Plan your health data model

Install this skill

npx skills add StanfordSpezi/SpeziVibe --skill health-data-model-planning

Or install all skills: npx skills add StanfordSpezi/SpeziVibe --skill '*'. See the Getting Started guide for tool-specific options.

Explore this skill in browser chat. Each button opens a new conversation; bring any earlier briefs with you and save the results yourself.

Try in ClaudeTry in ChatGPT

Need to paste the prompt yourself?

Browser prompt for health-data-model-planning

Please act as the SpeziVibe `health-data-model-planning` skill and walk me through it interactively.

STEP 1 — Fetch the skill instructions from this URL and read them carefully:
https://raw.githubusercontent.com/StanfordSpezi/SpeziVibe/main/skills/health-data-model-planning/SKILL.md

If the skill references other files or other skills by name, fetch them from the same repository:
- Files inside this skill (e.g. `references/foo.md`): https://raw.githubusercontent.com/StanfordSpezi/SpeziVibe/main/skills/health-data-model-planning/<relative-path>
- Other skills (e.g. `biodesign-needs-finding`): https://raw.githubusercontent.com/StanfordSpezi/SpeziVibe/main/skills/<skill-name>/SKILL.md

STEP 3 — Walk me through the skill interactively. When a step would normally save a markdown file to a project, show the file content in a code block instead so I can copy it. Don't simulate my answers — wait for me to respond.

If you cannot fetch the skill instructions or their reference files, ask me to paste them before proceeding.
Ask me to paste or attach any existing briefs you need. Do not claim to have read files, run commands, created a project, or tested an integration from this chat. Help me prepare the decisions and instructions, then hand off file and code work to my coding tool.

If your chat can’t open the skill link, open the instructions in a new tab and paste them into the chat.

Plans health data entities, relationships, and governance choices before committing to a storage layer. Biased toward FHIR for clinically meaningful or shareable data, but allows opt-out with justification.

What It Covers

• Core entities (person, observation, questionnaire, task, condition, medication, consent, care plan)
• Required and optional attributes
• Entity relationships and versioning needs
• Lifecycle states (draft, active, completed, cancelled, archived)
• FHIR resource fit assessment
• Terminology recommendations (LOINC, SNOMED CT, RxNorm, ICD-10)
• Library recommendations for the target stack (e.g., apple/FHIRModels for Swift)
• Governance and data-quality notes

Output

Produces docs/planning/data-model-brief.md.

Example

AI:   What does your app track or coordinate? And who creates,
      reads, and updates that data?

You:  Medication schedules, adherence logs, and vital signs.
      Patients log adherence, nurses review it, doctors
      adjust prescriptions.

AI:   For each of those, here's a starting data model:

      Medication Schedule
      — Required: medication name, dose, frequency, start date
      — Optional: end date, refill reminders, prescriber
      — FHIR fit: MedicationRequest
      — Lifecycle: draft → active → completed | cancelled

      Adherence Log
      — Required: medication ref, timestamp, taken (yes/no)
      — Optional: reason skipped, notes
      — FHIR fit: MedicationAdministration
      — Lifecycle: recorded (immutable)

      Should adherence data be shareable with external EHRs,
      or is it self-contained within the app?

Limitations

• Does not assume a specific database or framework
• Does not implement FHIR models — recommends existing libraries
• Requires clinical input to validate terminology choices
• Does not address implementation details like database indexes or API pagination