Published on 16/11/2025
Pharmacy, Lab, and Imaging Manuals That Sites Can Run—and Inspectors Can Trust
Purpose, Scope, and the Global Regulatory Frame
Pharmacy, lab, and imaging manuals are the operational heart of a clinical trial. They translate protocol intent into step-by-step actions for pharmacists, coordinators, technicians, and imagers—so safety is protected, blinding is maintained, and endpoints are measured consistently across countries and vendors. Well-written manuals reduce deviations, temperature excursions, sample losses, and unusable images; poor manuals do the opposite. Your goal is a set of documents that sites can run without guesswork and
Principled anchors. A quality-by-design posture—concentrating controls on what is critical to participant safety, rights, and endpoint integrity—is consistent with the ICH Good Clinical Practice principles. In the U.S., expectations around investigator responsibilities, informed consent, safety reporting, and trustworthy electronic records form the background against which operational manuals are judged; program teams often orient to the public materials within FDA clinical trial oversight resources. For EU/UK studies, operational practice and transparency duties are informed by high-level notes accessible through the European Medicines Agency’s clinical trial guidance. Ethics and fairness in execution are underscored by the World Health Organization’s research ethics materials. For Japan and Australia, align style and expectations with PMDA clinical guidance and the TGA clinical trial guidance so multinational instructions remain coherent.
Where manuals begin and end. Manuals operationalize the protocol; they do not silently change it. If a manual must introduce or revise a safety-critical rule (dose preparation, temperature limits, timing windows, acquisition parameters), the protocol or a controlled addendum should say so. Cross-references must be explicit: the protocol tells the reader where authoritative details live; the manual repeats only what staff need at the bench, scanner, or specimen station.
Audience and readability. Write for the person performing the task. Use imperative voice, numbered steps, iconography where appropriate, and margin cues for critical-to-quality items (CtQ). Provide a one-page “quick start” per procedure (e.g., thaw-prep-dispense, spin-aliquot-ship, scan-reconstruct-upload) and place rationale in shaded callouts so users understand why a step matters.
Inspection posture. Auditors and inspectors typically ask: Are the instructions feasible at real sites? Do they preserve blinding? Are temperature and timing controls meaningful and monitored? Can any number in the CSR be traced to a step here with ALCOA++ evidence (attributable, legible, contemporaneous, original, accurate—plus complete, consistent, enduring, available)? Are signatures captured with the meaning of each approval (“Pharmacy content accuracy,” “Imaging physics approval,” “Biorepository acceptance”)? Manuals should make these answers immediate.
Design for decentralized and hybrid trials. Home health visits, direct-to-patient (DTP) shipments, and remote imaging introduce extra failure modes. Manuals must specify identity checks, cold-chain controls, kit swaps, courier pick-up contingencies, scanner acceptance tests at affiliated community centers, and how to document events when the sponsor cannot be on site. Treat remote steps as first-class procedures, not footnotes.
What “good” looks like. A good manual is version-controlled, searchable, accessible (semantics for screen readers), translation-ready, and illustrated with photos or schematics where words alone invite error. It bundles a checklist, labels, and specimen/shipping worksheets; it names owners; it embeds escalation paths; and it rehearses the five-minute retrieval drill from manual step → source record → dataset or image → analysis table.
Pharmacy Manual: IMP Accountability, Compounding, Blinding, and DTP Logistics
Define scope and roles. State which products are covered (investigational product, comparators, rescue, placebo), who prepares and dispenses (unblinded pharmacist vs. blinded nurse), and which steps are CtQ (dose calculation, dilution, infusion rate, chain of identity). Provide a single escalation tree (24/7 contacts) for clinical queries, temperature excursions, labeling errors, or suspected unblinding.
Receipt and storage. Include delivery acceptance checks (integrity, tamper seals, calibration of receiving probes), quarantine rules pending release, and mapping of storage locations to temperature ranges (ambient/refrigerated/frozen/ultra-low). Specify continuous monitoring, probe placement, calibration/offset records, and alarm thresholds. Provide an excursion decision tree (time-out-of-range × temperature × product stability), required documentation, and disposition (use, return, quarantine, destroy) with lot-level traceability.
Accountability and reconciliation. Provide log templates that capture lot/batch, unique kit IDs, randomization linkage, date/time/user, dispensed quantity, returns, and destructions. Require dual verification for CtQ steps and reconciliation at participant, visit, and site close-out levels. For controlled substances, add vault rules, witness signatures, perpetual inventory calculations, and jurisdictional nuances. Link records to IWRS/IRT to avoid divergence between physical and system inventory.
