One SurveyJS form’s final score used A_total, B_total, and C_total even though none was defined. 1 affected file.
Artifact IDs: entidades_publicas_es
Retained review · July 2026
Four defect shapes across seven of 64 reviewed files.
FormPath reviewed a retained convenience set assembled from public-source SurveyJS definitions, REDCap data dictionaries, and FHIR Questionnaires—including eight hand-authored transcriptions of public hospital web forms. Seven files contained one of four supported logic-defect shapes within its stated evidence boundary.
This is a validation record, not a prevalence study. The files were not randomly sampled. Site copies, dated revisions, and three cross-format representations of the same clinical questionnaires remain separate artifacts because translation, revision, and site drift were part of what the review examined.
64artifacts reviewed
7files with a published supported defect
4distinct published defect shapes
3form ecosystems, with different evidence boundaries
The denominator
Every count resolves to a published manifest record.
Cohorts describe where files came from and why they were collected. They are not quality grades. “Affected” means the file contains one of the four published, supported defect shapes below; warnings and externally controlled behavior are not counted.
The public provenance manifest records all 64 stable artifact IDs, their paths and SHA-256 hashes inside the retained archive, result-record hashes, and defect-shape assignments.
| Review cohort | Artifacts | Affected files | Evidence boundary |
|---|---|---|---|
| SurveyJS public-repository definitions Definitions found in public repositories through a multilingual search, plus vendor examples and three clinical FHIR Questionnaires converted with their application’s own SurveyJS converter. | 27 | 2 | Native SurveyJS evidence |
| REDCap program and training dictionaries Publicly available programme, research, and training data dictionaries collected for adapter validation. | 6 | 0 | Translated REDCap logic |
| REDCap RADAR-CNS site dictionaries Eleven site-specific dictionaries from the retained RADAR-CNS multi-site collection. | 11 | 4 | Translated REDCap logic |
| REDCap eLwazi revision pair Two dated exports of one public catalogue project, retained as separate files for revision analysis. | 2 | 0 | Translated REDCap logic |
| FHIR project and maintained-example Questionnaires Questionnaires from public platform, standards-example, clinical-screening, and maintained project repositories. | 10 | 0 | Translated FHIR logic |
| FHIR hospital web-form transcriptions Eight public FHIR Questionnaires hand-authored from real hospital web forms in one retained repository. | 8 | 1 | Translated FHIR logic |
Artifact means a reviewed file, not an independent organization, deployment, or study. In particular, the retained set includes site copies, two dated revisions, and three FHIR Questionnaires also reviewed after conversion through their application’s SurveyJS converter.
What counted
Four concrete shapes—not seven different root causes.
One unavailable-choice-code problem appeared in four REDCap site files. The retained definitions establish the mismatch, but not when or why those values diverged. Counting files and counting defect shapes answer different questions, so the review reports both.
One SurveyJS visibility rule omitted the field-reference braces and also misspelled the expected value, leaving its question unreachable. 1 affected file.
Artifact IDs: mortgage_data
Four REDCap dictionaries required ethnicity code 17 while none of their supplied choice lists included that code. The same translated-model contradiction affected four files. 4 affected files.
Artifact IDs: radarmsdepvhirs1 · radarmsdisvhirs1 · radarepikcls1 · radarepiuklfrs1
One retained FHIR transcription used vision-20-40 in enableWhen while its source checklist offered vision. 1 affected file.
Artifact IDs: web_steadi_referral
Evidence boundaries
The same headline does not imply the same runtime evidence.
SurveyJS conclusions use SurveyJS as the native evaluator. REDCap and FHIR conclusions are about supported logic translated from the supplied definition into FormPath’s tested internal model.
| Ecosystem | Reviewed | Affected files | Defect shapes | How conclusions were checked |
|---|---|---|---|---|
| SurveyJS | 27 | 2 | 2 | Native SurveyJS evidence |
| REDCap | 19 | 4 | 1 | Translated REDCap logic |
| FHIR Questionnaire | 18 | 1 | 1 | Translated FHIR logic |
A translated proof is not a replay in a native REDCap or FHIR renderer. Unsupported functions, remote choices, and application-controlled behavior are separated for review rather than treated as false.
Inspect the ledger
The public manifest exposes provenance—not third-party forms.
