Standardized Vocabularies, Code Sets & Guideline Sources

Comprehensive map of clinical vocabularies, ontologies, structured guideline content, and decision-support datasets relevant to Starlight Practice (pediatric DPC EMR + parent app + AI substrate).

Verification note. WebFetch and WebSearch were unavailable in the session that produced this document. License terms and pricing below reflect the author's best knowledge as of early 2026 and must be re-verified before any binding commercial use — especially CPT, DSM, SNOMED CT International, and any "research-only" dataset. Items where I have moderate or low confidence are flagged VERIFY.

Reading Guide

• Priority — must-have (ship-blocking interoperability), nice-to-have (improves UX/CDS), wait-until-needed (defer until a feature requires it).
• Cost — free, free-with-license (no fee but click-through or attribution), academic-only, paid (commercial fee required).
• License risk — flagged inline where the wrong assumption could create legal exposure.

---

1. Clinical Vocabularies & Code Sets

1.1 ICD-10-CM — Diagnosis Codes (US Clinical Modification)

• Description. US clinical-modification diagnosis codes maintained by CDC/NCHS and CMS. Required on every claim and most clinical encounters.
• URL. https://www.cms.gov/medicare/coding-billing/icd-10-codes and https://www.cdc.gov/nchs/icd/icd-10-cm/
• Access. Direct download (XML, tabular text, PDF, Excel addendum); also bundled in UMLS.
• Format. Tabular flat files; XML; CSV; an official ClaML XML release.
• Cost. Free. US Government work, no royalties.
• Update cadence. Annual full release (October 1, fiscal-year aligned); April 1 mid-year addendum since 2022.
• License. Public domain (US Govt work). Safe to embed.
• Pediatric specificity. Full pediatric coverage including perinatal (P00–P96), congenital (Q00–Q99), and child-specific Z-codes.
• Priority. must-have.

1.2 ICD-10-PCS — Inpatient Procedure Codes

• Description. CMS-maintained inpatient procedure code system (7-character alphanumeric).
• URL. https://www.cms.gov/medicare/coding-billing/icd-10-codes
• Access. Direct download.
• Format. Tabular flat files, XML.
• Cost. Free (US Govt).
• Update cadence. Annual (Oct 1).
• License. Public domain.
• Pediatric relevance. Low for outpatient DPC; only relevant if Starlight ever handles inpatient billing or hospitalist workflows.
• Priority. wait-until-needed.

1.3 CPT — Current Procedural Terminology   ⚠️ PAID / RESTRICTIVE

• Description. AMA-owned procedure codes used for outpatient billing. The most commercially sensitive code set we will touch.
• URL. https://www.ama-assn.org/practice-management/cpt
• Access. AMA Intelligent Platform (formerly CPT API), AMA bookstore, third-party redistributors (Find-A-Code, Codify by AAPC).
• Format. CSV/Excel data files, JSON via API, PDF code book.
• Cost. Paid — significant. Requires a CPT Distribution License from AMA for any product that displays or transmits CPT descriptors. Pricing tiers vary by:
  • Product type (EHR, billing system, claims clearinghouse, mobile app).
  • Number of end users / seats.
  • Whether full descriptors, short descriptors, or only codes are displayed.
  • Typical commercial EHR licensing has historically run mid-four to low-six figures USD per year, scaling with user count. VERIFY current pricing directly with AMA Intelligent Platform sales — the AMA restructured licensing in 2024–2025.
• Update cadence. Annual (Jan 1) plus quarterly Category III addenda.
• License. Proprietary AMA copyright. Cannot redistribute. Cannot store full descriptors in a public repo. Cache must be access-controlled per the license. Audit clauses are real.
• Pediatric specificity. Includes well-child preventive visit codes (99381–99395), developmental screening (96110), behavioral/emotional screening (96127), vaccine administration (90460–90474), and counseling codes — all critical for DPC pediatric billing if/when we bill insurance for ancillary services.
• Priority. must-have if we bill any third-party payors; wait-until-needed if pure-cash DPC. Many DPC practices avoid CPT in patient-facing displays specifically to dodge the license. We may be able to use HCPCS-only public subsets and short ICD-10-CM codes for cash-pay encounters and only license CPT when we add a billing module.
• Note. Some workarounds: (a) use only the numeric CPT codes without descriptors in internal data and rely on the AMA-licensed billing service's output; (b) require each clinic to bring their own CPT license. VERIFY with healthcare counsel.