Compounding and preparation. Present one-page recipes with prerequisites (hand hygiene, PPE, laminar flow class), diluent options, final concentration, container/infusion set specifications, and compatibility notes. State hold times by condition (room temperature/ refrigerated) and light protection needs. For weight-based doses, embed calculators or tables, rounding rules, and max volumes. Include inline photos of acceptable final appearance and a troubleshooting box (e.g., precipitation).
Blinding safeguards. Explain how look-alike/feel-alike risks are mitigated: identical packaging, separate unblinded prep areas, opaque over-bags, barcoded labels without treatment codes, and scripted interactions between unblinded and blinded staff. Provide an emergency unblinding pathway, roles, documentation, and post-event impact assessment. Define how returned product is handled without revealing allocation.
Labeling and language. Supply label text elements (protocol code, kit ID, storage, beyond-use date/time, caution statements), font sizes, and translations needed locally. For home administration, include tamper-evident seals, at-home pictorial instructions, and hotline numbers. Require a label verification step with second-checker initials and printed timestamps.
Direct-to-patient (DTP) and home use. Define identity checks at delivery (two identifiers), shipment conditions (shipper type, validated pack-outs, minimum coolant mass), shock/tilt indicators, and patient instructions if a delivery window is missed. Provide return and destruction logistics for unused product; clearly separate clinical advice from logistics instructions to avoid practicing medicine outside the site.
Temperature-excursion playbook. Include a matrix of permit/assess/quarantine actions by product and excursion duration, with contacts for stability experts. Describe how to capture chain-of-temperature (logger ID, start/stop time, download, attachment to accountability). Require CAPA for repeated excursions and define vendor credits or at-risk fees where courier performance contributes.
Documentation and training. Build in checklists and sign-off blocks with the meaning of signature (“Prepared as per recipe,” “Verified concentration and label,” “Released to blinded staff”). Provide a short training deck and a change log; require acknowledgment of new versions and record refresher training after major amendments or stability updates.
Lab Manual: Collection, Processing, Chain-of-Custody, and Biobanking
Specimen map. Start with a table that ties every protocol assessment to a specimen type (serum, plasma, whole blood, urine, PBMCs, tissue), tube color/additive (e.g., K2EDTA), draw volume, fasting state, time-post-dose, processing steps, and shipping mode. Mark CtQ samples (primary endpoint, PK anchors, genomic material) with bold symbols and require dual checks.
Collection and processing. Specify order of draw, inversion counts, clot time, centrifuge settings (g-force, temperature, duration), aliquot volumes, and label conventions (participant ID, visit, timestamp, aliquot #, barcode). Provide photos of acceptable hemolysis/volume and a decision tree for redraws. For PBMCs and cell-based assays, add time-to-process limits and cryopreservation recipes (DMSO %, controlled-rate freeze, storage temperature, vial type).
Chain-of-custody and documentation. Include custody forms that travel with the specimen (who, when, condition, temperature, corrective actions). Require contemporaneous entries and barcoded scans into the lab portal/EDC to preserve ALCOA++ attributes. Tie specimen IDs to eCRF fields to avoid downstream mismatches; mandate periodic reconciliation between lab listings and EDC.
Shipping and stability. Provide validated pack-outs for ambient, refrigerated, frozen, and dry-ice shipments. State coolant quantities by expected transit time and ambient range; require at least one temperature logger per shipment. Provide UN/DOT/IATA dangerous-goods notes where applicable (dry ice as a dangerous good) and a photo guide to box assembly and labeling. Include courier pick-up cut-offs, weekend/holiday rules, and contingency plans (local storage limits, alternate courier, cancel/rebook instructions).
Reference ranges and critical values. For safety labs, explain how reference intervals are provided and versioned, how age/sex adjustments are handled, and which alerts trigger immediate PI contact. For exploratory biomarkers, state the plan for units, conversions, lot-to-lot normalization, and documentation of kit lot numbers. Provide a brief method validation summary when local labs are permitted.
Genetics and future use. If genomic or long-term storage samples are collected, specify consent checks, coding/pseudonymization, re-identification restrictions, and repository accession rules. Define long-term freezer monitoring, alarm responses, and disaster recovery (backup power, spare freezers, offsite storage). Provide a withdrawal-of-consent workflow for destruction or cessation of future use.
Home or remote collection. For decentralized designs, provide finger-stick kits, saliva kits, or dried blood spot instructions with photos, video links, and error-proofing (desiccants, capillary fill lines). Describe identity verification, packaging, drop-off/pick-up, and how to document failed attempts without shaming participants. Include stability limits for mail-in samples.
