Why Sample-Time Review Matters
A concentration record can be linked to the correct subject, sample, and assay result yet still be unusable for time-dependent analysis if its collection time is absent or wrong. A small time error can matter when concentrations change rapidly; a larger error can place a sample on the wrong side of a dose event. The objective of review is not to make every record look plausible. It is to establish what the source evidence supports, correct verifiable errors, and identify values that cannot be determined reliably.
The distinction between nominal and actual sampling time is essential. A nominal time describes the planned schedule; an actual time describes when the specimen was collected. The actual time may differ from the plan and remains the relevant observation when it is documented. Replacing an absent actual time with the scheduled time can conceal a delay rather than resolve it.
Before investigating the time itself, confirm that the record represents a collected specimen and that it is linked to the intended subject and sample. A missed collection is not a collected sample with a missing timestamp. Likewise, a laboratory receipt time is not automatically the collection time: transport and processing can separate those events. Earlier material on linking subjects, samples, doses, and occasions and on distinguishing nominal from actual sampling times provides the record structure and time concepts used here.
What Counts as Evidence?
Review the source fields that bear directly on the sample and its chronology. Depending on the study, these may include a collection worksheet, a specimen label, a clinic or nursing record, an electronic source-system timestamp, visit information, and the laboratory accession record. Related dose-event records can help test chronology, but they do not by themselves establish when a specimen was collected.
Evidence must be linked to the same subject and specimen, and its meaning must be clear. A timestamp for specimen receipt, data entry, or laboratory processing should not be treated as a collection timestamp merely because it is complete. If two credible collection records disagree, a value is not correctable simply because one agrees better with the planned schedule or appears more pharmacologically plausible.
For a sample collected at timestamp \(T_s\), an elapsed time relative to a documented reference timestamp \(T_0\) is calculated as
Both timestamps must use compatible date, clock-time, and time-zone conventions, and the analysis must express the difference in the specified unit. For example, a difference of 77 minutes is \(77/60=1.283\) hours. The reference event matters: elapsed time since a dose is not interchangeable with elapsed time since a study origin. A sample collected before a dose may correctly have negative time when dose time is the reference.
A Defensible Review Sequence
Check that the row represents a collected sample and that subject, sample identifier, occasion, and related fields are consistent. If the sample was not collected, do not create a collection time.
Inspect records that document collection and determine whether each timestamp refers to collection, receipt, entry, or another event. Confirm that the source belongs to this specimen.
Use the documented collection timestamp and the correct reference event. Check date, time zone, event ordering, and unit conversion before comparing the result with the analysis field.
Record whether the time is verified as entered, corrected from evidence, confirmed as missing, or unresolved. Do not use the planned time as a substitute for an unknown actual time.
Review other records from the same source, visit, or data transfer for repeated offsets, missing fields, or event-label confusion. A systematic pattern may indicate a process problem, but does not justify unsupported individual corrections.
Worked Examples
Worked Example: Recovering a Missing Time From a Collection Record
A hypothetical subject has a documented dose reference time of 08:00. A linked collection worksheet identifies the same subject and sample and records collection at 09:17. The analysis dataset has no actual sample time, while the nominal sample time was 1 hour after the dose. Assume the source timestamps have been checked for date and time-zone consistency and both refer to the same study day.
The source establishes that the sample was collected. The elapsed time is calculated from the actual collection timestamp, not copied from the nominal schedule:
The missing analysis value can be populated as 1.283 hours because the linked collection record supports that value and the reference event is known. The nominal time of 1 hour is not the appropriate replacement: it differs from the documented collection time by 17 minutes. The correction should be traceable to the source record.
Worked Example: Investigating an Out-of-Range Time
In a hypothetical record, the analysis dataset reports a sample time of 2.54 hours after a dose at 08:00. The nominal visit was at 24 hours, so the value appears inconsistent with the plan. The specimen label and collection worksheet both link the sample to the subject and record collection at 09:24 the following day. Assume the date and clock-time fields are consistent.
