Institutional Research
Arcametric separates site-local identity management from research export data. Standard research exports use synthetic identifiers, structured clinical fields, coded assessments, and governed event records. Direct identifiers and site-local identity maps remain outside the standard research export boundary.
| Data Class | Standard Research Export | Boundary Rule |
|---|---|---|
| Patient name | No | Remains in the clinic's local identity system |
| Date of birth | No | Age bands or age at session may be used when approved |
| Address, phone, email, MRN | No | Excluded from standard research exports |
| Site-local subject map | No | Controlled by the clinic or site of care |
| Synthetic subject ID | Yes | Stable within the approved dataset |
| Session ID | Yes | Pseudonymous encounter-level key |
| Coded outcomes and safety data | Yes | Structured fields only, scoped to the approved export |
| Narrative clinical notes | No | Not part of the standard research export model |
Research exports are organized around a small set of stable grains. This lets principal investigators separate subject-level baselines, session-level interventions, event-level safety records, assessment results, and follow-up windows without receiving site-local identity records.
Arcametric uses standardized clinical vocabularies at the point of data entry where applicable, to support consistent exports for analysis in tools such as SPSS, R, Python, and SAS. This does not imply ICH/GCP “trial-grade” completeness for every field; validate mappings for each protocol.
| Mapping Rule | Use | Description |
|---|---|---|
| Exact external code | Preferred | The external code fully represents the exported concept |
| External code + local qualifier | Conditional | A standard code is used with an Arcametric qualifier or context field |
| Validated instrument | Assessment | A recognized instrument is exported with score and timepoint metadata |
| Local concept | Scoped | No exact external mapping is asserted in the export dictionary |
Each exportable variable should be documented with both machine-readable and reviewer-readable metadata. This makes the dataset easier to evaluate for IRB review, data-use agreements, statistical analysis plans, and reproducible cohort definitions.
| Dictionary Field | Type | Description |
|---|---|---|
| variable_name | Text | Machine-readable export header |
| display_label | Text | Human-readable label for reviewers |
| grain | Categorical | Site, subject, session, dose event, safety event, assessment, or follow-up |
| value_type | Categorical | Integer, numeric, categorical, timestamp, boolean, or identifier |
| unit | Text | UCUM or local unit label, when applicable |
| timepoint | Categorical | Baseline, dosing, recovery, follow-up day, or longitudinal interval |
| standard_or_vocabulary | Text | RxNorm, MedDRA, CTCAE, LOINC, SNOMED, UCUM, validated instrument, or local |
| source | Categorical | Form, device, practitioner entry, patient report, or calculated field |
| export_rule | Categorical | Included, excluded, conditional, or protocol-specific |
| notes | Text | Caveats, validation requirements, or interpretation limits |
Datasets are delivered as flat, machine-readable files. The schema below represents a standard longitudinal outcome trajectory export for observational interventional mental health research. It does not, by itself, imply causal efficacy, trial-grade completeness, or protocol-level validation.
| Field Label | Type | Description |
|---|---|---|
| subject_id | De-identified reference | Synthetic subject identifier within the approved dataset |
| site_id | De-identified reference | Pseudonymous site or practice identifier, when included |
| session_id | De-identified reference | Unique identifier for the treatment encounter |
| substance_rxnorm_code | Integer | Standardized RxNorm code for the compound, when applicable |
| route_code | Categorical | Route of administration code or controlled value |
| cumulative_dose_mg | Numeric | Total milligrams administered during the encounter |
| baseline_phq9_total | Integer | Depression severity score at baseline |
| followup_day | Integer | Interval of the follow-up assessment |
| followup_phq9_total | Integer | Depression severity score at follow-up interval |
| ae_meddra_code | Integer | Standardized adverse event code, when present |
| ae_ctcae_grade | Integer | 1-5 severity grade per CTCAE v5.0, when present |
| collection_window_flag | Categorical | On-window, early, late, unknown, or not applicable |
| instrument_completion_pct | Numeric | Percent of required instrument items completed |
| export_inclusion_flag | Categorical | Included, excluded, conditional, or pending protocol review |
Research exports should identify whether required fields were collected, skipped, unavailable, outside the protocol scope, or excluded by governance rules. These metadata fields help analysts avoid treating absent values as negative findings.
| Field | Type | Use |
|---|---|---|
| collection_status | Categorical | Complete, partial, unavailable, skipped, or not applicable |
| source_status | Categorical | Clinician-entered, patient-reported, device-derived, calculated, or imported |
| timepoint_status | Categorical | On-window, early, late, unknown, or not applicable |
| instrument_completion_pct | Numeric | Percent of required instrument items completed |
| export_inclusion_flag | Categorical | Included, excluded, conditional, or protocol-specific |
| site_validation_status | Categorical | Validated, unvalidated, site-supplied, or not applicable |
| Research Question | Required Domains | Boundary Note |
|---|---|---|
| How do PHQ-9 and GAD-7 scores change across follow-up windows? | Baselines, assessments, follow-up | Observational unless paired with a protocol-specific causal design |
| Are adverse events associated with route, dose, or baseline risk category? | Dosing, safety events, baselines | Safety event coding and export scope must be protocol-approved |
| What follow-up completion patterns predict durable response? | Follow-up windows, outcomes, collection metadata | Absent values must be interpreted through collection status fields |
| How do session-level safety events relate to post-session deterioration signals? | Safety surveillance, follow-up assessments | Requires agreed definitions for deterioration and observation windows |
| Which variables are suitable for cross-site aggregation? | Synthetic identifiers, coded outcomes, coded safety events | Direct identifiers and site-local identity maps remain excluded |
Part 10 describes how Arcametric preserves the meaning of governed clinical concepts, data elements, instruments, and external mappings so a research team can determine what each field represents, how a value was produced, and whether similar values represent the same clinical fact or distinct observations.
