Two continents wrote the same list. The readiness review checks it.
NIH's rigor and reproducibility criteria and the EQIPD Quality System were written independently, for different research cultures, by people asking the same question: what has to be true of a preclinical study before its result can be believed? Below is their overlap, requirement by requirement, and what Study Manager does about each — verified, available, or planned.
NIH Rigor and Reproducibility
Since 2025, Rigor and Feasibility is one of two numerically scored factors in NIH review. The four rigor criteria: rigor of prior research; rigorous experimental design including randomisation and blinding; consideration of biological variables; authentication of key resources. In 2026 NIH launched a Replication and Reproducibility Initiative and asked the field, through Highlighted Topic 66, for "tools that help researchers implement and document rigor practices."
NIH Simplified Peer Review Framework; OER Nexus, 28 Apr 2026; Highlighted Topic 66, posted 27 Apr 2026.EQIPD Quality System
A lean quality system for non-regulated preclinical research: 18 core requirements, free to use, from a consortium founded by 29 institutions in eight countries and now involving more than 100 organizations. For a study that informs a formal knowledge claim it requires the hypothesis, the analysis and the sample-size calculation to be "defined and documented before starting the experiments," and randomisation and blinding to be implemented, with exceptions justified.
Bespalov et al., Introduction to the EQIPD quality system, eLife 2021. Study Manager is not EQIPD-certified; this mapping is ours and we would welcome their review.| Requirement | NIH | EQIPD | Study Manager today | Status |
|---|---|---|---|---|
| Hypothesis and analysis fixed before the study | Rigorous experimental design; the Approach is scored. | Hypothesis "must be pre-specified"; analysis "defined and documented before starting the experiments." | The model family and the comparisons to be reported are chosen with the design. The engine refuses to size what hasn't been decided and says what deciding would cost. The same assumptions run the analysis. | Verified2026-09-15 |
| Sample size justified in advance | Named in reviewer guidance; an underpowered preliminary result is the margin applications fail on. | "Sample size calculation must be defined and documented before starting." | Power, effect size, covariate adjustment, unequal allocation, multiplicity, non-inferiority — computed through the analysis service and recorded with every assumption and its provenance. | Verified2026-09-16 |
| Randomisation | Named explicitly in the rigor criteria. | "Should be implemented, exceptions must be justified and documented." | Seeded stratified permuted-block allocation at enrollment on sex and the design's declared factors; litter and arrival weight checked regardless; sequence recorded and never deleted; re-allocation refused once data exists. | Verified2026-09-17 |
| Blinding / allocation concealment | Named explicitly in the rigor criteria. | Same rule as randomisation. | Blinded treatments in the record; a distinct unblinding authority not implied by admin; every Data Collection read surface redacts arm identity until an unblinding record exists; ARRIVE readiness checks for both. | Verified2026-09-17 |
| Exploratory or confirmatory, declared up front | Rigor of prior research; preliminary data assessed for what it can support. | Core requirement 10: "declare in advance whether a study is intended to inform a formal knowledge claim." | Desired evidence strength is set at planning — exploratory, planning, confirmatory, high-confidence — and a confirmatory study must pre-declare its minimum meaningful difference before it can proceed. | Available — shown as steps |
| Biological variables and hidden sources of variability | Sex, age, genetic background considered in design; Topic 66 asks for "hidden sources of variability." | Core requirement 15: risk assessment of "factors affecting the generation, processing and reporting of research data." | Sex is always a stratum and arms can be sex-specific; litter and arrival weight are checked whether declared or not; pilot import computes the baseline correlation; simulation-based reproducibility for longitudinal designs; after collection, a readout of whether the variance assumption held. | Available — shown as steps |
| Outcomes traceable to data; changes documented | Transparency of methods and analysis. | Core requirements 6 and 8: data records and changes documented; "reported research outcomes must be traceable to experimental data." | One record from design through analysis; observations carry an amendment chain with who, when and a controlled reason; the SEND transformation report traces every submitted value to its source; the analysis run stores a hash of the rows it read. | Verified2026-09-19 |
| All repetitions disclosed, whatever the outcome | Replication and Reproducibility Initiative: make null findings "available, reportable, and searchable." | Core requirement 9. | The portfolio holds every study, including completed and archived ones. Marking a study as a repetition of another and surfacing null results as such is planned. | Planned |
| Authentication of key resources; materials and equipment fit for use | Fourth rigor criterion. | Core requirements 13 and 14. | Species, strain and supplier are recorded with the animal; reagent and equipment authentication in the record is planned by customer demand. | Planned |
Rigor should not be reconstructed at publication.
ARRIVE. Readiness checks for randomisation, blinding and sample size at design; design and analysis summaries exported from the record. We say "supports ARRIVE-oriented readiness and reporting," never "ARRIVE compliant."
3Rs. Reduction is what the design step does — the covariate that halves the cohort, the control weighted √k : 1, the study sized once. Refinement and replacement are considerations in the record, not a scored module. We say "considerations," not "evidence," until a product behaviour exists.
SEND. A SENDIG 3.1.1 submission package — 19 domains, Define-XML 2.1, Dataset-JSON 1.1, an nSDRG and a transformation report tracing every value — built from the approved design and collected data, terminology from SEND CT and never guessed, validated against the FDA SEND rule set (Pinnacle 21 Community: 0 rejects, 0 errors). Not "FDA-accepted" until one has been.
Highlighted Topic 66, answered.
The statistical engine was developed under NIH SBIR Phase I and Phase II awards [year · numbers · IC]. Against NIDA's stated priorities — implement and document rigor practices; harmonize design and analysis; identify hidden variability; share protocols and common data elements — the first two are verified, the third is partial, the fourth is the shape of a Direct-to-Phase-II aim under PA-27-100 (receipt 5 January 2027).
Request a technical briefingA quality system needs somewhere to live.
EQIPD deliberately prescribes objectives, not solutions. Study Manager is one place those objectives can be met in the course of ordinary work rather than in a separate binder: the pre-specified hypothesis, the documented sample size, the recorded allocation and the traceable outcome are what the workflow produces. We are mapping the study record to the 18 core requirements. If your unit runs the EQIPD system — or is deciding to — we would like to compare notes and be told where the mapping is wrong.
Talk to us about the mapping