1.4 SNOMED CT — Clinical Concepts

• Description. ~360k+ active concept SNOMED International ontology covering findings, disorders, procedures, body structures, organisms, etc. The most expressive clinical vocabulary in use.
• URL. https://www.nlm.nih.gov/healthit/snomedct/ (US Edition); https://www.snomed.org/ (International).
• Access. US Edition: download from NLM after free UMLS Metathesaurus License. International edition: requires SNOMED International member-country affiliate license.
• Format. RF2 (Release Format 2) — pipe-delimited TSV with reference sets; also FHIR terminology server distributions (e.g., Snowstorm, OntoServer).
• Cost.
  • US Edition (CORE Subset and full): free for any use within the United States under the US National Library of Medicine's no-cost license — including commercial use, including SaaS hosted in the US. VERIFY the latest UMLS license terms; NLM has occasionally tightened the redistribution clause.
  • International / non-US: paid affiliate license via SNOMED International. Member countries (e.g., UK, Canada, Australia) have national distributions. If Starlight ever serves Canadian or Mexican families, we need the affiliate license for those jurisdictions.
• Update cadence. US Edition: twice yearly (March, September). International: twice yearly (Jan, July).
• License. SNOMED Affiliate License (free in US via UMLS). Forbids resale of the terminology itself; embedded use in software is allowed.
• Pediatric specificity. Excellent — full pediatric findings, developmental milestones, neonatal disorders. Pediatric extension content is integrated.
• Priority. must-have for clinical CDS, problem-list normalization, and AI grounding.

1.5 LOINC — Lab Tests, Observations, Document Types

• Description. Logical Observation Identifiers Names and Codes — Regenstrief-maintained vocabulary for lab tests, vital signs, clinical documents, survey instruments (PHQ-9, M-CHAT-R, etc.), and panels.
• URL. https://loinc.org/
• Access. Free download after creating a Regenstrief account.
• Format. CSV, Access DB, FHIR CodeSystem, OWL.
• Cost. Free, including commercial use.
• Update cadence. Twice yearly (June, December).
• License. LOINC License — permissive, similar to a name-attribution license. Do not modify codes/short names; attribution required. Safe to embed.
• Pediatric specificity. Excellent — includes pediatric vitals (head circumference, BMI percentile by age), screening instruments (PHQ-9-modified for adolescents, M-CHAT-R, ASQ-3 panels), newborn screening panels.
• Priority. must-have.

1.6 RxNorm — Medication Concepts

• Description. NLM-maintained normalized naming for clinical drugs, ingredients, brand names, dose forms.
• URL. https://www.nlm.nih.gov/research/umls/rxnorm/
• Access. Free monthly download (RRF format) and free REST API (https://rxnav.nlm.nih.gov); also bundled in UMLS.
• Format. Rich Release Format (RRF) flat files; JSON via RxNav REST.
• Cost. Free.
• Update cadence. Weekly (Wed) for the API; monthly full file release.
• License. No-cost UMLS license. Public domain core where applicable; some source vocabularies inside UMLS have restrictions (see UMLS source-restriction matrix).
• Pediatric specificity. General drug coverage; pediatric dosing not encoded here — pair with Lexicomp Pediatric or Harriet Lane (paid, see §3).
• Priority. must-have for medication list, e-Rx, drug-interaction lookups.

1.7 NDC — National Drug Code (FDA)

• Description. Three-segment product identifier for every drug marketed in the US.
• URL. https://www.fda.gov/drugs/drug-approvals-and-databases/national-drug-code-directory
• Access. Free download (Excel/text); also queryable via openFDA API.
• Format. CSV/Excel, JSON via openFDA.
• Cost. Free.
• Update cadence. Daily (FDA NDC Directory); openFDA is delayed 1–2 days.
• License. Public domain.
• Priority. must-have for prescription accuracy, especially for pharmacy interoperability via NCPDP SCRIPT.

