Beyond the certificate: a batch record checklist

A certificate of analysis answers a narrow question: what was measured on the sample that was tested. The batch record answers a wider one: how the lot in front of you was assembled, by whom, in which room, and on what date. The two are usually filed together, which is why they get read as if they were interchangeable. They are not, and the gap becomes visible the first time you have to explain a result to someone who was not at the bench.

The certificate is a summary, not the source

The certificate is compiled from the record, never the other way round. Each value on it has an entry behind it, and that entry names the operator, the instrument and the time. When a reviewer asks how a figure was obtained, the answer sits in the record. The certificate establishes only that a figure exists, not how it was produced.

A practical test costs a minute: take any figure from the certificate and try to follow it back to the entry that produced it. If the trail stops at the certificate, the file is incomplete no matter how many pages it holds. Running that test on the first lot of a study tells you what the rest of the file will look like.

Dispensing and weighing entries

The dispensing section states what came out of stock and what the balance displayed: component, target quantity, quantity actually dispensed, the balance identifier, and both dates, weighing and use. Differences between target and actual belong here, with the correction and its reason where one was made. This is the section pulled when a concentration will not reconcile, and again when a later lot has to be set against an earlier one. Reading it takes a minute and saves an afternoon.

Two dispensing entries for the same component can also be set against each other, which is useful when one batch was split across two weighing sessions. Read side by side, they expose a routine slip as well: a net quantity logged after the container went back on the balance, with the tare value stranded in a separate note.

Material input and the fill calculation

The input table lists every component charged to the vessel, its lot number, and the order of addition. Beside it sits the calculation that produced the fill volume or the vial count, normally with the arithmetic shown rather than only the result. This is where an overage is either documented or missing. It is also the field that explains why two lots built from the same stated amount can yield different numbers of vials.

Equipment, containers and room conditions

Before a component is charged, the record should show what held it: vessel identifier, cleaning state and the date of that cleaning, and whether the container was dedicated to one material or shared across several. If a residue question is ever raised, this line is the first thing pulled. A blank cleaning status is not a clean vessel; it is an unanswered question.

Temperature and relative humidity sit in the same part of the document, each with the time of reading and the device that produced it. Where a room is monitored continuously, the record usually cites the monitoring output instead of repeating every value. Those readings matter when a run has to be bracketed against the conditions it was made under, and when two runs months apart are being compared by someone who was not there for either.

Deviations and remarks

Every record needs a place for what did not go to plan: an interrupted power supply, a balance that drifted, a filter that clogged, an entry corrected after the fact. That place is not evidence of poor work; it is evidence that the work was watched while it happened. A record with no deviation entries across many consecutive lots deserves a second look, because a blank section usually means the section was ignored rather than that nothing occurred.

• Dispensing entries, when a concentration or a yield will not reconcile
• Input table, when two lots from the same stated amount differ in output
• Cleaning status, when a container is reused or a residue is queried
• Room readings, when a run must be bracketed against its conditions
• Deviation notes, when a reviewer asks what changed and at which step

The list is not a filing rule. It is a map of which page answers which question, and knowing it in advance is the difference between a five-minute check and a week of correspondence.

The release signature

The last page carries the decision: who reviewed the record, against which criteria set, and on what date. A signature without a named criteria set is decorative. Read the two together, and note whether the reviewer is also the person who performed the work; keeping those roles apart is the entire purpose of the page.

Taken in sequence, these sections answer three things a certificate cannot: what went in, what happened while it was there, and who accepted the outcome. A study that retains only the certificate retains answers to a much smaller set of questions, and usually discovers this at the least convenient moment.