For registries
An EHR feed for your cancer registry software.
Clinical Extract pulls your registry's cases from your EHR's certified bulk FHIR API and returns them as a CSV mapped to your data dictionary, ready for your registrars.
Diagnosis, histology, biomarkers, stage, treatment and follow-up, each column traced to its FHIR element.
1 The crosswalk
NAACCR items completed from the EHR.
Clinical Extract exports the coded diagnosis, histology, biomarker, stage, treatment and follow-up data for every case in the cohort and returns a CSV mapped to the NAACCR data dictionary, where each element completes its registry item. Cancer registry data that used to be read out of the chart arrives as coded columns: topography and histology from Condition, ER, PR and HER2 from Observation by LOINC code, systemic therapy from MedicationRequest by RxNorm, surgery from Procedure, and TNM stage from the staging Observations.
The registry decides when to pull, and which cases and items. Casefinding starts from a diagnosis list or a pathology-coded Condition set, the cohort builder returns the count, and the export runs for that cohort alone. This reverses the HL7 Central Cancer Registry Reporting guide, where the EHR pushes a report.
Biomarkers and staging come out as named columns through the oncology lens: ER, PR, HER2, PD-L1, EGFR, KRAS, BRAF, ALK, MSI and TNM clinical and pathologic.
2 Follow-up
Follow-back from a roster of MRNs.
Paste the cases due for annual follow-up: MRNs, or last name, first name and date of birth, up to 500 rows a run. Roster pull grades every row (certain, probable, possible, multiple, none), you confirm the matches, and one export returns the latest vital status, date of last contact, recurrence and treatment for every confirmed case, mapped to the follow-up items the crosswalk marks in crimson.
The matching is the EHR's own: an exact MRN first, then the FHIR Patient/$match operation on name, date of birth and sex, then a demographic search. How a row gets its grade shows the path.
3 Registry types
Trauma, cardiac and quality registries use the same connection.
A clinical registry of any kind draws on the same six kinds of FHIR data. Cancer registries lead: diagnosis, histology, biomarkers, stage, treatment, follow-up. Trauma registries take the injury encounter, its procedures and outcomes. Cardiac registries (NCDR) take procedures, medications and results. Quality registries take the numerator and denominator data of each measure. Patient registry software of every kind imports the CSV; Clinical Extract is the EHR feed behind it, one read-only connection per EHR, each registry keeping its own variable list.
4 Your registry software
Fits the cancer registry software you already run.
The CSV comes with a data dictionary and maps to the NAACCR layout (item numbers and names, version 25). The tumor registry software you run imports it directly, and your abstractors start from a case with the coded fields already filled. Every cancer registry software vendor's product reads that layout.
With Clinical Extract, cancer registry automation means the coded fields are taken from the EHR as recorded, with no model inferring them, and the abstractor reviews what needs the notes. Where an AI cancer registry workflow reads notes, it reads them on top of a case that is already coded. Abstraction starting from the coded record shows the split, with the fields and the minutes per case.
5 Questions
Questions from registries
What is cancer registry software?
The system a cancer registry abstracts, stores and reports its cases in, laid out on the NAACCR data items. Clinical Extract is the EHR feed behind it: it pulls the coded fields of every case from the chart and delivers a CSV in the NAACCR layout that the registry software imports.
What is cancer registry data?
One record per reportable case: the primary site and histology, the stage at diagnosis, the biomarkers, the first course of treatment, and the follow-up items updated every year (vital status, date of last contact, recurrence). Most of it is already coded in the EHR, and Clinical Extract exports that coded part.
What can a central cancer registry pull from an EHR?
Everything the certified bulk FHIR API carries for a cohort: Condition (diagnosis, ICD-10-CM and SNOMED CT), Observation (biomarkers and staging by LOINC), MedicationRequest (systemic therapy by RxNorm), Procedure (surgery and radiation) and Encounter (dates of contact). The crosswalk in Figure 1 maps each one to the NAACCR item it completes.
Does Clinical Extract do the abstraction?
It fills every field that coded EHR data answers and shows the FHIR source next to each value. The abstractor starts at the judgment calls that need the notes. Abstraction starting from the coded record shows the split for each field.
Which registries besides cancer?
Trauma, cardiac (NCDR) and quality registries draw on the same six kinds of FHIR data, so one connection per EHR serves all of them, each registry keeping its own variable list. Cancer registries lead.
How does follow-up work?
Paste the cases due for follow-up as MRNs, or names with dates of birth. Roster pull grades every row, you confirm the matches, and one export returns the latest vital status, date of last contact, recurrence and treatment for every confirmed case.
Next
See Clinical Extract run on one of your studies.
Tell us which EHR you run and what the study or registry needs. We reply within one business day to set a meeting time.