Authoring the Site Playbook: Roles, Workflows, Checklists, and Tools That Make the Right Action the Easy Action

Who this document is for. The Playbook is written for investigators, sub-investigators, coordinators, pharmacists, research nurses, and site administrators. Write in imperative voice with numbered steps, role callouts, and screen/label snippets where misclicks have consequences. Use a consistent iconography for critical-to-quality (CtQ) steps that protect safety, rights, and endpoint integrity.

Structure that sites can actually run. Organize the Playbook into seven short, searchable clusters: (1) Study Overview & Contacts—24/7 safety line, IWRS/IRT hotline, courier account, imaging/lab help desks; (2) Pre-Screening & Eligibility—quick-check grid (criteria, source location, threshold, who confirms), with scripts for sensitive topics; (3) Informed Consent—version control, interpreter/witness rules, remote consenting identity verification, copy-to-participant steps; (4) Randomization & First Dose—eligibility confirmation, randomization steps, kit pick-up, IP checks; (5) Visits & Assessments—required/optional procedures, windows, “if missed, then” logic, tele-visit instructions; (6) Safety & Escalation—what to report, clocks, first-hour checklist, who approves emergency unblinding; and (7) Close-Out—IP reconciliation, returns/destructions, records finalization, archiving.

Delegation, training, and attestations. Include a delegation matrix template and role descriptions. Define who may randomize, who may dispense, who may trigger a code break, and who signs off forms. Provide a training log and brief scenario-based checklists (e.g., “tele-visit fails,” “device firmware update,” “participant arrives outside window”). Signatures throughout the Playbook should capture the meaning of approval (e.g., “PI discussed consent and answered questions,” “Pharmacist verified kit and label,” “Coordinator completed identity check”).

Source documentation and ALCOA++ discipline. Spell out minimum source expectations by procedure (vitals, ECG, imaging order, dosing) and how contemporaneity is preserved for remote data (device time sync, server stamps). Teach how to correct errors without overwriting: who can change, reason for change, and how audit trails are reviewed. Provide examples of acceptable abbreviations and prohibit PHI/PII in free text that migrates to vendor portals.

Decentralized and hybrid operations. For at-home elements, include identity verification steps, packaging instructions for mail-back kits, what to do if a courier window is missed, and how to document failed attempts without shaming participants. Provide instructions for mobile nursing, local labs/imaging, and remote ePRO/eCOA troubleshooting. Clarify when a home visit can “satisfy” a clinic visit and how to record substitutions without creating silent deviations.

Pharmacy and temperature control touchpoints. Summarize receipt, storage, probe placement, alarm thresholds, and excursion decision trees (assess → document → disposition). Provide photos or diagrams of acceptable pack-outs and a short guide to reading logger outputs. Map returns/destroys to accountability and IWRS/IRT events so physical stock and system stock never diverge.

Courier exceptions and supply resilience. Provide a simple algorithm for late pickup, dry-ice shortfall, or “address closed” errors. Teach the difference between site-initiated and sponsor-initiated reshipments, and which messages to send to participants to avoid biasing adherence behavior. Include a micro-budget policy for parking or ride vouchers to avoid no-shows in high-burden visits.

Five-minute retrieval drills. Once per month, pick a scenario—“eligibility lab out of window,” “kit destroyed after excursion,” “tele-visit missed,” “emergency unblinding”—and practice retrieving the entire evidence chain: Playbook step → user action → system audit trail/screenshot → eTMF filing. Record pass/fail and corrective actions; recurring defects should drive design changes (template wording, screen prompts), not just retraining.

Writing the IWRS/IRT Guides: Randomization, Supply, Unblinding, and System Integrity

Purpose and scope. The IWRS/IRT Guides turn supply strategy and blinding design into step-by-step, inspection-ready instructions. They cover: user roles and authentication; site activation and user provisioning; randomization schemes (ratio, blocking, stratification, minimization if applicable); kit types and label logic; initial and ongoing resupply; returns, quarantines, and destructions; emergency unblinding; and data interfaces to EDC, safety, and eTMF.

Role-based access and provisioning. Define roles (PI, unblinded pharmacist, coordinator, sponsor viewer, depot) and their permissions. Enforce least-privilege: coordinators can schedule randomization but cannot view treatment codes; pharmacists can dispense and reconcile but cannot alter allocation rules; sponsors have blinded oversight unless firewall status grants unblinded access for a designated statistician. Require multi-factor authentication and password rotation consistent with institutional norms.

Randomization that resists error. Document the exact steps for allocation: eligibility confirmation, pre-randomization checks, handling screen failure after randomization, and how mis-randomizations are detected and resolved. Include screenshots of the allocation page, error states, and recovery prompts. For stratified designs, provide a one-page table of strata and examples so sites pick the same category for the same scenario.

Supply logic that keeps clinics running. Explain initial provisioning and resupply triggers (predictive, min/max, just-in-time). Provide formulas and examples for fast vs. slow enrollers and for devices with consumables. Require depot/site reconciliation frequencies and define automatic holds when discrepancies exceed thresholds. Teach how to quarantine kits (temperature excursion, damaged shipper, label error) and how to release or destroy after stability review—with all steps reflected in audit trails.

Direct-to-patient (DTP) and hybrid supply. Specify identity checks on delivery (two identifiers), reship criteria, and what happens if the window is missed. Provide tamper-evident seal rules, pictorial instructions for lay users, and return logistics. Distinguish operational messages to participants from any treatment-revealing content; teach staff scripts that do not leak allocation.

