Why a good SOP earns its keep

A standard operating procedure is more than a document that keeps an assessor happy. Done well, it is your laboratory's memory: a written record of how a method is actually performed, by whom, with which materials, and to what acceptance criteria. In clinical diagnostics this underpins patient safety — a mislabelled sample, an extended lysis step or an unrecorded reagent change can all shift a result. In biotechnology and research settings, the same discipline protects the integrity of data that may later support a regulatory submission or a patent.

Assessors working to ISO 15189 for medical laboratories, or ISO/IEC 17025 for testing and calibration, rarely ask simply whether you have an SOP. They ask whether bench practice matches the document, and whether staff can explain why a step exists. That gap between paper and practice is where most findings begin, so it is worth treating SOP writing as a technical skill rather than an administrative chore.

Start with the task, not the template

Before opening a document, watch the method being performed. Stand at the bench, ask the person doing the work where they hesitate, and note the small workarounds that never made it into the last version. Then agree the scope: one SOP should cover one process, with a clear beginning and end. An extraction SOP should not quietly absorb the downstream quantification step.

Every controlled document needs a header block that makes it traceable. At minimum, include:

• Title and unique identifier — for example, SOP-MOL-014, so the document can be cited in training records, deviation reports and audit trails.
• Version number, approval date and effective date — kept deliberately separate to allow a training window.
• Author, reviewer and approver — with named roles rather than a department alone.
• Scope, applicability and any exclusions — which sample types, instruments and sites the procedure covers.
• Review date and a short change history — so a reader can see what altered and why.

Writing steps people can actually follow

Write in the imperative and put one action per step. "Pipette 200 µL of lysate into a 1.5 mL tube and vortex for 10 seconds" is usable; "the lysate should be transferred and mixed" is not. Be specific about volumes, concentrations, temperatures, timings, centrifugation speeds, plate formats and equipment identifiers. State tolerances where they matter, and state the acceptance criteria that tell an analyst whether the run has passed.

Avoid phrases such as "as appropriate", "standard conditions" and "as required". They transfer the decision back to the reader and make the procedure impossible to audit. Where a judgement call is genuinely needed, describe what that judgement looks like and who is permitted to make it.

Include the practical detail that surrounds the science: required personal protective equipment, the relevant COSHH assessment reference, waste disposal routes and what to do if a spill occurs. Set out sample requirements, storage conditions and rejection criteria, along with the limits of the method and the action to take if a control fails. A competent new starter should be able to follow the document from beginning to end without asking a colleague for the missing pieces.

Version control and document management

Version control fails when copies multiply. Keep a single controlled master in your document management system, and treat printed bench copies as short-lived. If paper copies are needed for a cleanroom or a category 3 containment lab, stamp them as controlled, log their issue, and withdraw them when the procedure changes. Superseded versions should be removed from circulation, with one archived copy retained for traceability.

A simple numbering scheme works well. Use a major increment for changes to principle, reagents, instrument settings or acceptance criteria, and a minor increment for typographical corrections or updated contact details. Every change should be logged with its date, description, reason and author. Most laboratories set an annual or biennial review cycle, but any of the following should trigger an earlier revision:

• A change of instrument, reagent supplier or kit formulation.
• An updated kit insert or manufacturer's instructions.
• An incident, near miss or out-of-specification result traced to the method.
• New or revised regulatory or accreditation guidance.
• Persistent poor performance in external quality assessment.

Training, competency and read-and-understood records

An SOP that nobody has been trained against is not really in force. Build a training matrix that maps each member of staff to the procedures they are authorised to perform, and keep it current. A robust training record usually combines several elements: induction reading, supervised practice, direct observation of the trainee performing the whole method, and a signed competency assessment by a suitably qualified assessor.

Read-and-understood signatures have their place, but they prove awareness rather than competence. Competency should be reassessed at defined intervals, after a prolonged absence from the bench, and whenever the SOP is revised. Always record which version of the document someone was trained against — it is the first thing an assessor will look for when a procedure has recently changed.

Keeping SOPs alive

The best procedures evolve. Make it easy for staff to flag a step that no longer matches reality, and review those comments in team meetings rather than letting them sit in an inbox. Deviations, audit findings, incident reports and quality control trends are all useful fuel for revision, and feeding them back into the document shows that your quality system is working as intended.

Write in plain English, keep sentences short, and let the document be as long as the method demands and no longer. A procedure that people can follow confidently at seven in the morning, on a busy day, with a full sample run, is one that will hold up under scrutiny — and that is the whole point.