A reactor in Block C finishes its reflux at 2.40 in the morning. The operator notes the temperature on a scrap of paper. The batch record sits on a desk two rooms away, and the second reactor is already calling for a solvent charge. He fills the record at 5.30, from memory and that scrap. The QA reviewer notices that three readings share the same handwriting slant, the same pen and the same minute. QA parks the lot and opens an investigation.
It heads the list of strains on the API makers page, and it is rarely about honesty. A missed or late entry in an active ingredient plant usually starts with how the shift is staffed and laid out.
What a slip looks like beside a reactor
- Charging quantities written after the drum is already empty, with the weighing slip stapled on later.
- In-process readings such as pH, temperature or pressure entered in one tidy column just before handover.
- Done-by and checked-by signed by the same hand, or by a checker who arrived after the step.
- Centrifuge spin times and dryer loss on drying samples copied from a neighbour's sheet for the previous batch.
- Overwritten figures, correction fluid or a torn page in an equipment logbook.
- Times that clash with the DCS trend or the balance printout attached to the record.
Why the slip is usually a staffing problem
Reactor work is lumpy. A reaction can idle for hours and then demand three actions in ten minutes: a dosing start, a sample pull and a jacket temperature change. If one operator covers two vessels at that moment, something gives, and the paper loses first. The fault looks like carelessness in the record. On the floor it was a roster decision.
Checkers are the second gap. Many blocks run nights with a single shift chemist who is also expected to clear the centrifuge, answer stores and take QC calls. A checker who cannot be at the vessel will sign afterwards. Within a few months that becomes the custom, and new operators learn it as normal.
Layout plays its part too. Where the batch record lives on a desk by the block office, every entry costs a walk. Operators batch their writing to save those walks. Where the record or a clipboard stage sheet sits at the vessel, entries tend to be made at the moment.
What the slips cost
A held batch ties up a reactor, a dryer or a warehouse bay until QA closes the investigation. The block head and the reviewer lose days writing it up. Customers who audit the site read those deviation files and ask about them. And when management replies by issuing the operator a warning, it often loses someone who took years to train. The colleague who replaces him learns the same shortcut in the same roster, and the cycle restarts.
Repairs that begin on the floor
- 1
Sort three months of deviations by shift, step and role
Look for clusters. Slips that pile up on nights, at charging or during dryer unloading point to a staffing or layout cause rather than one careless person.
- 2
Walk one night shift with the batch record
The block head or HR lead stands in the block from midnight to six and notes every moment an operator has two things to do at once. That list is the real root cause document.
- 3
Put an entry point at each vessel
A stage sheet on a clipboard at the reactor, a fixed pen and a wall clock in sight. Remove the reason to carry readings in a pocket.
- 4
Roster a checker who is free to check
On busy campaigns, name a second person per shift whose first duty is to witness charges and critical steps. Write it into their role, not as a favour.
- 5
Practise entries before the first live batch
Joiners fill a mock batch record on a dummy reactor run, plus one deliberate correction struck through properly. QA reviews it with them line by line.
- 6
Change the response to a slip
After a slip, the first question becomes what blocked the entry, asked by the supervisor within the shift. Repeated, deliberate falsification still goes through the disciplinary route, but most slips never were that.
Roles that make checking possible
Many API units write one job description for every block operator, so nobody owns witnessing. A clearer ladder helps: reactor assistant, operator, senior operator able to check, and shift chemist who reviews. Each level gets the steps it may perform and the steps it may verify. Writing those levels and linking pay to them is covered under job levels in API blocks. Done well, a senior operator sees checking as earned status, not unpaid extra work.
Training that sticks past the first month
A read-and-sign session on documentation rules fades within weeks. What holds is short, regular practice tied to the block's own forms. A fifteen-minute huddle each month on one real correction, shown on screen with names removed, teaches more than a yearly lecture. The refresher planning notes for API units lays out a refresher rhythm that fits campaign schedules.
- Show one anonymised slip each month and ask the shift how it could have been avoided.
- Rotate who demonstrates the correct correction method, so operators teach as well as learn.
- Pair each joiner with a senior operator for their first three campaigns, not just their first week.
- Record attendance at huddles in the training file, so the habit is visible to auditors and to the block head.
Further reading
The wider pattern across tablets, capsules and API units appears in paperwork errors and training gaps. Recurring people issues across the group are gathered on the pharmaceutical manufacturing page. For an outside look at block rosters and role levels, describe the site in the requirements wizard.



