical writing, regulatory affairs. Sponsor remains regulatory sponsor; FSP is staff-augmentation.
Regional specialists.
Domestic-led CROs deep in their geography. Often preferred for that region's regulatory pathway (CDSCO, NMPA, ANVISA, COFEPRIS) and patient population.
Decentralised operators.
Site-less or site-light CRO/SMO hybrids that deliver decentralised study delivery directly to participants. Most have pivoted toward hybrid (some site, some decentralised) since 2023.
The non-delegable sponsor obligation.
ICH E6(R3) Annex 1 §3.18 (Sponsor Quality) is explicit: outsourcing does not transfer regulatory accountability. The sponsor remains responsible for quality, data integrity, safety, and TMF completeness regardless of who runs which task. Vendor oversight plans, governance committees, and quality agreements are not optional.
Site networks & SMOs.
Aggregating sites for sponsor + CRO efficiencySite Management Organisations (SMOs) and dedicated investigative-site networks aggregate dozens to hundreds of sites under one operational and contracting umbrella. They reduce sponsor/CRO overhead per site, accelerate start-up, professionalise CRC staffing, and provide cross-site visibility. The trade-off: less heterogeneity than a country-spread of independent sites; concentration risk if a network has an inspectorate finding.
Interoperability & standards.
CDISC · HL7 FHIR · ICH M11The substrate is held together by data standards. CDISC covers the trial data lifecycle (CDASH for case report forms, SDTM for tabulation, ADaM for analysis, SEND for non-clinical, Define-XML for metadata). HL7 FHIR covers EHR-to-EDC and clinical data exchange. ICH M11 Clinical Electronic Structured Harmonised Protocol (CeSHarP) was finalised at Step 4 in 2024 and rolling into FDA/EMA/PMDA implementation in 2026 — the first machine-readable harmonised protocol template.