1.8 UCUM — Units of Measure

• Description. Unified Code for Units of Measure (Regenstrief). Canonical machine-readable units (e.g., mg/dL, [degF], kg/m2).
• URL. https://ucum.org/
• Access. Free download (XML).
• Format. Single XML file (ucum-essence.xml) plus implementation libraries (ucum-lhc for JS).
• Cost. Free.
• Update cadence. Infrequent (years between revisions).
• License. Permissive Regenstrief license; commercial use allowed with attribution.
• Priority. must-have — required by FHIR for any quantity value.

1.9 VSAC — Value Set Authority Center

• Description. NLM's authoritative repository of clinical value sets used in CMS quality measures, USCDI, and CDC programs.
• URL. https://vsac.nlm.nih.gov/
• Access. Free with UMLS license; FHIR Terminology Service API and SVS SOAP API.
• Format. FHIR ValueSet/CodeSystem JSON; SVS XML; spreadsheet exports.
• Cost. Free.
• Update cadence. Continuous — value sets update on author timelines.
• License. Free under UMLS terms; component vocabularies (CPT inside a value set) are still subject to their own licenses — CPT-containing value sets cannot be redistributed without a CPT license. Important caveat.
• Pediatric specificity. Includes Bright-Futures-aligned screening value sets (developmental screening, depression screening, BMI), CMS pediatric quality measure value sets.
• Priority. must-have for grounding our screening/CDS logic in published, traceable value sets.

1.10 HL7 FHIR R4 (and R4B/R5)

• Description. Healthcare data interchange standard. R4 is the regulatory baseline (ONC Cures Act); R4B adds subscriptions and minor fixes; R5 is current but not yet mandated.
• URL. https://hl7.org/fhir/R4/ and https://www.hl7.org/fhir/
• Access. Free download; reference servers (HAPI FHIR open source, public test servers at tx.fhir.org, hapi.fhir.org).
• Format. JSON, XML, Turtle; StructureDefinition profiles.
• Cost. Free.
• Update cadence. Specification: major every ~2 years; profiles and ImplementationGuides ongoing.
• License. CC0 for the specification itself; HL7 trademarks for "FHIR".
• Pediatric specificity. US Core profiles cover most pediatric needs; AAP has been involved in pediatric IGs. See also: HL7 Pediatric Care WG, IPA (International Patient Access) IG.
• Priority. must-have — our wire format.

1.11 USCDI — US Core Data for Interoperability

• Description. ONC-published minimum data set every certified EHR must support. Currently USCDI v4 (2023 final), USCDI v5 in progress (VERIFY).
• URL. https://www.healthit.gov/isa/united-states-core-data-interoperability-uscdi
• Access. Free download (PDF, JSON, machine-readable manifest).
• Format. PDF spec; JSON manifest; mapped to FHIR US Core IG.
• Cost. Free.
• Update cadence. Annual (USCDI vN+1 published each year).
• License. Public domain.
• Pediatric note. USCDI does not yet have rich pediatric-specific data classes (growth percentiles, immunization forecast). Pediatric gaps are covered by the AAP-influenced pediatric supplement and USCDI+ Pediatric (in progress as of early 2026, VERIFY).
• Priority. must-have.

1.12 CVX & MVX — Vaccine Codes (CDC)

• Description. CVX (vaccine product) and MVX (manufacturer) codes used in HL7 immunization messaging and FHIR Immunization resources.
• URL. https://www2a.cdc.gov/vaccines/iis/iisstandards/vaccines.asp?rpt=cvx
• Access. Free download (HTML tables, downloadable text/Excel) and PHIN VADS.
• Format. Text, Excel, FHIR CodeSystem.
• Cost. Free.
• Update cadence. As needed when new vaccines are licensed (multiple updates per year).
• License. Public domain (US Govt).
• Pediatric specificity. Critical — pediatric immunization documentation depends on CVX/MVX.
• Priority. must-have.

1.13 HCPCS — Healthcare Common Procedure Coding System

• Description. CMS-maintained billing codes. Level I = CPT (paid, see §1.3). Level II = J-codes for drugs/DME/supplies and services not in CPT (free, public domain).
• URL. https://www.cms.gov/medicare/coding-billing/healthcare-common-procedure-system
• Access. Free download for Level II.
• Format. Excel, text.
• Cost. Level II free; Level I (CPT) paid.
• Update cadence. Annual + quarterly updates for Level II.
• License. Level II is public domain.
• Pediatric specificity. Vaccine product codes (90xxx are CPT, but some immunoglobulin J-codes are HCPCS), pediatric DME (apnea monitors, nebulizers, hearing aids).
• Priority. must-have for billing module; nice-to-have otherwise.

