Choosing Software to Run a Proficiency Testing Scheme: A Checklist for PT Providers

Why the software matters

A proficiency testing scheme is judged on more than its statistics. An ISO/IEC 17043 assessor follows a round from the test items you prepared to the report each participant received, and asks for the evidence at every step. When that evidence sits in email threads and spreadsheets, it has to be pieced together before every assessment. The system you run the scheme on decides whether that record is kept as the round runs.

The questions below apply to any system. Each one ends with how EqasOnline handles it.

1. Can you describe your scheme in your own terms?

Your tests, measurands, units and sample types should be yours to define, whatever you test. A system with fixed fields for one sector gets in the way of a scheme in another.

In EqasOnline: you build quantitative or qualitative tests with your own analytes and units, and sample types are your own list; nothing clinical is pre-set.

2. Are test items checked before they ship?

Homogeneity and stability are part of the scheme's evidence. Ask whether the check is recorded against the batch, and whether a failing batch can still be dispatched.

In EqasOnline: an ISO 13528 Annex B homogeneity and stability check runs before samples ship. A failing batch cannot be dispatched until a remedy is recorded, and the evidence exports to Excel and PDF for your assessor.

3. Who enters the results?

Every result re-typed from an email or a returned form is a chance for a transcription error that someone then has to trace. Results should come from the laboratory that measured them, against the units and parameters you set.

In EqasOnline: participating labs enter their results in their own portal. Implausible values and unit slips are queried at entry, and the lab's confirmation is stored. Labs can edit until the deadline; your staff can correct results afterwards.

4. Do the statistics show their working?

Ask which assigned-value method is used, what happens with small groups, and whether you can print the calculation for an assessor.

In EqasOnline: one server-side engine scores every round the same way, following ISO 13528. It uses Algorithm A for 12 or more results and the median and MADe for 3 to 11; no z-score is given from fewer than three results, and the reason is recorded. Each result gets a z-score and % bias, and the band is always given in words: satisfactory, questionable or unsatisfactory. A detailed evaluation report prints every Algorithm A iteration.

5. Do you control when reports go out?

Participants should see their results only when you release them, and an issued report should not change quietly afterwards.

In EqasOnline: reports carry your logo, seal, signatures and document control code, and go out on your release date; one click notifies every lab. Issued score and analysis reports are versioned and SHA-256 recorded. An amendment carries a written reason, notifies the lab and keeps the earlier version on record.

6. What will your assessor find?

Complaints, appeals and corrective actions need an owner and an outcome, and changes to data need a trail.

In EqasOnline: out-of-range results open a corrective-action record your reviewer accepts or returns, and complaints and appeals run in a thread that closes with an outcome and a reason. An append-only audit trail keeps before-and-after values for three years and exports to CSV.

7. Can you take your data with you?

Your scheme's records are yours. Ask what can be exported, and in which formats.

In EqasOnline: results export to Excel, the analysis and the audit log to CSV, and reports to PDF.

The checklist

• Can we define our own tests, measurands, units and sample types?
• Is the homogeneity and stability check recorded, and does it hold back a failing batch?
• Do participants enter their own results, with checks at entry?
• Is the assigned-value method stated, including for small groups?
• Can we print the calculation for an assessor?
• Do we decide when reports are released, and are issued reports versioned?
• Are corrective actions, complaints and appeals tracked to an outcome?
• Is there an append-only audit trail we can export?

See it on your own scheme

The best test is your own scheme structure. Book a 30-minute demo and we will run one full round in both portals, from item check to issued report.