CROSSWALK is Arcametric's semantic mapping method for relating external research frameworks, consensus sets, registries, terminology systems, reporting requirements, and other structured sources to governed Arcametric concepts. Each external source retains its own identity, version, requirements, and provenance; it does not define the canonical Arcametric concept. The taxonomy below describes intended meaning and relationships. The presence of a field or code does not establish production implementation, interoperability, clinical validity, or regulatory acceptance.
The semantic model follows six rules. Arcametric uses authoritative external terminology where an exact standard exists. Where a shared external code would not preserve the intended meaning, Arcametric uses a governed local concept and records the mapping relationship rather than asserting equivalence. Experience and safety remain separate facts because a difficult or intense experience is not necessarily an adverse event. Occurrence, provenance, appraisal, measurement, and safety classification remain conceptually distinct. Stable concept identity remains separate from display labels and external mappings. Patient-facing surfaces use plain labels, not internal codes or database identifiers.
For this dictionary, the relevant data elements and concept families are grouped into six layers. Each layer answers a different question about the treatment record.
The taxonomy holds several distinctions apart by design. Subjective safety is not an adverse event: a patient can feel unsafe without a clinician-classified adverse event, and an adverse event can exist independently of the patient's global safety appraisal. Personal meaningfulness is not MEQ-30: meaningfulness is a standalone appraisal, and MEQ-30 is a validated instrument with separate provenance. The global quality-of-life rating is not WHOQOL: the Arcametric global rating is a separate governed concept, and existing WHOQOL data retains its own meaning. Intentions are not mindset: intentions are stated aims, and mindset is the patient's psychological orientation entering treatment. Treatment expectancy is a governed preparation-context measure, and Arcametric defines its numeric representation. Preparation exposure is a process measure of time spent preparing, not a required therapeutic cadence.
The table below records the meaning of each field. Each meaning quotes Arcametric's approved field definition.
| Field | Layer | Approved meaning |
|---|---|---|
| Touch boundaries discussed | L1 | "Confirms that touch preferences and boundaries were discussed before the administration session, including what forms of supportive touch are acceptable or off-limits." |
| Medication interactions reviewed or discussed | L1 | "Confirms that possible interactions between the planned treatment and the patient's current medications were reviewed with the patient. This records the discussion, not a medication order or interaction finding." |
| Altered-state consent restriction explained | L1 | "Confirms that the patient was told that permissions and boundaries established before the altered state remain the reference point during the session, rather than introducing new permissions while acutely altered." |
| Intentions | L2 | "Captures the patient's stated reasons or aims for treatment using structured intention themes. It records preparation context, not a predicted outcome." |
| Mindset | L2 | "Captures the patient's current psychological orientation entering treatment as part of set and setting. It supplements, rather than replaces, the patient's stated intentions." |
| Treatment expectancy | L2 | Captures the patient's treatment expectancy before the session on Arcametric's canonical zero-to-ten scale. |
| Preparation exposure | L2 | "Total time spent in preparation before the treatment session, recorded in minutes. This measures preparation exposure without requiring a specific number of visits." |
| Harm versus benefit | L3 | "Captures the patient's overall appraisal of whether the experience was more harmful or more beneficial." |
| Goals met | L3 | "Captures the patient's appraisal of how fully the experience met the goals they had going into treatment." |
| Subjective safety | L3 | "Captures the patient's overall sense of how safe they felt during the experience. This is separate from clinician-recorded adverse events." |
| Personal meaningfulness | L3 | "Captures how personally meaningful the patient considers the experience. This is a standalone outcome and is not an MEQ-30 score." |
| Well-being | L3 | "Captures the patient's overall appraisal of well-being following the experience." |
| Global quality of life | L3 | "Captures the patient's overall quality of life" on a zero-to-one-hundred scale. "This is a separate global rating and is not the existing WHOQOL instrument score." |
Arcametric applies external terminology by domain, where the standard fits the fact. Drug identity uses RxNorm where appropriate. Units and measures use UCUM, the Unified Code for Units of Measure. Diagnosis classification uses the ICD family according to use case and jurisdiction, and clinical terminology may also use SNOMED CT. Assessment observations use LOINC where the exact instrument and form identity is verified. Safety and adverse events use MedDRA and CTCAE where licensing permits. HL7 FHIR, a healthcare data-exchange specification, is transport infrastructure, not terminology.
For experiential treatment phenomena, the reviewed work identified no broadly adopted shared terminology that adequately represents the relevant non-pathologizing concepts. Where that holds, Arcametric governs a local concept and maintains the mapping relationship transparently rather than forcing the concept into a safety dictionary.
The limits of this section are explicit. Arcametric's experiential vocabulary is a governed local vocabulary, not an official industry standard. Arcametric makes no FHIR conformance claim. It does not publish instrument identifiers for forms whose exact implementation has not been separately verified and approved. It does not treat MedDRA as the primary identity for a therapeutic experience. The presence of a code, here or elsewhere, does not by itself establish implementation, interoperability, clinical validity, or regulatory acceptance.
Source evidence and provenance are maintained for each external mapping. Depending on the concept and source, supporting evidence may include peer-reviewed consensus work, validated instrument documentation, health authority or regulatory materials, terminology standards, and other domain-specific literature. External sources inform mapping and provenance; Arcametric governs its own canonical concepts, data elements, applicability, value domains, and product representation.
This appendix preserves concepts from the source taxonomy that sit outside the evergreen core export dictionary. Inclusion here does not represent a standard research export, clinical claim, or protocol-ready dataset without separate validation.
Send questions about the data model or the export formats to hello@arcametric.com.