1.14 DSM-5-TR — Mental Health Diagnoses   ⚠️ PAID

• Description. APA's Diagnostic and Statistical Manual, Text Revision (2022).
• URL. https://www.psychiatry.org/psychiatrists/practice/dsm
• Access. Print/eBook purchase, DSM-5-TR online subscription, APA licensing for software embedding.
• Format. Book, PDF, structured XML for licensees.
• Cost. Paid. Book ~$220 print; institutional/software embedding licenses negotiated case-by-case (typically four to five figures USD/yr depending on scope). VERIFY.
• Update cadence. Major revisions every ~10 years; text revisions every few years.
• License. APA proprietary copyright. Do not embed DSM criteria text in our app without a license. Diagnosis names alone (e.g., "ADHD, predominantly inattentive presentation") are descriptive but the structured criteria sets are copyrighted.
• Pediatric relevance. ADHD, autism spectrum disorder, anxiety, depression, conduct disorders — central to pediatric behavioral health.
• Workaround. ICD-10-CM F-codes are the billable equivalents and are public domain. Cross-walks DSM-5-TR ↔ ICD-10-CM are published by APA. We can use ICD-10-CM codes plus AAP/Bright Futures structured screening (PHQ-9, SCARED, Vanderbilt) without DSM text.
• Priority. wait-until-needed — start with ICD-10-CM F-codes + screening tools; revisit if we add a behavioral-health module.

---

2. Pediatric-Specific Structured Content

2.1 Bright Futures (AAP / HRSA)

• Description. AAP's national health-promotion and prevention guidelines for infants, children, and adolescents — defines well-child visit content from prenatal through age 21.
• URL. https://brightfutures.aap.org/ and https://www.aap.org/en/practice-management/bright-futures/
• Access.
  • Periodicity Schedule (one-page chart of recommended screenings/services by age): free PDF.
  • Bright Futures Guidelines, 4th Edition (full clinical guidance text): free PDF download from AAP since 2017 (HRSA-funded).
  • Bright Futures Tool and Resource Kit, 2nd Edition: paid AAP product (~$100–$300 retail; site-license pricing for EHR vendors negotiated). Contains the actual visit forms, parent handouts, and structured templates.
• Format. PDF for guidelines; structured forms in the paid Tool Kit (PDF + Word). No official FHIR or JSON release — community-built FHIR PlanDefinitions exist (e.g., NACHC, Logica) but are not authoritative. VERIFY any claim of an official machine-readable release.
• Cost. Guidelines free; Tool Kit paid; structured FHIR derivative work — none official as of early 2026 (VERIFY).
• Update cadence. Periodicity schedule updated annually-ish; full guidelines every several years.
• License. AAP copyright; the free PDFs allow clinical use but redistribution and derivative works require AAP permission. Do not republish the periodicity chart inside our product UI without checking the AAP's redistribution policy.
• Pediatric specificity. The single most important pediatric content source.
• Priority. must-have. Plan to build our own structured representation (visit-by-visit checklist) derived from the periodicity schedule, citing AAP. Engage AAP for a Tool Kit license once we're past MVP.

2.2 ACIP Immunization Schedules (CDC)

• Description. Advisory Committee on Immunization Practices recommendations — the authoritative US childhood and adolescent immunization schedules.
• URL. https://www.cdc.gov/vaccines/hcp/imz-schedules/ and https://www.cdc.gov/vaccines/acip/
• Access.
  • Human-readable: free PDF/HTML schedules (child 0–18, catch-up).
  • Machine-readable: CDC publishes the schedule as a FHIR ImmunizationRecommendation–compatible artifact and an XML/JSON release annually, plus the CDS Hooks-compatible CDS (Clinical Decision Support) for Immunization (CDSi) Logic Specification with reference data files (https://www.cdc.gov/vaccines/programs/iis/cdsi.html). This is gold.
  • CDC's mobile app (CDC Vaccine Schedules) and STC's open-source ICE (Immunization Calculation Engine) consume CDSi.
• Format. PDF, XML, JSON (ImmunizationRecommendation); CDSi: XML supporting data + spec PDF.
• Cost. Free.
• Update cadence. Annual (Feb) + interim updates as ACIP votes on new vaccines.
• License. Public domain.
• Pediatric specificity. Critical.
• Priority. must-have. Use CDSi reference engine output to drive forecasting in the EMR.

