Problems

Goal of Extraction

The Core Clinical Problems Library is a condition-focused clinical data extraction tool designed to identify, extract, and ontologically map active and historical health problems from unstructured clinical narratives and patient health record documentation. It extracts patient diagnoses, clinical conditions, and health issues documented across clinical notes, problem lists, assessment sections, and medical histories, standardizing representation through SNOMED CT concept mapping with associated confidence metrics.

Extraction Methodology

The extraction process operates at the sentence level, capturing each problem with precise character-level locations (start and end positions) that enable direct traceability back to the “document”. When a patient’s health problem appears multiple times within a single document, each instance is extracted and recorded separately to preserve the full documentary record. The library performs extraction on a document-by-document basis; there is no automatic cross-document reconciliation or cross-patient data merging.

Note: A document is determined by our intelligent document splitting feature. Medical PDFs can contain a single document or multiple documents. A document’s type can provide clinical context important for the data extraction. This feature separates and categorizes an input PDF into individual documents by document type to prepare them for data extraction. An input PDF can remain one document or can be split into multiple documents through this process. A “document” in the context of this data extraction refers to the documents identified after the intelligent document splitting process.

Interpretation & Validation

Every extracted health problem is assigned a confidence score (0.0–1.0 scale) to quantify extraction certainty. Problems identified with hypothetical or conditional language are flagged as hypothetical to distinguish them from confirmed diagnoses, while content representing clinical guidance is flagged separately.

Clinical Considerations

Temporal relevance is critical for clinical decision-making. Distinguishing between active versus historical conditions directly impacts patient management and care planning. Anatomical specificity, including laterality and site information, has significant implications for clinical management and risk stratification. Clinical judgment should be applied when interpreting extracted results.

Target Concepts

Coding System: SNOMED CT

All extracted problems are mapped to SNOMED CT (Systematized Nomenclature of Medicine Clinical Terms), with each problem assigned a valid, active SNOMED concept identifier (6–18 digit numeric codes), along with its associated full SNOMED concept description. This structure provides both the standardized data needed for system integration and the contextual information needed for informed clinical interpretation.

Clinical Concepts Extracted

Problems are clinically meaningful conditions that have been observed, assessed or diagnosed by a clinician. This includes diagnoses, clinical conditions, acute and chronic diseases, complications, and comorbidities. Associated clinical attributes—including temporality, anatomical site, and laterality—are captured as they appear in source documentation.

We define problems with SNOMED concepts. Only concepts that fall into one of these clinical categories are included:

• Diseases and syndromes
• Acquired, congenital, or anatomical abnormalities
• Injuries or poisoning
• Mental or behavioral dysfunction
• Neoplastic processes (cancers and tumors)

This means incidental findings, morphologic descriptors, and vague qualifier terms are excluded — we are looking for named, diagnosable conditions.

The concept must be affirmatively stated. Only conditions that are positively asserted in the text are extracted. Negated conditions (e.g., “no history of diabetes”) and speculative language (e.g., “rule out pneumonia”) are not included.

Concepts in the “Review of Systems” section are excluded because this section captures what the patient reports as symptoms rather than what a clinician has assessed or diagnosed.

Note: Problems should only be drawn from sections of the chart that document active clinical assessment. If there are other sections that indicate patient reported conditions, as opposed to clinician assessed or diagnosed conditions, we recommend filtering these records out in a post processing step.

The concept must be a specific, named condition. A small number of overly broad or non-specific SNOMED concepts are excluded by design — for example, generic terms like “disease (disorder)” or “chronic disease (disorder)” add no diagnostic specificity to a problem list.

The extracted concept must meet a confidence threshold. The extraction engine assigns a confidence score to each identified concept. Only concepts where the system is greater than 80% confident in the mapping to a SNOMED concept are included.

Extracted Fields

Problem Extraction Fields

Example of Extracted Output

• Note: This example demonstrates the content of the output, not the structure

Problem Illustrative Output Example

• Example JSONL object showing a typical extraction result

{
  "filename": "patient_1.pdf",
  "subdoc_id": "patient_1.pdf-0007",
  "manifest_id": 7,
  "measure": "problems",
  "chartdate": "2022-02-19",
  "subject_id": "11110001",
  "processing_date": "2024-03-08T00:00:00",
  "data": [
    {
      "sentence": "rface. There is 3mm of | impaction at the dorsal aspect of the fracture, with mild ulnar positive variance. A non- | displaced fracture of the ulnar styloid is also present. |  | IMPRESSION: |  | 1. Right scapular body fracture.  | 2. Right wrist fractures.  |  | The ***scapular body fracture*** was communicated to Dr. Hayward in the ER at 9:15 a.m. |  | Attending Physician: Rajni Ahmad |  | Page 1 of 1",
      "section_name": "IMPRESSION",
      "entity_text": "scapular body fracture",
      "entity_start": 1230,
      "entity_end": 1252,
      "concept_id": "40643005",
      "description": "Fracture of scapular body (disorder)",
      "entity_confidence": 0.96,
      "anatomic_site": "scapular body",
      "laterality": "right",
      "experiencer": "patient",
      "temporality": "9:15 a.m.",
      "is_hypothetical": false,
      "is_guidance": false
    }
  ]
}