Drug Retail Data Provider: 16 Cataloged Datasets · Head-to-head
NLM Clinical Table Search Service vs Civil Rights Data Collection (CRDC)
Which drug retail data provider: 16 cataloged datasets data fits your job: NLM Clinical Table Search Service, or Civil Rights Data Collection. API, files, or your warehouse. Daily, weekly, or hourly.
NLM Clinical Table Search Service
Civil Rights Data Collection (CRDC)
Where the fields line up
No shared field names. These two answer different questions.
| Field | NLM Clinical Table Search Service | Civil Rights Data Collection |
|---|---|---|
DISPLAY_NAME | Drug name and route combination for one prescribable concept - the default search and display column of the RxTerms table. | not in this set |
STRENGTHS_AND_FORMS | Pre-built menu of strength-and-form strings available for the drug, one entry per purchasable configuration such as a capsule, tablet or suspension concentration. | not in this set |
RXCUIS | RxNorm concept identifiers aligned element-for-element with the strength-and-form menu, so every purchasable configuration carries its own join key. | not in this set |
SXDG_RXCUI | Single RxNorm identifier for the entity the display name represents - drug plus intended route, independent of any one strength. | not in this set |
DISPLAY_NAME_SYNONYM | Brand names and alternate spellings recorded against the concept, carried without strength associations. | not in this set |
code | Leading identifier column on the coding and registry tables: an ICD-10-CM or HCPCS code, or an NPI number. | not in this set |
name | Human-readable label paired with the code - the full diagnosis description or the registered provider name. | not in this set |
Additional fields on request | Full provider-panel columns for the registry tables (taxonomy, practice and mailing addresses, deactivation status), the remaining lookup tables in the same code-plus-name shape, synonym arrays flattened to one row per alias, and derived views such as pharmacy-type segmentation - pinned down when your sample is cut. | not in this set |
entity_ID | not in this set | documented |
lonG_NCES_ID | not in this set | documented |
entity_Name | not in this set | documented |
leA_Name | not in this set | documented |
Coverage, side by side
| NLM Clinical Table Search Service | Civil Rights Data Collection | |
|---|---|---|
| Geographic | United States by construction — US clinical terminologies and the national provider registry | All fifty states, Washington DC and Puerto Rico — the universe of public schools and districts receiving federal financial assistance |
| Temporal | Current editions, release-stamped per table: RxTerms cut 202608, ICD-10-CM fiscal 2026 edition | Six published waves, 2011-12 through 2021-22 (2019-20 skipped); state and national tables from 2000, district-level archives to 1968, 2023-24 files expected 2026 |
| Granularity | One row per prescribable concept, coding term or registered provider | One observation per school or district, per measure, per student subgroup, per collection wave |
What each contains
They tie on 1 attribute. Pick by fit, not by loyalty.
| NLM Clinical Table Search Service | Civil Rights Data Collection | |
|---|---|---|
| Publisher | U.S. NLM — the National Library of Medicine's Lister Hill National Center for Biomedical Communications | US Department of Education, Office for Civil Rights |
| Subject lens | Clinical vocabulary and provider reference: drug display names with strength-and-form menus, ingredients, diagnosis and procedure codes, LOINC questions, UCUM units and the NPI practitioner and organization registries | K-12 civil-rights census: discipline, harassment, restraint and seclusion, course access, retention, staffing and enrollment, every figure split by student subgroup |
| Unit of analysis | One row per prescribable concept, coding term or registered provider | One observation per school or district, per measure, per student subgroup, per collection wave |
| Geographic coverage | United States by construction — US clinical terminologies and the national provider registry | All fifty states, Washington DC and Puerto Rico — the universe of public schools and districts receiving federal financial assistance |
| Temporal coverage | Current editions, release-stamped per table: RxTerms cut 202608, ICD-10-CM fiscal 2026 edition | Six published waves, 2011-12 through 2021-22 (2019-20 skipped); state and national tables from 2000, district-level archives to 1968, 2023-24 files expected 2026 |
| Detail level | Term-level attributes: strengths, forms, synonyms, RxNorm identifiers, codes and provider addresses | Count-level outcomes: suspensions, arrests, harassment reports, course enrollments and staff headcounts per school per subgroup |
| Documented fields | Seven-field core, definitions verified against live responses during the August 2026 research pass, further columns on request | Thirteen delivered columns covering the entity-measure-subgroup grammar, all verified, with fuller per-year codebooks available as context |
| Scale | About 25 tables: 9,326 RxTerms display names, 2,342 ingredients, 74,719 ICD-10-CM codes, 8,893 HCPCS codes, roughly 109,000 LOINC rows and 9.37 million NPI records | 17,821 local educational agencies and 97,575 schools in the 2020-21 wave alone, hundreds of measures per school-year |
| Best for | Normalizing product catalogs, seeding site search and building pharmacy-footprint baselines | Auditing discipline, course-access and climate equity by race, sex, disability and English-learner status |
What each does better
the NLM Clinical Table Search Service
Normalization at purchase granularity. The pre-built strength-and-form menu makes duplicate listings collapse mechanically: two supplier spellings of the same amoxicillin tablet resolve to one display name, one configuration array, one set of RXCUIS. Typed identifiers, enumerable arrays and a fixed seven-field core mean feature engineering starts immediately — one reason ML model training teams tag the record relevance 3.
Many vocabularies under one grammar. Learn the shape once and it reads across the whole shelf: drugs (RxTerms), ingredients, diagnoses (ICD-10-CM) and procedures (HCPCS), clinical questions (LOINC), measurement units (UCUM), and the people and organizations involved (NPI registry). Joins that would otherwise be fuzzy-matching projects become column mappings onto RxNorm identifiers.
