Smarter hospitals start here

Amorphous turns your clinical notes into structured, queryable data your teams can act on.

View Demo

Around 80% of the clinically relevant information in a hospital sits in free text.

Notes, discharge summaries, surgical reports. Structuring it multiplies the data points your teams can actually analyse and leverage.

Discharge SummaryCardiology · Page 1 of 12026-09-27

Background of hypertension1 and type 2 diabetes2, maintained on metformin3. Previous appendectomy4 in 1998. Documented penicillin allergy5. Admission troponin T6 34 ng/L with LVEF7 38%. Impression: decompensated congestive heart failure8.

8 Concepts Coded4 Systems
  • 1HypertensionI10ICD-10
  • 2Type 2 diabetesE11.9ICD-10
  • 3Metformin6809RxNorm
  • 4Appendectomy80146002SNOMED CT
  • 5Penicillin allergy91936005SNOMED CT
  • 6Troponin T6598-7LOINC
  • 7Ejection fraction10230-1LOINC
  • 8Congestive heart failureI50.9ICD-10

Clinical notes become insights

The Data Engine reads clinical notes and turns them into coded, structured data. Striata sits on top and answers questions in plain language.

source

Clinical notes

Unstructured notes, discharge summaries and reports.

Discharge SummaryTXT

Background of hypertension and type 2 diabetes, maintained on metformin.

Previous appendectomy in 1998. Documented penicillin allergy.

Admission troponin T 34 ng/L with LVEF 38%. Impression: decompensated congestive heart failure.

Data Engine

Structured data warehouse

Ingests text and writes coded records.

Clinical conceptCode
  • ConditionHypertensionICD-10I10
  • MedicationMetforminRxNorm6809
  • ObservationEjection fraction38%LOINC10230-1
Striata

Answers

Questions asked in natural language.

Question

What does the discharge summary show?

Clinical Summary
BackgroundHypertension
MedicationMetformin
Ejection fraction38%
LVEF
Source: Discharge Summary

Cohort-driven reproducibility

Ask Striata the same question twice and it outputs identical cohorts. Each report surfaces the scripts and data behind it, so any analysis can be rerun, audited, or handed to a reviewer.

Post-MI patients on secondary-prevention therapy at 12 months
Run 01
  • Cohort1,284
  • Excluded216
  • Median age67
Run 02
  • Cohort1,284
  • Excluded216
  • Median age67

Data Engine understands meaning

Most systems detect keywords. Amorphous captures the clinical semantics of the sentence.

Clinical note

01
…discussed starting apixaban. Patient declined. Bisoprolol stopped after 3 years…

Clinical note

02
History: HTN (on lisinopril), appendectomy 2010. Admitted for acute CHF exacerbation…

Clinical note

03
…father died of colon cancer. Possible pneumonia on imaging…

If a system cannot tell these apart, every answer built on top of it is quietly wrong.

How accurate is it?

Inter-note agreement
0–0%

How often clinicians agree with each other when structuring the same clinical notes.

0.0%

How often clinicians agree with the output of our pipeline when they review it.

Clinicians agree with the pipeline more often than clinicians agree with each other.

Ground-truth recovery
0%

Hidden findings recovered by Striata in ground-truth tests. We inject known facts into a large dataset and ask Striata to identify the biases.

Example query

Find diabetes patients whose HbA1c stayed above 9% on two consecutive tests with no change in therapy in between.

Internal benchmarks on clinician-annotated data. Full results available soon.

Use cases

Amorphous tools replace tasks that hospital teams do manually

Clinical

Which breast-cancer patients developed febrile neutropenia, and was treatment aligned with protocol?

60min

Time to first IV antibiotics

NHS Borders’ febrile-neutropenia pathway specifies IV antibiotics within one hour, without waiting for blood results.

NHS Borders (opens in a new tab)

Financial

Where do documented diagnoses, procedures or complications point to incomplete DRG coding and an understated case mix?

5%

Additional tariff income identified

A two-hospital NHS audit of 8,888 emergency admissions identified £816,977 in additional tariff income after coding review. Published in 2016.

Imperial College Healthcare NHS Trust (opens in a new tab)

Operational

Where are patients waiting too long across the oncology pathway?

10working days

Booking to diagnostic tests

Latvia’s Green Corridor specifies this window for state-funded diagnostic tests when cancer is suspected.

National Health Service of Latvia (opens in a new tab)

Clinical trials

Which patients match a trial phenotype across diagnosis, biomarkers, treatment history and free text?

98%

Precision extracting patient attributes

A single-centre study extracted eight patient and tumour characteristics from 640 lung-cancer meeting reports using rule-based NLP.

Tours University Hospital (opens in a new tab)

Audit and reporting

Which patients with abnormal findings lack documented follow-up?

25%

State-funded follow-up recorded

After positive or suspicious cervical screening results in the 2023 audit. Privately funded follow-up was not captured.

State Audit Office of Latvia (opens in a new tab)

Works with existing infrastructure

Amorphous reads from your EMR and feeds your existing warehouse with structured, coded data compatible with FHIR R4 and OMOP. Your current infrastructure becomes more valuable.

Your infrastructure

EMR System

Notes, summaries and reports stay where they are.

Amorphous layer

Amorphous

Structures and codes the free text, continuously.

FHIR R4OMOP CDM

Your infrastructure

Data Warehouse

Receives structured, coded records in open standards.

Why Amorphous

Depth

18 entity types that capture nuanced attributes, coded to SNOMED CT, ICD-10, RxNorm, LOINC and ATC.

Flexibility

Support for any ontology that your healthcare system works with, such as national procedure ontologies or regional ICD-10 variants.

Sovereignty

All processing inside the EU on European infrastructure. Customer data is never used to train models.

Common questions

More questions, including certifications, audits, breach handling and sub-processors, are answered on the Compliance page.

Deploy Striata now.Let's talk.

For partnerships

hello@amorphous.health
Get in touch