The source timestamp gives:
The data value 2.54 does not match the source-derived 25.40 hours and is consistent with a misplaced decimal digit. Because the source records support a specific collection time, 25.40 hours is a defensible correction. The fact that this is later than the nominal 24-hour visit does not make it invalid: actual collection can be delayed. The review corrects the transcription discrepancy, not the delayed visit.
Worked Example: Leaving Conflicting Collection Times Unresolved
A hypothetical sample is linked to a dose at 07:00. The analysis record gives 2.70 hours. A collection worksheet records 09:12, while a nursing note linked to the same sample records collection at 09:21. A laboratory record shows receipt at 09:42. No source note explains the nine-minute disagreement between the two collection records.
The two possible elapsed times are:
The 09:42 receipt time cannot resolve the discrepancy because it documents receipt, not collection. Neither collection record is shown to be more authoritative or to be a transcription of the other. Therefore, the analysis value of 2.70 hours is unsupported, but the evidence also does not establish whether 2.200 or 2.350 hours is correct. The time should remain unresolved pending clarification. Choosing the value closer to the nominal schedule, averaging the two times, or using the receipt time would create a value not supported by the evidence.
Diagnostics, Alternatives, and Limitations
Useful diagnostics include counts of missing, corrected, and unresolved times; comparisons of actual with nominal times; and checks for repeated discrepancies by site, visit, source system, or data-transfer batch. Inspect both extreme values and suspiciously repeated values. A cluster of identical offsets may point to a time-scale or field-mapping problem, while isolated discrepancies may be record-specific. These patterns guide investigation; they do not prove what happened for an individual sample.
A plausible value is not necessarily a verified value, and an unusual value is not necessarily an error. A sample can be collected outside its planned window, before a dose, or at a time that appears surprising. Conversely, a value that falls within the expected visit window may still be wrong. Plausibility screening is a reason to check evidence, not a substitute for it.
| Approach | Appropriate use | Main limitation |
|---|---|---|
| Source-based correction | A linked source record supports a specific collection time and elapsed-time calculation. | Cannot resolve a conflict when credible records disagree without a basis for choosing between them. |
| Use nominal time | Describes the planned schedule or supports a check against actual timing. | Does not establish the actual collection time and can conceal deviations. |
| Use a plausibility bound | Flags values for review when a study has a justified rule for identifying suspicious records. | A bound cannot determine the correct time or establish that an outlying sample is erroneous. |
| Leave time unresolved | Evidence is missing, ambiguous, or conflicting. | The record may not support time-dependent analysis until clarified; fabricating a value is not a remedy. |
A common mistake is to replace a missing actual time with the nominal time because the sample was intended for that visit. The reasoning confuses the planned collection with the observed collection. The correction is to seek collection evidence; if no reliable evidence exists, retain the time as missing or unresolved under the dataset’s documented rules. Another mistake is to treat a negative elapsed time as automatically impossible. Its meaning depends on the chosen time origin and event ordering: a sample before a dose can legitimately precede that dose, whereas a negative value may be inconsistent with a different, study-wide origin.
Review also has limits. Source records may be incomplete, timestamps may have been entered after collection, and several records may inherit the same underlying error. A calculated elapsed time can be arithmetically precise while the source time itself is uncertain. Do not report more precision than the source supports, and do not infer the collection time from a concentration profile or from the timing that would make a model fit better.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- A sample was collected at 14:30, and its documented dose reference time was 14:00 on the same day. The analysis time is blank. What elapsed time is supported?
- A nominal sample is scheduled at 4 hours, but a verified collection record gives 5 hours after the dose. Which time describes the actual sample?
- A sample time is \(-0.20\) hours relative to a dose. Must it be rejected? State what must be checked.
- Two linked collection records give different times, and neither is shown to be more reliable. Should the value be corrected to the time closer to nominal?
Solutions. 1. The elapsed time is \(14{:}30-14{:}00=30\) minutes, or \(30/60=0.500\) hours; the linked collection record supports this value. 2. The actual time is 5 hours because the source documents actual collection; 4 hours is the nominal schedule. 3. No. A negative time can be valid if the sample was collected before the dose used as the reference. Check the time origin, linked event, date, and event ordering before judging it. 4. No. Nominal proximity does not establish which source is correct. Leave the time unresolved unless further evidence supports one value.