Emergency unblinding—minimal information, maximal safety. Give the authorization matrix (who requests, who approves), the steps in the system code-break feature, and a sealed-envelope backup if the system is down. Require two-person verification, reason capture, and automatic alerts to the Safety Physician and the sponsor’s liaison. If only the treating clinician needs the code, keep the rest of the site and sponsor blinded; if the clinical decision can be made without knowing the other arm’s identity, reveal only the participant’s assignment.

Interfaces and data hygiene. Map field-level exchanges to EDC (subject IDs, visits, dosing dates), safety (unblinding flags), and eTMF (exportable audit logs, code-break reports). Require synchronized clocks across systems to preserve contemporaneity. Provide a naming convention for exports and mandate checksum verification on files transferred to the eTMF.

Validation and UAT. Include positive/negative test cases: randomization with missing stratum, resupply when stock is quarantined, code-break when authorization fails, and device-kit pairing errors. Document expected messages and blocker states so sites can recognize a configuration issue in minutes. Store UAT evidence with signatures that state the meaning of approval (“Configuration verified,” “Clinical accuracy approved”).

Metrics embedded in the system. Track and display: median time from “randomize” click to kit issue, rate of mis-randomizations, stockout near-misses, quarantine frequency by reason, DTP reship rate, excursion-related waste, and age of unresolved supply queries. Use red/amber/green thresholds and auto-escalation rules that open a ticket and add the metric to weekly risk huddles.

Governance, Vendor Oversight, Metrics, and a Ready-to-Use Checklist

Small-team governance with meaning of signature. Name a Site Operations Lead (accountable for the Playbook), an Unblinded Pharmacy Lead (accountable for kit flow), a Supply Chain Lead (depot → site), an IWRS/IRT Owner (configuration and interfaces), a Safety Physician (unblinding oversight), and Quality (ALCOA++ verification). Approvals should record the meaning of signature (“Operational feasibility verified,” “Firewalls confirmed,” “Audit trail sufficiency approved”).

Vendor and CRO oversight. Build obligations into quality agreements and statements of work: immutable logs; role-based access; synchronized clocks; audit-export formats; DTP identity verification rules; temperature-logger specifications; query turnaround SLAs; and participation in monthly five-minute retrieval drills. Require shared dashboards for red/amber/green metrics and credits or at-risk fees for persistent red items.

KPIs that predict control.

• Timeliness: days from site activation to first randomization; median time from kit arrival to site release; response time to courier exceptions.
• Quality: first-pass acceptance of randomization and dispensing records; percent of visits on window; logger download completeness; proportion of code-breaks with full documentation on first pass.
• Consistency: divergence between system stock and physical stock; rate of eligibility misclassifications detected post-randomization; repeat causes of quarantine.
• Traceability: five-minute retrieval pass rate for Playbook step → IWRS/IRT event → eTMF artifact; alignment of timestamps across platforms.
• Effectiveness: reduction in deviation categories tied to supply or randomization; time-to-green after CAPA; inspection observations related to IWRS/IRT or site execution.

Common pitfalls—and durable fixes.

• Playbooks that read like policy, not instructions. Fix by using imperative steps, screenshots, and decision trees; move rationale to callouts.
• Quiet edits to screens or labels. Fix by enforcing change control with redlines and by filing “what changed and why” memos; re-train only after the eTMF shows effective dates.
• Stockouts and late resupply. Fix with predictive triggers, min/max rules tuned to enrollment velocity, and quarantine-aware resupply logic.
• Over-use of emergency unblinding. Fix with explicit examples/non-examples, two-person authorization, and least-information rules.
• Courier and DTP blind spots. Fix with photo guides to pack-outs, identity-check scripts, and a reship algorithm that preserves windows without biasing behavior.
• System–paper divergence. Fix with weekly reconciliation of physical counts to IWRS/IRT and eTMF artifacts; escalate mismatches over a defined threshold.

30–60–90-day rollout. Days 1–30: publish Playbook and IWRS/IRT templates; finalize role matrices and meaning-of-signature blocks; complete UAT with positive/negative cases; configure dashboards and thresholds. Days 31–60: activate first sites; run tabletop drills for emergency unblinding, courier exceptions, and temperature excursions; rehearse five-minute retrieval for two evidence chains per site; tune resupply triggers. Days 61–90: scale to all regions; implement weekly risk huddles; add vendor scorecards; close CAPA items with design changes (screen prompts, label text, kit assortments), not only retraining.

Ready-to-use checklist (paste into your SOP or site initiation pack).

• Study Overview & Contacts complete; 24/7 safety/IWRS/IRT/courier lines verified; escalation tree posted.
• Pre-Screening grid approved with thresholds, sources, and who confirms; scripts localized and trained.
• Consent steps, interpreter/witness rules, remote identity checks, and copy-to-participant procedure documented.
• Randomization steps with screenshots; strata table and examples; mis-randomization recovery documented.
• Pharmacy receipt/storage, probe placement, alarm thresholds, and excursion decision tree posted; label verification step defined.
• Resupply logic (min/max or predictive) configured; quarantine/release/destroy pathways documented; weekly reconciliation scheduled.
• Emergency unblinding authorization matrix live; two-person verification and least-information rules trained; sealed-envelope backup secured.
• DTP workflow defined (identity, tamper seals, reship rules, return logistics) with participant-safe scripts.
• Interfaces mapped to EDC/safety/eTMF; export naming and checksum rules enforced; clocks synchronized.
• Dashboards active; KPIs/KRIs reviewed weekly; five-minute retrieval drill passed for two random scenarios; CAPA closes with design fixes.

Bottom line. When a Site Playbook and IWRS/IRT Guides are authored like controlled code—imperative steps, clear roles, CtQ emphasis, firewalled unblinding, solid supply logic, synchronized systems, and evidence you can retrieve in minutes—sites execute confidently, the blind stays intact, and your study remains inspection-ready from first patient screened to last kit destroyed.