📚 Part of the OBX-319 Regulatory Dossier — Reader's Guide. This article shows the live document; edits to the source appear here automatically.

---

Data Management Plan (OBX319-301)

Document ID: TMF-002
Version: 1.0
Change History: 1.0 — Initial issue.
Standard(s): ICH E6(R3)

Data Management Plan — OBX319-301

Defines the data-handling lifecycle for OBX319-301: EDC build and validation, CRF/eCRF design, edit checks, data review and query management, medical coding (MedDRA/WHODrug), SAE reconciliation, external-data handling, and database lock, delivering CDISC SDTM/ADaM datasets. ICH E6(R3); CDISC.

Scope and study context

OBX319-301 is a Phase 3, randomized, double-blind, placebo-controlled study of OBX-319 — an anti-CD19 × anti-CD20 humanized IgG1 B-cell-depleting bispecific monoclonal antibody, produced in Chinese hamster ovary (CHO) cell culture and administered subcutaneously — in adults with moderate-to-severe active Systemic Lupus Erythematosus (SLE) on background standard of care. A total of 480 subjects are randomized 1:1:1 to High-dose OBX-319 (162), Low-dose OBX-319 (158), or placebo (160) and are treated and assessed over 52 weeks. This plan governs every category of clinical data supporting the Week 52 primary endpoint (SRI-4 response with low disease activity, defined as SLEDAI-2K ≤ 4) and the SLEDAI-2K disease-activity, serologic (anti-dsDNA, complement C3/C4), pharmacodynamic (CD19+ B-cell count), pharmacokinetic, and immunogenicity data, and defines the standards, systems, ownership, and procedures required to deliver an inspection-ready, locked database and CDISC-conformant datasets for the Biologics License Application under 21 CFR 601. Baseline disease activity is moderate-to-severe (SLEDAI-2K approximately 11), and the plan is written to preserve the double blind across all data streams through database lock.

Roles, responsibilities, and data-management systems

Data management is executed by the contract data-management provider under the oversight of the sponsor, Virtual Biopharma Inc. Named roles include the Lead Data Manager, Clinical Data Coordinator, Database Developer/Programmer, Medical Coder, and the Data Management Standards/CDISC lead, with defined interfaces to Clinical Operations, Biostatistics, Statistical Programming, Pharmacovigilance, Clinical Pharmacology, and the central and specialty vendors. Data are captured in a validated, 21 CFR Part 11-compliant Electronic Data Capture (EDC) system with a secure, time-stamped audit trail, role-based access control, electronic signatures, and system-generated version control. The Interactive Response Technology (IRT/IWRS) governs randomization and drug supply; safety events are managed in the sponsor's pharmacovigilance safety database. A Data Management File retains all specifications, validation records, transfer agreements, coding dictionaries, review documentation, and lock approvals. Study-specific data-handling conventions, self-evident corrections, and the discrepancy-management workflow are defined and version-controlled before first-subject-in.

EDC build and validation

The EDC study database is built from approved specifications (CRF specification, edit-check specification, and data-transfer specifications) and validated under a documented computer-system-validation lifecycle: requirements, design, build, and formal User Acceptance Testing (UAT) with traceable test scripts covering data entry, derivations, edit checks, dynamic/branching forms, visit structure, and role-based access. Test evidence, defect logs, and a validation summary are approved before the database is released to production. Any post-production change is controlled through change control, re-tested, and re-approved, with migration of affected data verified. The visit schedule in the build reflects the 52-week double-blind treatment period and the follow-up assessments, including the B-cell, pharmacokinetic, and immunogenicity sampling windows.

CRF/eCRF design and data capture

The eCRF is designed to capture the full disease-activity, safety, pharmacodynamic, and biologic-specific dataset for a B-cell-depleting bispecific antibody, and includes:

• Disease activity and response: SLEDAI-2K components, BILAG-2004 organ-domain scoring, Physician's Global Assessment (PGA), and the elements required to derive the SRI-4 responder endpoint and the SLEDAI-2K ≤ 4 low-disease-activity target at Week 52.
• Background standard-of-care therapy: corticosteroids (with dose and taper capture), antimalarials, and immunosuppressants, recorded to support protocol-defined concomitant-medication and steroid-taper rules.
• Study drug administration and accountability: subcutaneous dosing records, injection-site observations, and drug-accountability reconciliation against the IRT.
• Safety: adverse events, serious adverse events, and adverse events of special interest aligned to the identified class risks of B-cell depletion — serious and opportunistic infections, hypogammaglobulinaemia (with immunoglobulin, notably IgG, monitoring), and injection/infusion-related reactions.
• Laboratory and serology: central-laboratory hematology/chemistry, anti-dsDNA, and complement C3/C4.
• Pharmacodynamics: CD19+ B-cell counts by flow cytometry, the primary pharmacodynamic marker of depletion.
• Pharmacokinetics and immunogenicity: serum OBX-319 concentration sampling and anti-drug-antibody (ADA) sampling, with collection date/time and relationship to dosing recorded to support the target-mediated (TMDD) pharmacokinetic and immunogenicity analyses.

Data are entered at site with expectation of timely entry after each visit; required-field, format, and range controls operate at point of entry.

