Every oncologist and life sciences researcher knows the uncomfortable truth about Real-World Evidence (RWE) in oncology: the most critical clinical details almost never live in structured billing claims.

While diagnosis codes (like ICD-10) capture that a patient was diagnosed with a solid tumor, they remain completely silent on the real-world clinical nuance that dictates disease progression and patient outcomes. What was the exact substance strength, dosing schedule, and route? When and why did a dose delay or reduction occur? How did the patient navigate the hospital system between oncology clinics, emergency departments, and specialized pain management units?

Over 80% of actionable oncology data points remain locked inside unstructured clinical text, consultation notes, pathology reports, and discharge summaries.

Ahead of the ESMO Congress in Madrid, we are highlighting what becomes possible when advanced Clinical Natural Language Processing (NLP) activates unstructured hospital records and standardizes them into the OMOP Common Data Model (CDM) across European hospital networks.

Activating Clinical Data at Hospital Scale

Turning raw, heterogeneous hospital electronic health records (EHRs) into regulatory-grade oncology cohorts is not a simple data conversion. It requires a robust, privacy-by-design data engineering pipeline deployed directly within participating centers.

Through IOMED’s Health Data Space Platform, unstructured clinical notes, lab results, and hospital pharmacy records were transformed into standardized, research-ready cohorts across Spain:

19,412 multi-center oncology EHR profiles screened → 16,674 validated adult oncology patients → 13,450 solid tumor patients followed longitudinally across five primary indications:

Colorectal Cancer: 4,187 patients (31.1%)
Lung Cancer: 3,347 patients (24.9%)
Breast Cancer: 3,207 patients (23.8%)
Prostate Cancer: 2,583 patients (19.2%)
Bladder Cancer: 1,795 patients (13.3%)
Alongside 2,087 hematological malignancy patients harmonized across lymphoma, multiple myeloma, and leukemia.

Longitudinal Depth & Data Quality

Unlike administrative registries with truncated follow-up or missing baseline records, standardizing hospital EHRs under the OMOP CDM provides deep longitudinal visibility. Across the 13,450 solid tumor patients, the cohort encompasses 33,110+ active person-years of oncology follow-up, with a mean follow-up of 4.5 years post-diagnosis (median 5.5 years, up to 7.0 years). Furthermore, patients had a median baseline history of 19.1 years in their hospital records, ensuring true incident cohort definition and comprehensive baseline comorbidity profiling.

Mapping the Real-World Healthcare Resource Utilization (HCRU) Footprint

A major challenge in Health Economics and Outcomes Research (HEOR) is understanding how oncology patients utilize acute and specialized hospital resources beyond their scheduled chemotherapy infusions.

By harmonizing encounter records, inpatient admissions, and departmental specialty visits into standard OMOP CDM tables, we mapped the multi-year healthcare utilization footprint across the solid tumor cohort:

Acute Inpatient & Emergency Care Demand

Overall, 64.6% of solid tumor patients (8,684 patients) required at least one inpatient hospitalization during follow-up, and 62.1% (8,356 patients) visited the Emergency Department. When examining acute resource consumption across tumor types, critical variations emerge.

Cross-Specialty Outpatient Care

Cancer care extends far beyond medical oncology suites. Overall, 39.3% of solid tumor patients were tracked across specialized outpatient clinics, averaging 2.1 specialized consultations per year:

Beyond ICD Codes: Over 1.11 Million Enriched Prescribing Records

The most significant leap forward achieved through Clinical NLP is moving past simple drug mentions to granular, linked prescribing observations.

Within this cohort, 1,113,771 linked prescribing detail records—capturing substance strength, dose, frequency, and quantity—were extracted directly from hospital text and mapped to OMOP RxNorm across 137 standardized oncology and supportive therapies

Attending ESMO in Madrid? Let’s Unlock Oncology RWE Together

The European Health Data Space (EHDS) is setting a new standard for secondary use of health data. But having access to hospital data is only the beginning—data must be clinically activated, standardized, and deep enough to address regulatory and HTA evidence requirements.

IOMED’s Health Data Space Platform bridges this gap by deploying privacy-by-design infrastructure (ISO 27701:2019 certified) directly within hospital networks. Through our platform, life sciences teams and clinical researchers gain access to rapid feasibility counts, aggregated characterization, and research-ready patient-level datasets.

‍