Public definitions have different redistribution terms. The manifest therefore publishes stable identifiers, archive paths, source and result hashes, and finding assignments without shipping the definitions or analyzer results themselves. The totals on this page are recalculated directly from those 64 manifest records.
SurveyJS public-repository definitions · 27
Retained archive result path: results/formpath-results.json
Result SHA-256: cd93ebe7a444a4e9bdadec8240eb7e3b344667e3598a18d5598443c3265cf079
Artifact IDs: ansa_aod_clinical · asbi_nidaqs2usaudit · asbi_usaudit · asbi_whoaudit · covid19_survey · customer_satisfaction_tpl · entidades_publicas_es · event_feedback_tpl · finishing_projects_survey · fluvial_ined_research · fullstack_demo_survey · income_declaration_sample · leicester_cough · mgma_es · mortgage_data · mortgage_start · qtypes_showcase · rand36_health_survey · react_creator_sample · scored_survey_howto · sjs_health_wellness · sjs_legacy_angular_demo · sjs_market_research · sjs_pdf_demo · sjs_usability · survey_config_es · xuaun_gpsa_research
REDCap program and training dictionaries · 6
Retained archive result path: results/redcap-results.json
Result SHA-256: 1f7ea03097643db24867c4c3ee89a6a82b850a8c36a94cf8da875b6a93c074c9
Artifact IDs: aou_basics · aou_cope_202007 · aou_social_determinants_of_health · genomics_thailand · nuscript_clinical · vandy_trainingstudy
REDCap RADAR-CNS site dictionaries · 11
Retained archive result path: results/redcap-wild-results.json
Result SHA-256: 50e6461b1541891c874b8d8554cdc6e8ccb29d550505c30dda9f044aea0c5e64
Artifact IDs: radarepikcls1 · radarepiuklfrs1 · radarmddcibers1 · radarmddkcls1 · radarmddvumcs1 · radarmsdeposrs1 · radarmsdepregionhs1 · radarmsdepvhirs1 · radarmsdisosrs1 · radarmsdisregionhs1 · radarmsdisvhirs1
REDCap eLwazi revision pair · 2
Retained archive result path: results/redcap-wild-results.json
Result SHA-256: 50e6461b1541891c874b8d8554cdc6e8ccb29d550505c30dda9f044aea0c5e64
Artifact IDs: elwazi_2024_08 · elwazi_2024_10
FHIR project and maintained-example Questionnaires · 10
Retained archive result path: results/fhir-results.json
Result SHA-256: fd27acfff35d698fda32cc8338dcc430f92befb1b5b5cf31e995da2fdab9bf60
Artifact IDs: google_covid_screener · nlm_framingham · nlm_ussg_family_health · opensrp_covid19_assessment · opensrp_patient_registration · sdc_skip_logic_demo · who_measles_outbreak · xval_questionnaire_nidaqs2usaudit · xval_questionnaire_usaudit · xval_questionnaire_whoaudit
FHIR hospital web-form transcriptions · 8
Retained archive result path: results/fhir-results.json
Result SHA-256: fd27acfff35d698fda32cc8338dcc430f92befb1b5b5cf31e995da2fdab9bf60
Artifact IDs: web_audit · web_audit_c · web_bwh_aerd_allergy · web_cc_union_intake · web_mgh_derm_intake · web_mgh_hrscc · web_stanford_valleycare_gyn · web_steadi_referral
Retained archive: formpath-handoff (5).zip · SHA-256: 235bdeae6a0e82d38441b0acc523dd1904fde351951c7b9f1165ebb5e281b701
How to read this
Three cautions belong beside every total.
Does seven of 64 estimate a defect rate?
No. This was a convenience set assembled to exercise different form shapes and adapters, not a random sample. The result says what FormPath found in these retained files.
Why count site copies, revisions, and converted versions separately?
FormPath analyzes files. Site adaptation, revision drift, and translation can make two related representations behave differently, so each reviewed artifact remains visible in the denominator and none is presented as an independent study.
Does “not affected” mean a form was verified?
No. It means the retained review did not include that file among these four published defect shapes. Unsupported logic, external behavior, warnings, and other untested behavior remain outside that statement.
Check your own definition
Replace an aggregate with evidence about your form.
FormPath runs locally and labels native, translated, unsupported, and externally controlled conclusions separately.
Open the analyzer