Edit checks and data validation

Programmed (automatic) and manual edit checks enforce completeness, range, logical consistency, and cross-form/cross-visit coherence. Study-specific checks include range and trend checks on CD19+ B-cell counts, anti-dsDNA and C3/C4, and immunoglobulin values; consistency between dosing records, visit dates, and pharmacokinetic/ADA sampling times; concomitant-medication and corticosteroid rules; adverse-event date/severity/seriousness logic; and derivation-supporting checks for SLEDAI-2K and the SRI-4 components. The edit-check specification is validated during UAT, and check performance is monitored throughout conduct. Centralized data-review listings and, where applicable, statistical monitoring supplement the programmed checks.

Data review and query management

Ongoing clinical, safety, and data-management review is performed against a data-review plan using system listings and status metrics. Discrepancies raise queries in the EDC; queries are routed to sites, resolved, and closed with a full audit trail, and query-aging and outstanding-data metrics are tracked to lock-readiness targets. Blinded medical/data review is conducted so that emerging data are cleaned without compromising the treatment blind (see unblinding-sensitive data handling below). Protocol deviations are identified, categorized, and reconciled with Clinical Operations for consistent capture in the analysis datasets.

Medical coding (MedDRA / WHODrug)

Adverse events and medical history are coded to MedDRA, and concomitant and prior medications to WHODrug, using version-controlled dictionaries with a documented auto-encoding and manual-review workflow and medical review of coded terms. Coding conventions are defined for the modality's key safety terms — serious and opportunistic infections, hypogammaglobulinaemia, and injection/infusion-related reactions — and for the background SLE therapies (corticosteroids, antimalarials, immunosuppressants). Dictionary versions are fixed at defined points, with re-coding and impact assessment controlled if an up-versioning occurs before lock.

External and vendor data handling

External data are received under approved data-transfer specifications and reconciled against the EDC and IRT before lock. External sources include the central laboratory (hematology, chemistry, serology), the specialty flow-cytometry laboratory (CD19+ B-cell enumeration), the bioanalytical laboratory (serum OBX-319 concentrations supporting the TMDD pharmacokinetic analysis), the immunogenicity laboratory (tiered ADA testing — screen, confirm, titer, and neutralizing-antibody assessment), and the IRT. Each transfer is format-validated, reconciled on key identifiers, sample dates/times, and expected record counts, and discrepancies are queried and resolved. Transfer frequency, file structure, and reconciliation criteria are documented per vendor.

Unblinding-sensitive data handling

Because OBX-319 produces near-complete depletion of CD19+ B cells on the active arms — declining from approximately 210 to approximately 7 cells/µL — while placebo values remain essentially unchanged, the flow-cytometry B-cell data, serum drug concentrations, and ADA results are potentially unblinding. These data streams are therefore designated restricted access and are firewalled from the blinded study team: they are received, reconciled, and reviewed by a segregated group and are not returned to sites or exposed to blinded reviewers until the blind is broken at database lock. Access controls, restricted-data transfer routes, and unblinding-protection procedures are documented and audited, consistent with the double-blind design.

SAE reconciliation

Serious adverse events recorded in the clinical (EDC) database are reconciled against the pharmacovigilance safety database on a defined periodic basis and prior to lock. Reconciliation compares event term, onset/resolution dates, seriousness criteria, causality, and outcome; discrepancies are documented, queried, and resolved so that the two systems are consistent. Given the identified risks of B-cell depletion, particular attention is paid to serious and opportunistic infections and to events associated with hypogammaglobulinaemia.

Database lock

Lock readiness is confirmed against a pre-lock checklist: all data entered and cleaned, queries closed, external data reconciled, coding complete and approved, SAE reconciliation complete, protocol deviations finalized, and outstanding-item logs cleared. A soft (interim) lock precedes final review; upon sign-off by Data Management, Biostatistics, Clinical, and Pharmacovigilance, the database is hard-locked, treatment codes are released, and the locked datasets are transferred to Statistical Programming. Any post-lock change is governed by a controlled database-unlock/re-lock procedure with documented justification and re-approval.

CDISC deliverables (SDTM/ADaM)

Locked data are transformed to CDISC SDTM and analysis-ready ADaM datasets using controlled terminology and version-controlled specifications, with Define-XML v2.0 metadata and the accompanying Study Data Reviewer's Guides (cSDRG for SDTM, ADRG for ADaM). SDTM coverage includes the disease-activity and questionnaire data (SLEDAI-2K, BILAG-2004, PGA), response derivation supporting SRI-4 and the SLEDAI-2K ≤ 4 low-disease-activity endpoint, laboratory and serology (including anti-dsDNA and C3/C4), the CD19+ B-cell pharmacodynamic findings, pharmacokinetic concentrations, and immunogenicity results, with corresponding ADaM datasets (subject-level, response, laboratory/biomarker, pharmacokinetic, and immunogenicity) supporting the Week 52 analyses. Deliverables are validated against CDISC conformance rules and prepared to the applicable FDA study-data submission standards for the BLA.

Quality control, audit trail, and records retention

Quality control is applied throughout — validated build and edit checks, independent review of coding and derivations, reconciliation of external and safety data, and documented lock approvals — and all changes are captured in the system audit trail. Data-management documentation, specifications, validation evidence, transfer agreements, dictionary versions, and lock records are retained in the Data Management File for the regulatory retention period and are available for inspection, consistent with ICH E6(R3) and applicable regulatory requirements.