2.3 CDC Growth Charts + WHO Growth Standards

• Description. Percentile reference curves for length/height, weight, head circumference, BMI by age and sex.
• URL. CDC: https://www.cdc.gov/growthcharts/. WHO: https://www.who.int/tools/child-growth-standards.
• Access. Free download — LMS parameter tables (CSV) for both CDC (2-20 years) and WHO (0-2 years recommended by AAP).
• Format. CSV/Excel LMS tables; SAS macros.
• Cost. Free.
• Update cadence. Stable — CDC charts last updated 2000; WHO 2006/2007.
• License. Public domain (CDC) / WHO permissive use.
• Pediatric specificity. Core EMR feature.
• Priority. must-have. Use WHO 0–24 mo and CDC 2–20 yr per AAP recommendation.

2.4 USPSTF Recommendations

• Description. US Preventive Services Task Force evidence-graded preventive recommendations (Grade A/B/C/D/I).
• URL. https://www.uspreventiveservicestaskforce.org/ and the Prevention TaskForce app/API at https://epss.ahrq.gov/.
• Access. Free web; AHRQ's Prevention TaskForce API provides JSON output keyed by patient age/sex/risk factors.
• Format. Web pages, PDFs; AHRQ JSON API.
• Cost. Free.
• Update cadence. Continuous — recommendations updated as evidence reviews complete.
• License. Public domain.
• Pediatric specificity. Limited but important: depression screening (12+), behavioral counseling for obesity, vision screening, tobacco/sexual-health interventions for adolescents.
• Priority. nice-to-have — overlap with Bright Futures; useful for patient-facing "why this screening" explanations.

2.5 AAP Red Book — Pediatric Infectious Disease   ⚠️ PAID

• Description. Authoritative AAP reference on pediatric infectious diseases.
• URL. https://redbook.solutions.aap.org/
• Access. AAP Red Book Online subscription (per-clinician or institutional).
• Format. Web/HTML, PDF chapters; mobile app.
• Cost. Paid. Individual subscription ~$200/yr; institutional negotiated. Software-embedding API not publicly offered.
• License. AAP copyright. Cannot embed.
• Priority. wait-until-needed. Link out to AAP Red Book Online for clinicians who subscribe individually; do not embed text.

2.6 HEEADSSS & Other Adolescent Frameworks

• Description. HEEADSSS (Home, Education, Eating, Activities, Drugs, Sexuality, Suicide/Depression, Safety) is a public-domain interview framework.
• URL. Original — Goldenring & Rosen 2004; widely reproduced.
• Access. Free; framework itself is not copyrighted but specific published forms may be.
• Format. Various PDFs and clinic-specific implementations.
• Cost. Free for the framework structure.
• License. Framework concept is freely usable; AAP/SAHM versions of the form may have copyright on layout. Build our own template.
• Priority. nice-to-have — easy to implement as a structured template.

2.7 Validated Screening Tools

Tool	Domain	Cost / License	Notes
M-CHAT-R/F	Autism, 16-30 mo	Free for clinical use (mchatscreen.com); copyrighted by Robins, Fein, Barton — cannot modify.	LOINC: 73831-0. Required permission for translations.
ASQ-3   ⚠️	Developmental, 1-66 mo	Paid — Brookes Publishing (~$295 starter kit, per-administration licenses for digital).	Most-used Bright-Futures-recommended tool.
PEDS	Developmental	Paid — pedstest.com licensing.	Lower-cost ASQ alternative.
PHQ-9 / PHQ-A	Depression, adolescent	Public domain (Pfizer released).	LOINC 44249-1; widely embedded.
GAD-7	Anxiety	Public domain (Pfizer released).	LOINC 69737-5.
SCARED	Pediatric anxiety	Free for non-commercial; commercial use requires permission from authors.	VERIFY for our commercial use.
Vanderbilt ADHD	ADHD, parent & teacher	Free; AAP-distributed.	LOINC panels exist.
CRAFFT 2.1+N	Adolescent substance use	Free for clinical use; Boston Children's holds copyright.	Must include attribution.
Edinburgh Postnatal Depression Scale (EPDS)	Maternal depression at well-child	Free with attribution.	LOINC 99046-5.
PSC-17 / PSC-35	Psychosocial	Public domain (Mass General).
SWYC (Survey of Well-being of Young Children)	Combined developmental + behavioral + family	Free under SWYC license; cannot modify.	Built into many state Medicaid programs.

