A unit that ran for years can fail days after its annual strip-down. Intrusive tasks can leave faults behind, and many failure modes are not related to age, so a fixed overhaul cannot prevent them. Keep wear-out parts on the calendar and open everything else on condition evidence.
An out-of-specification result falls outside the specification or acceptance criteria and starts with a laboratory assessment. An out-of-trend result is inside the limits but breaks the pattern of earlier results. EU GMP Chapter 6 says both should be addressed and investigated, which needs data recorded so that trends can be evaluated.
WHO's technology transfer guideline (TRS 1044, Annex 4) ends a transfer with a report: results against the acceptance criteria written beforehand, deviations and changes assessed, and a conclusion stating whether the transfer is complete, authorised by the responsible persons. Stability studies, commitments and new suppliers continue after the signature.
EU GMP Annex 15 (2015) accepts concurrent validation in exceptional circumstances, where there is a strong benefit-risk ratio for the patient. The decision must be justified, documented in the validation master plan and approved by authorised personnel. A project that is running late does not meet that test.
A first site visit teaches a junior engineer more than a month of drawings, with some preparation: the latest drawing revision, induction and PPE, a tape, torch and camera, three things to find written down, and notes written up the same day. Look above a ceiling and inside a plant room.
EU GMP Annex 1 (2022) says airflow pattern studies should be performed at rest and in operation, with video recordings retained. A usable study starts as a plan: locations and interventions from the risk assessment, defined runs, fixed camera and smoke positions, a script with acceptance criteria per shot, and a QA review.
A fire damper that drops on a failed fusible link cuts the supply to a branch while the AHU still shows healthy on the BMS. Each damper needs an access door on the side the link is reset from, a closed-position switch wired to the BMS, and a recorded test on the fire code's interval.
Each SKU left on a transfer list has to be covered by validation and stability work. Before the transfer, list the SKUs with volume and market, group them by formulation, process and pack line, retire or merge the tail, choose the extremes where bracketing is justified, and order the work by value and risk.
A protocol line that says 'as per design' leaves the tester to choose the document and the tolerance on test day. EU GMP Annex 15 (2015) says protocols should define the acceptance criteria. Write a number with its unit and tolerance, the method and room state, and the source document with its revision.
Pure water has a resistivity of 18.18 MΩ·cm at 25 °C because its own ions conduct, so no treatment plant can deliver more. That is 0.055 µS/cm, against 1.3 µS/cm for the USP Purified Water stage 1 limit. Resistivity sees only ions, so UPW specifications add TOC, particles, dissolved oxygen and silica.
Two mixers of the same volume are not equivalent if one tumbles and the other uses a ribbon. The FDA's SUPAC guidance speaks of equipment of the same design and operating principles, of the same or a different capacity. WHO's guidelines on technology transfer say the comparison should cover working principle, capacity, make and model.
An SOP that becomes effective the moment it is signed puts the next shift in deviation before anyone is trained. The 2011 text of EU GMP Chapter 4 says the effective date should be defined; the gap after approval is where training, the swap of controlled copies and the register update take place.