Provider-side depth merchandising rarely gets. The registry separates 1,959,660 organizations from 7,415,294 individuals, with taxonomy and address detail riding on the organization rows. Filter to pharmacy taxonomies and the store-count baseline behind any retail-expansion model falls out directly — and newly enrolled organizations surface market entrants hiding inside reference data.
the Civil Rights Data Collection
A census, not a sample. Reporting covers the universe of federally assisted public schools — 17,821 agencies and 97,575 schools in 2020-21, justice facilities, charters and alternative programs included — so a district benchmark needs no extrapolation, and the smallest alternative program reports on the same form as the largest suburban system.
Subgroup resolution most sources only promise. Nearly every measure splits male, female and total across seven race and ethnicity categories, then again for English learners, students with disabilities under IDEA and Section 504-only students where collected. Restraint and seclusion report IDEA and non-IDEA counts separately, so compliance questions bring their own denominators.
Climate measures found almost nowhere else. Out-of-school and in-school suspensions, expulsions with and without educational services, corporal punishment, referrals to law enforcement, school-related arrests, harassment and bullying reports on the basis of race, sex and disability, physical and mechanical restraint, seclusion — the numbers behind every major federal civil-rights inquiry into American schooling, in one consistent schema.
Trend lines, not anecdotes. Six comparable waves run 2011-12 through 2021-22, long enough to test whether a district's Algebra I gate actually moved after a policy change; state and national tables reach back to 2000 and district-level archives to 1968.
Where they're equivalent
More than their shapes suggest. Both field dictionaries were verified during research, both score 8/10, and both ship complete written dictionaries rather than inferred schemas.
Both key their rows to national identifier systems — RxNorm concept identifiers and NPI numbers on one side, NCES IDs and CRDC entity IDs on the other — so joins outward become column mappings instead of reconciliation projects. Both are federal reference records built for scrutiny: mandatory reporting underneath the census, maintained terminologies underneath the vocabulary. And both ride in the drug-retail slice while answering questions far from the aisle — multi-industry publishers index into many industries at once, which is precisely why a pairing this asymmetric exists to be made.
The verdict
Verdict: sample both — different instruments pointed at different questions that happen to share a catalog shelf, and neither substitutes for the other.
Take the NLM Clinical Table Search Service if your question names a thing. What is this SKU actually called at every strength and form on the market, which organizations hold pharmacy taxonomies in this metro, which diagnosis and procedure codes key dispensing records onto the merchandising view — anything answered by what is this called and what is it. Data scientists and market researchers both score it relevance 3.
Take Civil Rights Data Collection (CRDC) if your question names a place and a disparity. Whether suspension patterns shifted after a policy change, which schools offer calculus and physics and which arrest their students, how enrollment composes across a catchment — anything answered by who experienced what, how often. Journalists and academics live here; the story usually sits in a cross-tab nobody has queried yet.
To see how either sits in the wider slice, start from the ranked best drug retail datasets shortlist.
Sample both, pick by fit. See NLM Clinical Table Search Service · See Civil Rights Data Collection
Or take both in one feed
Yes — stacked, they cover both halves of a location question neither settles alone: what operates here, and who lives with the consequences. A defensible workflow: build the pharmacy footprint first from NPI organization taxonomies and addresses aggregated to metros and states, then overlay Civil Rights Data Collection (CRDC) enrollment composition for the community frame around every store cluster — catchment demographics drawn from mandatory federal counts rather than modeled estimates.
Join discipline decides whether the merge holds. There is no shared key — an RxNorm identifier never names a school district and an NCES ID never names a capsule strength — so relate the two through geography deliberately, rolling provider rows up to whatever level the census reports before comparing anything. Mind the definitions next: census counts arrive privacy-protected, with reserve codes marking suppressed values in the 2020-21 files, and RxTerms synonym arrays flatten to one row per alias only when asked. Browse the rest of the shelf at the drug retail data hub.
Datadory normalizes each record to its documented dictionary before anything ships, delivered daily, weekly, or hourly — your call. Request a sample of either and inspect real rows before anyone commits. Or take both in one feed.
API, files, or your warehouse. Daily, weekly, or hourly.
Fair questions
Is the NLM Clinical Table Search Service better than the Civil Rights Data Collection (CRDC)?
Better at different jobs. The lookup service answers identification questions: 9,326 RxTerms display names with strength-and-form menus, 74,719 ICD-10-CM codes and 9.37 million NPI provider records under one seven-field grammar. CRDC answers measurement questions: discipline, harassment, course access and staffing counted for 17,821 districts and 97,575 schools, split by race, sex, disability and English-learner status. Vocabulary against census — most analyses eventually want both.
Do the two field dictionaries overlap?
In role, not in columns. Both anchor every row on a national identifier — an RxNorm concept identifier like 1152901 against a 12-digit NCES school ID — and both pair it with a readable label and an explicit time marker. Beyond that they diverge by design: seven fields describing terms and their purchasable configurations against thirteen describing schools, measures and student subgroups. Nothing joins them directly.
Which dataset reaches further back?
CRDC, by decades. Six comparable waves run 2011-12 through 2021-22, state and national tables start in 2000, and district-level archives reach back to 1968. The lookup tables ship as release-stamped current editions — RxTerms cut 202608, ICD-10-CM fiscal 2026 — so history accumulates from successive deliveries rather than published waves.
Can Datadory deliver both datasets together?
Yes — separately or merged onto one calendar, delivered daily, weekly, or hourly, your call. Each arrives normalized to its documented dictionary — seven lookup fields on the NLM side, thirteen census columns on the CRDC side — with sample rows attached for inspection before anything ships. Or take both in one feed.