• Priority. must-have for the public-domain tools (PHQ-9, GAD-7, Vanderbilt, PSC, M-CHAT-R, SWYC). wait-until-needed for ASQ-3 (paid) — many DPC pediatricians use M-CHAT + Vanderbilt + SWYC and skip ASQ.

2.8 Care Frameworks (PRECEDE-PROCEED, etc.)

• Description. PRECEDE-PROCEED, Chronic Care Model, Patient-Centered Medical Home — conceptual frameworks, not code sets.
• License. Concept-level frameworks are not copyrightable; specific published diagrams/forms may be.
• Priority. nice-to-have as documentation conventions, not data sources.

---

3. Drug Interaction & Decision Support Content

3.1 DrugBank

• URL. https://go.drugbank.com/
• Access. Academic XML download (free with academic license); commercial API via DrugBank Online (paid).
• Cost. Academic free; commercial paid (five to six figures USD/yr depending on use; VERIFY with DrugBank sales).
• Format. XML, JSON, REST API.
• Update cadence. Continuous (commercial); periodic (academic file).
• License. CC BY-NC 4.0 for academic XML — explicitly non-commercial. A commercial product cannot legally use the academic file. This is a common mistake.
• Priority. wait-until-needed for drug-drug interactions; if budget allows, license commercially. Otherwise rely on RxNorm + FDA labels + paid pediatric-dosing reference (see 3.6).

3.2 Open Targets

• URL. https://www.opentargets.org/
• Access. Free download (Parquet, JSON), GraphQL API.
• Cost. Free.
• License. CC0 for Open Targets data; embedded sources have their own licenses (some restrictive).
• Priority. wait-until-needed — drug-disease/target associations are more research than DPC.

3.3 DDInter, DDI-Corpus, and similar interaction datasets

• Description. Academic drug-drug interaction datasets used to train ML models.
• License. Varies — most are CC BY-NC or research-only. VERIFY per dataset.
• Priority. wait-until-needed — useful only if we're training our own DDI model rather than licensing a commercial one.

3.4 SIDER — Side Effect Resource

• URL. http://sideeffects.embl.de/
• Cost. Free for academic use; data extracted from FDA labels.
• License. CC BY-NC-SA — non-commercial. Cannot embed in a commercial product.
• Priority. wait-until-needed. Use openFDA labels directly (public domain) instead.

3.5 FDA Black-Box Warnings, REMS

• URL. https://www.fda.gov/drugs/drug-safety-and-availability/risk-evaluation-and-mitigation-strategies-rems and openFDA https://open.fda.gov/.
• Access. openFDA API; DailyMed for full structured product labels (SPL).
• Format. SPL XML; openFDA JSON.
• Cost. Free, public domain.
• Priority. must-have for any e-prescribing — pull boxed warnings from SPL, surface in CDS.

3.6 Pediatric Dosing References   ⚠️ PAID

Source	Format	Cost	Notes
Lexicomp Pediatric & Neonatal Dosage Handbook (Wolters Kluwer)	Web, app, API	Paid. Institutional licensing; embedded API in five to six figures USD/yr. VERIFY.	Industry standard for pediatric dosing.
Harriet Lane Handbook (Elsevier)	Book, app	Paid. Per-clinician subscription.	Beloved bedside reference; no commercial-embed API.
Micromedex NeoFax / Pediatrics (Merative)	Web/API	Paid.	Strong neonatal coverage.
Epocrates Pediatric	App	Freemium	Clinician-only; not embeddable.

• Priority. wait-until-needed but must-have before launch of any active prescribing module. This is the #1 patient-safety risk in pediatrics — wrong-dose-by-weight kills kids. Do not roll our own dosing logic from public sources alone.

