National HPV screening programmes have shifted from cytology-based to HPV-based primary screening across many regions, and tender documents have sharpened accordingly. Bids that stop at a CE mark and a "high-risk HPV DNA detected" claim are hard to evaluate against a genotype-specific triage algorithm, because the algorithm itself demands more. This review covers the four evidence layers procurement teams should verify before shortlisting an HPV assay bid.
Set the analytical floor first
Three analytical parameters carry most of the weight. Limit of detection should be reported per genotype or genotype group, not as a pooled figure, because performance on HPV16 and HPV18 — the genotypes that drive triage decisions — is what the algorithm depends on. Clinical sensitivity for cervical intraepithelial neoplasia of grade two or worse, and clinical specificity, should be evidenced against an accepted comparator in the intended-use population, not only against a reference PCR on convenience samples. Finally, check how the assay reports results: pooled high-risk positivity without partial genotyping cannot support an HPV16/18 triage pathway and will surface as a compliance failure at evaluation, however strong the underlying chemistry.
Genotype coverage and triage fit
Screening algorithms typically triage on HPV16 and HPV18 positivity with reflex cytology, and manage the remaining high-risk genotypes through follow-up intervals. That makes partial genotyping the minimum viable configuration for most programmes, with extended genotyping valuable only where the national algorithm uses it. Verify coverage against the genotype hierarchy the programme recognises, and confirm the assay quantifies or at least flags potential high viral-load infections where the algorithm differentiates. An assay whose genotype panel fits the algorithm exactly is worth more than a broader panel that the algorithm ignores — broader coverage that is not used adds validation burden without clinical return.
Grade the clinical evidence
Not all evidence in a bid deserves equal weight. Peer-reviewed clinical performance studies in screening populations sit at the top; manufacturer-sponsored studies with independent statistical analysis follow; internal analytical verification alone is a floor, not a case. For self-sampling workflows — now specified in a growing number of programmes — demand performance data per specimen type, since clinician-taken and self-taken samples behave differently at low viral loads. Longitudinal evidence, registry linkage, and head-to-head comparisons against the incumbent method in the target population are the strongest signals that claimed performance will survive contact with the programme. Where local population data is absent, ask for it to be generated as a condition of award rather than accepted on faith.
Programme integration requirements
- Daily workload fit: batch structure, hands-on time, and turnaround against the screening programme reporting timeline
- Reflex testing pathway: whether cytology triage runs on the same platform, a partner platform, or a separate workflow
- Laboratory information system interoperability, order codes, and result messages aligned to the national reporting format
- Quality assurance participation: proficiency testing schemes, external quality assessment enrolment, and documented corrective-action process
- Training and implementation support committed in the bid, with named personnel and timelines
Tender evaluation that works through these four layers in order — analytical floor, algorithm fit, evidence grade, integration — produces shortlists that survive implementation. Skipping the order, in particular grading evidence before checking algorithm fit, is how programmes end up with a well-evidenced assay that cannot serve the triage pathway they committed to.