Data flow and reconciliation. Map systems (EDC, lab portal, biobank LIMS) and show how participant, visit, and specimen IDs align. Require monthly cross-checks and a single owner to resolve mismatches. Provide file formats for listings and define who signs off on freeze-thaw events or batched results before database lock.
Training, QC, and CAPA. Include a short quiz or sign-off for phlebotomy and processing steps, periodic blind duplicates for precision checks, and trend charts (hemolysis rates, % on-time shipments). Recurrent defects should trigger CAPA with design fixes (clearer photos, simpler labels, revised pack-outs), not only retraining.
Imaging Manual: Acquisition Parameters, De-Identification, Central Read, and Readiness to Inspect
Scanner qualification and acceptance. Define vendor/model eligibility, field strength, coil requirements, calibration/phantom scans, and acceptance criteria before first participant. For CT, specify kVp/mAs ranges and reconstruction kernels; for MRI, list sequence families (e.g., T1, T2, FLAIR, DWI) and geometric constraints; for ultrasound, define presets and depth/focus ranges. Provide a run-sheet for techs with critical timing (e.g., contrast injection to acquisition delays).
Acquisition parameters that protect endpoints. Present locked parameters (slice thickness, FOV, matrix, b-values, cardiac gating settings) with tolerances and a rationale box (“keeps lesion conspicuity constant across scanners”). For longitudinal studies, emphasize repeatability (same scanner/preferred coil, same positioning aids, same contrast lot class when possible). For PET, specify uptake times, blood glucose thresholds, and SUV calculation rules.
De-identification and privacy. Provide DICOM header scrub rules (PatientName, PatientID, DOB, accession #), burned-in text detection, face de-facing for head scans where appropriate, and a “no PHI in free-text” checklist. Define how anonymization IDs map to the study ID (kept only with the imaging core) and how re-identification is prevented for readers. State secure upload paths, encryption, and checksum verification.
Central read interface. Align with the central read charter: reader qualifications, masking/blinding (timepoints, modality, treatment arm), read paradigms (independent dual, consensus, adjudication), lesion selection and tracking rules, and software versions. Provide reader training cases and drift checks across time. For device or algorithm-assisted reads, state version/control of software and audit trails for changes.
Quality control and query loop. Provide a ranked list of failure modes (motion, wrong sequence, missing phases, truncated FOV, metal artifact, wrong orientation), how to detect them before upload, and what to do if discovered later (re-scan windows and safety checks). Use a standard image QC form with pass/fail and defect codes. Define query turnaround SLAs and escalation when defect rates exceed thresholds.
Endpoint-specific guidance. For oncology (RECIST-like endpoints), define slice thickness, lesion measurability, and contrast timing. For neurology, specify susceptibility-artifact mitigation and DTI constraints. For orthopedics, require positioning aids and weight-bearing rules where relevant. For cardiology, define gating, heart rates, and breath-hold coaching scripts. Consistent parameters keep adjudication rates low and precision high.
Data flow and traceability. Map data from scanner → PACS → upload gateway → imaging core archive → reader workstations → read outputs → EDC. Require immutable logs, time-sync across systems, and a five-minute retrieval drill (participant/timepoint → DICOM series → reader measurement → CSR figure). Provide file-naming conventions and a checksum policy to detect silent corruption.
Vendor oversight and metrics. Flow requirements into quality agreements/SOWs: immutable edit logs, drift monitoring, reader calibration sessions, query SLAs, and access controls. Track KPIs/KRIs: first-pass acceptance rate, defect category trends, re-scan rates, time to query resolution, and consistency of parameters across scanners/sites. Link repeated red metrics to credits or at-risk fees; require a corrective roadmap with dates and owners.
Ready-to-use checklist (paste into your SOPs).
- Pharmacy: validated storage and alarm thresholds; excursion decision tree; compounding recipes; blinding safeguards; DTP identity and pack-out rules; full accountability and reconciliation templates.
- Lab: specimen map with CtQ flags; processing and aliquot photos; chain-of-custody and barcode flows; pack-outs and logger rules; genomic/biobank governance; reconciliation cadence with EDC/LIMS.
- Imaging: scanner qualification; locked parameters with tolerances; DICOM scrub and face-de-facing; central read alignment; QC/defect codes; upload security and checksums.
- All manuals: translation-ready, accessible formatting; version control and change logs; signatures with meaning; five-minute retrieval drill from step → data → table/figure.
Bottom line. Manuals are not appendices; they are how trials run. When pharmacy, lab, and imaging instructions are clear, feasible, and tied to rationale; when blinding, temperature, timing, and privacy are actively controlled; and when evidence trails are ALCOA++ reliable, sites perform confidently and regulators can verify quickly—study after study, country after country.