---

4. Guideline Aggregators

Source	Description	Cost	Embed-able?	Priority
Trip Database (tripdatabase.com)	Evidence-based search aggregator	Free tier + Trip Pro (~$60/yr individual)	Search-only; no bulk embed	nice-to-have (link out)
ECRI Guidelines Trust (guidelines.ecri.org)	Replaced AHRQ National Guideline Clearinghouse	Free with registration; commercial use requires ECRI license. VERIFY.	Limited	nice-to-have
PubMed Clinical Queries	NLM-curated PubMed filters for clinical questions	Free (NLM E-utilities API)	Yes	must-have for AI-grounded answers
DynaMed (EBSCO)	POC reference	Paid — institutional/per-clinician	No commercial embed	wait-until-needed
UpToDate (Wolters Kluwer)	POC reference	Paid — ~$559/yr individual; enterprise $ negotiated	UpToDate API exists, paid & gated	wait-until-needed
ClinicalKey (Elsevier)	POC reference + journals	Paid	No commercial embed	wait-until-needed
Cochrane Library	Systematic reviews	Mixed — some free (CDSR abstracts always free; full text free in many countries via Cochrane National Provisions, paid in US)	Subscription required for full-text US embed	nice-to-have (abstracts only)
BMJ Best Practice	POC reference	Paid	No commercial embed	wait-until-needed
AAFP / AAP / AHA / etc. society guidelines	Free PDFs from societies	Free (read); redistribution varies	Link out, do not republish	nice-to-have
GIN (Guidelines International Network)	International registry	Free directory	Link out	nice-to-have

---

5. Imaging & Waveform Datasets (Academic / Public)

Source	Description	Cost	License	Pediatric Relevance
MIMIC-IV (physionet.org/content/mimiciv)	BIDMC adult ICU EHR	Free w/ credentialing	PhysioNet Credentialed Health Data License (DUA) — research-only, no commercial use, no patient re-identification	Adult only; pattern-reference for our own data architecture
MIMIC-IV-ECG	12-lead ECG waveforms paired with MIMIC-IV	Free credentialed	Same DUA; research-only	Adult
MIMIC-CXR	Chest X-rays + radiology reports	Free credentialed	Same DUA; research-only	Adult, but useful priors
eICU-CRD	Multi-center adult ICU	Free credentialed	Research-only DUA	Adult
PhysioNet broadly	Many waveform/EHR datasets	Free credentialed	Per-dataset; most research-only	Some pediatric (e.g., neonatal HRV)
NIH ChestX-ray14 / ChestX-ray8	100k+ frontal CXRs	Free	CC0/permissive	Includes pediatric subset
CheXpert (Stanford)	224k CXRs	Free with DUA	Research-only	Mostly adult
Pediatric Pneumonia CXR (Kermany 2018)	~5k pediatric CXRs	Free	CC BY 4.0	Pediatric
RSNA Pediatric Bone Age	~14k hand X-rays for skeletal age	Free	Research-only DUA	Pediatric
NIH Pediatric MRI Data Repository	Normal pediatric brain MRI	Credentialed	Research-only	Pediatric
IXI dataset	Normal brain MRI (some pediatric)	Free	CC BY-SA 3.0	Mixed
BraTS Pediatric (BraTS-PEDs)	Pediatric brain tumor MRI	Free w/ registration	Challenge-only / research	Pediatric

• Priority. wait-until-needed. All of these are research-only — we cannot ship them inside a commercial product. They are useful for training our own AI models or for benchmarking; outputs of those models can ship.

---

6. Public Clinical Research Datasets

Source	Description	Access	License	Priority
MIMIC-IV	(see §5)	Credentialed	Research-only	wait
eICU-CRD	(see §5)	Credentialed	Research-only	wait
PEDSnet (pedsnet.org)	Multi-institution pediatric EHR network (CHOP-led); >7M children	Restricted — DUA + study approval; not a download	Highly restricted; collaboration only	must-track for partnerships, not ingestible
All of Us (NIH)	1M+ Americans, longitudinal	Researcher Workbench (cloud-only); Tier 1 free, Controlled Tier requires further training	Cannot exfiltrate; analysis stays in workbench	wait-until-needed for research collabs
N3C (NIH)	COVID-cohort EHR	Credentialed enclave	Cannot exfiltrate	wait
PCORnet	National research network; CDM	Network-only	Collaboration	wait
Synthea (synthetichealth.github.io/synthea/)	Synthetic patient data generator	Open source	Apache 2.0	must-have for dev/test
OMOP CDM / OHDSI	Common data model + analytics tooling	Open source	Apache 2.0 / permissive	nice-to-have for analytics layer

• Note on Synthea. Synthea has decent pediatric coverage including immunizations, well-child visits, and growth observations — perfect for seeding our dev/QA environments without HIPAA exposure.

