HIER × Robins Lab — Data Submission Templates
==============================================

One CSV per data domain. Each file has a header row plus one example row
(prefixed EXAMPLE_) so collaborators can see units and formats. Delete the
example row before returning the file. All identifiers should be replaced
with study-specific de-identified IDs (subject_id, scan_id, plan_id).

Master data dictionary:
- 00_data_dictionary.csv        Variable-level dictionary for every field in
                                templates 01–10. Columns: variable_name,
                                domain, form, type, units, allowed_values_or_range,
                                required, phi, derivation, source_system,
                                validation_rule, notes.

Data submission templates:
- 01_imaging.csv                Planning CT + adjunct imaging (per scan)
- 02_clinical_demographics.csv  Demographics and CV risk factors (per subject)
- 03_oncologic_treatment.csv    RT and systemic therapy exposures (per plan)
- 04_outcomes_longitudinal.csv  MACE and longitudinal cardiovascular events
- 05_biomarkers_labs.csv        Lipids, glycemia, hs-CRP, NT-proBNP, troponin
- 06_functional_pro.csv         ECG, echo, 6MWT, PROMIS-based PROs
- 07_physics_qa.csv             Scanner, kernel, reconstruction, phantom QA
- 08_medications.csv            Concomitant medications with RxNorm codes
                                and cumulative anthracycline exposure
- 09_adjudication.csv           Endpoints-committee adjudication log
                                (two-reviewer + tiebreaker)
- 10_biospecimen.csv            Biospecimen collection, processing, and
                                aliquot inventory (optional, if biobanked)

Conventions
-----------
- Encoding: UTF-8, comma-delimited. Line ending: LF preferred.
- Dates: ISO-8601 (YYYY-MM-DD). Timestamps: ISO-8601 with timezone.
- Missing values: leave the cell blank (do not use NA, NULL, -999, or 9999).
- Booleans: use "yes"/"no" strings, not 1/0 or true/false.
- Enumerations: lowercase snake_case; match the allowed_values column in
  00_data_dictionary.csv exactly.
- Identifiers: pseudonymous, unique across sites, stable within a subject.
  Do NOT include MRNs, names, DOBs, or accession numbers.
- PHI: no free-text PHI in notes columns. Date-shift dates per site policy;
  keep the shift constant per subject so intervals are preserved.
- DICOM UIDs may be included; they are not, on their own, PHI under HIPAA
  Safe Harbor when scan dates are shifted.

Versioning
----------
Templates are versioned v4.x. Header rows are stable within a major version.
Additions are appended as new trailing columns to preserve backward
compatibility. Breaking changes bump the major version and are announced
via the study coordinator.

Submission
----------
Return files via the SFTP endpoint provisioned per site under the executed
DUA. File name convention: {site}_{template}_{yyyymmdd}.csv, e.g.
DARTMOUTH_01_imaging_20260401.csv. A submission manifest (JSON) accompanies
each batch and lists file hashes (SHA-256), row counts, and the subject_id
range covered.

Contact: collaboration@radonccv.com