How to write a glove specification for a public tender
A glove tender is decided when the specification is written. Which standards to cite and in which editions, how to specify AQL, what documentation to require, and how to write award criteria that can actually be evaluated.
A glove tender is decided long before the bids are opened. It is decided when the specification is written, because the specification determines what the market is allowed to offer and what the evaluation is able to see. A vague specification does not produce a fair competition. It produces a competition on price alone, and it hands the award to whoever is most comfortable claiming things they will not be asked to prove.
This article is about the parts of a glove specification that decide outcomes: the standards you cite, the editions you cite them in, the documentation you require, and the award criteria you can actually evaluate. It is written from the supplier side of the table, which is exactly why it is worth reading. Suppliers know which clauses do work and which are decorative.
Key takeaways
What actually goes wrong in a glove specification?
Four failures account for most of it, and all four are avoidable at the drafting stage.
The citation with no edition. A requirement that reads "gloves shall comply with EN 455" cannot be failed. The series has five parts and several live editions, and a supplier holding a test report against a superseded edition complies with the sentence as written. This is the single most common gap, and it is invisible unless someone checks the edition line on the report.
The unverifiable adjective. "High quality", "durable", "soft", "environmentally friendly". None of these can be scored consistently across bidders, and an evaluation that tries will not survive a complaint.
The criterion with no evidence attached. If the specification asks for something but never says which document proves it, the requirement resolves to whoever asserts it most confidently.
The single-number award. Awarding points for a higher figure on one test is the most seductive mistake, because it looks objective. It is discussed below, because it deserves its own answer.
Which standards should the specification cite, and in which edition?
Cite the part, the edition and, where it matters, the clause. This is the spine of the document.
Two points from that table change how a specification should be written.
Only EN 455-1:2020+A2:2024 and EN 455-2:2024 are harmonised under the Medical Device Regulation, added by Commission Implementing Decision (EU) 2025/681 of 8 April 2025. Presumption of conformity attaches to those editions and not to their predecessors. A supplier with reports against EN 455-2:2015 is outside the presumption even where the underlying test was identical. If you want that legal effect, name the edition.
EN 455-5:2025 sets no limits. It standardises how extractable chemical residues are measured, using 75 % ethanol as the extraction medium, and it explicitly does not provide information on allergenic potential or user safety. That makes it genuinely useful in a specification, and useless in the way people often try to use it. You can require testing to EN 455-5 and require the results to be reported in µg/g with the limit of detection and limit of quantification stated. You cannot require a glove to "pass EN 455-5", because there is nothing to pass.
A related trap: EN 455-2 does not specify elongation. Any elongation figure on a datasheet comes from somewhere else, usually an ASTM material specification. A specification that demands an elongation minimum "per EN 455-2" is asking for something that standard does not contain, and every bidder will answer it differently.
For what each part covers in full, see our guide to the EN 455 series. For chemical protection, which is a different question from medical glove compliance, see EN ISO 374.
How should AQL be specified?
State the value, the inspection level and the sample size code letter, or you have not specified a sampling regime at all.
EN 455-1 requires standardised AQL tables, single or multiple sampling plans, general inspection level I, with minimum sample size and accept and reject numbers equivalent to sample size code letter L. The acceptable quality levels are 1.5 for examination gloves and 0.65 for surgical gloves.
Two things buyers routinely get wrong here.
The first is treating AQL as a quality rating. It is not. It is the maximum proportion of defective units a lot may contain and still be accepted, for one defect: leakage. It says nothing about comfort, chemical resistance, ageing or consistency between lots. We have written about that at length in what AQL actually means.
The second is specifying a tighter AQL than 1.5 in the belief that it buys better gloves. It mostly buys a larger sample and a higher rejection rate at inspection, which the supplier prices in. If your concern is batch-to-batch consistency, the effective lever is not a tighter AQL but a requirement that every lot is inspected before shipment, and that you receive the report. That is a different clause, and a far more useful one.
What documentation should the tender require?
Ask for documents that exist. A specification that demands paperwork nobody issues teaches bidders to write around the requirement.
More detail on what a buyer can and cannot verify under MDR is in our MDR article.
How do you write award criteria that can be evaluated?
The test is simple. For every criterion, write down the document or measurement that closes it. If you cannot, the criterion is decoration and it will be argued about later.
Do not award points for absolute force at break. This is the recommendation most likely to be resisted, so here is the basis for it. EN 455-2:2024 states in Clause 5.1 that different glove materials require different force at break requirements to achieve acceptable performance, and that absolute force at break values do not directly correlate with in-use performance. Material selection belongs in risk management, not on a points scale. The requirement should be a pass or fail against the threshold for the material: 9.0 N for surgical gloves, 6.0 N for examination gloves, 3.6 N for thermoplastic materials such as PVC and PE. Awarding above the threshold rewards a thicker, stiffer glove that users then complain about, and we have seen tenders won exactly that way.
Useful criteria to score instead, all evidence-backed:
The last one is worth the effort. A glove that documents perfectly and that nurses dislike will be worked around, and the workaround is usually double-dispensing, which costs more than the tender saved.
What about sustainability requirements?
Specify the test method and the conditions, or expect claims you cannot compare.
Biodegradability is the usual example. A glove described as biodegradable is normally tested to an anaerobic method such as ASTM D5511 or ASTM D5526, and those methods model different environments and produce different numbers. A criterion that says "biodegradable" invites every bidder to answer against whichever method flatters them. A criterion that names the method, the test duration and the percentage achieved is comparable across bids. We have set out the difference in reading a biodegradability claim, and the test data behind our own biodegradable nitrile glove is stated in those terms for the same reason.
The same discipline applies to carbon and life cycle claims: require the study, its scope and its boundaries rather than the headline figure. A life cycle assessment with a cradle-to-gate boundary is not comparable with a cradle-to-grave one, and the difference is invisible in a marketing number.
Questions we get asked
Can we name a brand in the specification? No, and it is rarely in your interest anyway. EU procurement rules do not permit a specification that favours a particular product without an "or equivalent" allowance, and in practice naming a brand imports its datasheet quirks into your requirement. Specify the performance and the evidence instead.
Should we require samples? Yes, and require them from production stock rather than a sample-room batch, with the lot number stated. Then evaluate them against the same criteria you published. Samples assessed informally are a common source of challenge.
Is a CE mark enough? No. For most gloves the CE mark reflects a self-certified declaration, not third-party testing of barrier performance. On sterile surgical gloves a notified body reviews the sterility aspects, not the whole product. The CE mark tells you a declaration exists; the test reports and their editions tell you what it is worth.
How long should a glove specification be? Shorter than most. Every clause you cannot evaluate is a clause someone will argue about. A specification that names the right standards in the right editions, sets pass or fail thresholds, lists the documents required and scores a small number of evidence-backed criteria will do more work than twenty pages of adjectives.
Where to start
If you are drafting one now, the fastest improvement is mechanical: go through the document and add an edition to every standard you cite, then delete every requirement for which you cannot name the proving document. Most specifications shrink by a third and get considerably harder to game.
Our quality team reviews PPE specifications for public buyers before publication at no cost, covering standards, editions, test methods and the documentation to require. If a draft would benefit from that, get in touch.
Seventeen years sourcing and supplying medical gloves and PPE to governments, UN agencies, NGOs and healthcare systems. Writes on glove standards, quality management and responsible sourcing for institutional buyers, working from the source standards rather than secondary summaries.
View profile on LinkedIn