---

7. Licensing Cheat-Sheet (the burn-list)

The mistakes most likely to bite us on launch day:

1. CPT — never display CPT descriptors without an active AMA license. If we are pure cash-pay DPC and avoid CPT entirely, we sidestep this; the moment we add insurance billing, negotiate AMA Intelligent Platform license before code ships.
2. DSM-5-TR criteria text — never copy criteria into our app. Use ICD-10-CM F-codes and validated screeners instead.
3. SNOMED CT outside the US — UMLS license covers US use only. Any international patient or clinician triggers an Affiliate License obligation.
4. DrugBank academic XML — CC BY-NC. Cannot ship commercial.
5. SIDER — CC BY-NC-SA. Cannot ship commercial.
6. Bright Futures Tool Kit — paid AAP product. Periodicity schedule structure is fact-like and we can re-implement, but exact AAP visual layouts and forms are copyrighted.
7. AAP Red Book content — never embed. Link out.
8. MIMIC / eICU / PhysioNet datasets — research-only DUA; never embed in product, only use for model training (and even then check the DUA for derivative-model commercial-use restrictions).
9. Lexicomp / UpToDate / DynaMed / ClinicalKey — paid POC references. Negotiate institutional API licenses before launching prescribing or active CDS.
10. VSAC value sets containing CPT — the value set is free, but redistributing CPT codes inside it still requires a CPT license. Use SNOMED/LOINC/ICD-10 value sets where possible.

---

8. Recommended Ingestion Order for Starlight Practice

Phase 1 — Wire Format & Core Codes (must-have, all free).

ICD-10-CM, SNOMED CT US Edition (via UMLS), LOINC, RxNorm, NDC, UCUM, CVX/MVX, FHIR R4 + US Core, USCDI, VSAC (non-CPT value sets), CDC + WHO growth charts, ACIP CDSi, public-domain screeners (PHQ-9, GAD-7, M-CHAT-R, SWYC, Vanderbilt, PSC).

Phase 2 — Decision Support Foundation (must-have for prescribing).

openFDA SPL (boxed warnings), ACIP CDSi forecasting engine integration, USPSTF AHRQ API, Bright Futures structured periodicity (our own derivative). License a commercial pediatric dosing reference (Lexicomp Pediatric or Micromedex NeoFax) before turning on active prescribing.

Phase 3 — Billing & Payor Integration (only when needed).

CPT license from AMA (if we add insurance billing), HCPCS Level II (free), CPT-containing VSAC sets.

Phase 4 — Advanced AI / Research (wait-until-needed).

Synthea for dev data; PEDSnet/All-of-Us for research partnerships; commercial DrugBank if our own DDI engine isn't adequate; AAP Red Book Online for clinician link-out.

Phase 5 — Specialty Modules (deferred).

DSM-5-TR licensing (only if we build a behavioral-health module beyond ICD-10-CM + screeners), AAP Tool Kit license, ASQ-3 / PEDS commercial screening tools.

---

9. Open Questions Requiring Verification

The following items I could not verify in this session and recommend confirming before relying on them:

1. Current AMA CPT licensing tiers and per-user pricing for an EHR product — contact AMA Intelligent Platform sales.
2. Whether USCDI v5 or USCDI+ Pediatric is finalized as of mid-2026 — check healthit.gov/isa.
3. DrugBank current commercial pricing — contact DrugBank sales.
4. SCARED commercial-use permission — contact authors at WPIC.
5. Bright Futures Tool Kit EHR-vendor site licensing — contact AAP licensing.
6. Lexicomp / Micromedex pediatric API embedding terms — contact Wolters Kluwer / Merative.
7. ECRI Guidelines Trust commercial use — read current ToS.
8. Whether the AAP has published an official FHIR PlanDefinition for the periodicity schedule since 2024 — none authoritative as of my last knowledge.

---

Last updated: 2026-05-07. Verify all licensing terms before commercial deployment.