A cold-chain record is more than a temperature graph. It is a connected account of the goods, the device, the route, the custody changes, any system alerts and the review that followed. A reviewer should be able to understand what the data covers, identify missing or transformed information and distinguish direct observation from interpretation.
Assemble the evidence set
Start with the shipment and product references used by the organization. Add the logger or sensor identifier, stated placement, activation and stop times, measurement file, calibration status, device configuration and export details. Include route events and timestamps, custody or handover notes, alert history, delivery documentation and relevant communications. Not every shipment will have every record; make the expected sources and any unavailable items explicit.
Retain original files in a controlled manner and keep a separate working copy for analysis. Record when a file was retrieved, who retrieved it, from which system and whether it was converted. If a graph is generated from raw samples, keep the underlying series and note any filters, averaging, omitted values or unit conversions. A screenshot may help show what a user saw, but it should not replace a machine-readable source file where one exists.
Example record layout
The rows below are an illustrative structure only. Their entries do not represent a real shipment, validated limits or product advice.
| Evidence item | Illustrative entry | Review question |
|---|---|---|
| Shipment reference | Example consignment | Does this match the package and dispatch documents? |
| Logger identifier | Logger A-01 | Is this the device linked to the shipment? |
| Measurement window | Dispatch to receipt | Do recorded start and stop times cover the journey? |
| Placement | Documented load position | Does the position support the stated interpretation? |
| Exceptions | Gap or alert noted for review | What evidence explains or remains unresolved? |
| Decision record | Assigned reviewer and rationale | Is the decision authorized and traceable? |
Reconcile the timeline
Align the measurement period with dispatch, route events, transfers and receipt. Do not assume that all systems use the same clock or time zone. Preserve original timestamps and record any normalization used to compare them. If a sample appears before loading or after receipt, determine whether the device was running outside the shipment period and state how that affects interpretation.
Look for discontinuities in the sensor series, unexpected changes in sampling interval, logger resets, duplicate records and delayed uploads. A gap should be described by its start and end where possible, along with the checks performed. Do not interpolate missing values unless a documented method is appropriate for the analysis; even then, label calculated data clearly and retain the actual observations.
Review handover evidence
Handover records connect events to responsibility. Compare receipt and dispatch times, parties, shipment identifiers, seals or packaging condition, visible device status and any noted delay or exception. A signature or scan may document custody without verifying environmental conditions. Conversely, a valid sensor record may not identify who had responsibility during a particular interval. Review both sources and state what each one establishes.
Assess source, integrity and context
For every item, ask who created it, what system or device produced it, whether it was changed and whether the person reviewing it can reproduce the transformation. Preserve version history and access controls according to organizational requirements. If a supplier report contains only a summary, ask whether the raw data and configuration can be provided. Be careful with charts that hide scale, omit intervals or display rounded values without showing the underlying data.
Context matters as much as completeness. Identify the product-specific requirements and the approved monitoring plan used to assess them. A sample value cannot establish a universal threshold. If a reading is outside the applicable requirement, follow the relevant quality procedure and route the issue to the authorized decision-maker. Avoid writing a final disposition into an evidence summary unless it is supported and approved.
Write conclusions in layers
A strong review states observations first: what a named device recorded, over what time period and at what position. It then describes system events: which alert rule triggered, if any, and when it was acknowledged. Finally, it states the reviewer’s assessment, references the procedure applied and records the responsible role. Where the evidence does not support a confident answer, say so plainly.
Records should be retained, protected and retrievable under the applicable quality system, contract and legal obligations. Retention duration depends on those requirements; there is no single period suitable for every product or organization. Periodically test that records can be exported and read after a system change. A record that exists but cannot be retrieved or interpreted is weak evidence at the moment it is needed.