Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
uid: string
states: list<item: struct<risk_id: string, state: string, confidence: double, severity: string, volatility_p (... 553 chars omitted)
  child 0, item: struct<risk_id: string, state: string, confidence: double, severity: string, volatility_profile: str (... 541 chars omitted)
      child 0, risk_id: string
      child 1, state: string
      child 2, confidence: double
      child 3, severity: string
      child 4, volatility_profile: string
      child 5, timeline: list<item: string>
          child 0, item: string
      child 6, resolution_evidence: string
      child 7, escalation_evidence: string
      child 8, last_updated: int64
      child 9, reason: struct<level_1_evidence: list<item: struct<text: string, type: string, source: string>>, level_2_int (... 302 chars omitted)
          child 0, level_1_evidence: list<item: struct<text: string, type: string, source: string>>
              child 0, item: struct<text: string, type: string, source: string>
                  child 0, text: string
                  child 1, type: string
                  child 2, source: string
          child 1, level_2_interpretation: list<item: string>
              child 0, item: string
          child 2, level_3_decision_rule: struct<rule: string, threshold: double>
              child 0, rule: string
              child 1, threshold: double
          child 3, reasoning_schema: struct<decision_type: string, input_signal: list<item: string>, severity_marker: string, causal_link (..
...
ce: double
      child 3, based_on_risks: list<item: string>
          child 0, item: string
      child 4, justification: string
      child 5, constraints: list<item: string>
          child 0, item: string
      child 6, reason: struct<level_1_evidence: list<item: struct<text: string, type: string, source: string>>, level_2_int (... 311 chars omitted)
          child 0, level_1_evidence: list<item: struct<text: string, type: string, source: string>>
              child 0, item: struct<text: string, type: string, source: string>
                  child 0, text: string
                  child 1, type: string
                  child 2, source: string
          child 1, level_2_interpretation: list<item: string>
              child 0, item: string
          child 2, level_3_decision_rule: struct<rule: string, constraints_applied: int64>
              child 0, rule: string
              child 1, constraints_applied: int64
          child 3, reasoning_schema: struct<decision_type: string, input_signal: list<item: string>, severity_marker: string, causal_link (... 50 chars omitted)
              child 0, decision_type: string
              child 1, input_signal: list<item: string>
                  child 0, item: string
              child 2, severity_marker: string
              child 3, causal_link: list<item: string>
                  child 0, item: string
              child 4, verification_anchor: string
          child 4, conclusion: string
          child 5, method: string
to
{'uid': Value('string'), 'recommendations': List({'rec_id': Value('string'), 'type': Value('string'), 'confidence': Value('float64'), 'based_on_risks': List(Value('string')), 'justification': Value('string'), 'constraints': List(Value('string')), 'reason': {'level_1_evidence': List({'text': Value('string'), 'type': Value('string'), 'source': Value('string')}), 'level_2_interpretation': List(Value('string')), 'level_3_decision_rule': {'rule': Value('string'), 'constraints_applied': Value('int64')}, 'reasoning_schema': {'decision_type': Value('string'), 'input_signal': List(Value('string')), 'severity_marker': Value('string'), 'causal_link': List(Value('string')), 'verification_anchor': Value('string')}, 'conclusion': Value('string'), 'method': Value('string')}}), 'provenance': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1816, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              uid: string
              states: list<item: struct<risk_id: string, state: string, confidence: double, severity: string, volatility_p (... 553 chars omitted)
                child 0, item: struct<risk_id: string, state: string, confidence: double, severity: string, volatility_profile: str (... 541 chars omitted)
                    child 0, risk_id: string
                    child 1, state: string
                    child 2, confidence: double
                    child 3, severity: string
                    child 4, volatility_profile: string
                    child 5, timeline: list<item: string>
                        child 0, item: string
                    child 6, resolution_evidence: string
                    child 7, escalation_evidence: string
                    child 8, last_updated: int64
                    child 9, reason: struct<level_1_evidence: list<item: struct<text: string, type: string, source: string>>, level_2_int (... 302 chars omitted)
                        child 0, level_1_evidence: list<item: struct<text: string, type: string, source: string>>
                            child 0, item: struct<text: string, type: string, source: string>
                                child 0, text: string
                                child 1, type: string
                                child 2, source: string
                        child 1, level_2_interpretation: list<item: string>
                            child 0, item: string
                        child 2, level_3_decision_rule: struct<rule: string, threshold: double>
                            child 0, rule: string
                            child 1, threshold: double
                        child 3, reasoning_schema: struct<decision_type: string, input_signal: list<item: string>, severity_marker: string, causal_link (..
              ...
              ce: double
                    child 3, based_on_risks: list<item: string>
                        child 0, item: string
                    child 4, justification: string
                    child 5, constraints: list<item: string>
                        child 0, item: string
                    child 6, reason: struct<level_1_evidence: list<item: struct<text: string, type: string, source: string>>, level_2_int (... 311 chars omitted)
                        child 0, level_1_evidence: list<item: struct<text: string, type: string, source: string>>
                            child 0, item: struct<text: string, type: string, source: string>
                                child 0, text: string
                                child 1, type: string
                                child 2, source: string
                        child 1, level_2_interpretation: list<item: string>
                            child 0, item: string
                        child 2, level_3_decision_rule: struct<rule: string, constraints_applied: int64>
                            child 0, rule: string
                            child 1, constraints_applied: int64
                        child 3, reasoning_schema: struct<decision_type: string, input_signal: list<item: string>, severity_marker: string, causal_link (... 50 chars omitted)
                            child 0, decision_type: string
                            child 1, input_signal: list<item: string>
                                child 0, item: string
                            child 2, severity_marker: string
                            child 3, causal_link: list<item: string>
                                child 0, item: string
                            child 4, verification_anchor: string
                        child 4, conclusion: string
                        child 5, method: string
              to
              {'uid': Value('string'), 'recommendations': List({'rec_id': Value('string'), 'type': Value('string'), 'confidence': Value('float64'), 'based_on_risks': List(Value('string')), 'justification': Value('string'), 'constraints': List(Value('string')), 'reason': {'level_1_evidence': List({'text': Value('string'), 'type': Value('string'), 'source': Value('string')}), 'level_2_interpretation': List(Value('string')), 'level_3_decision_rule': {'rule': Value('string'), 'constraints_applied': Value('int64')}, 'reasoning_schema': {'decision_type': Value('string'), 'input_signal': List(Value('string')), 'severity_marker': Value('string'), 'causal_link': List(Value('string')), 'verification_anchor': Value('string')}, 'conclusion': Value('string'), 'method': Value('string')}}), 'provenance': Value('string')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1683, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1869, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

uid
string
recommendations
list
provenance
string
7665777-11
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-7665777-11
7665777-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.97, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-7665777-1
7665777-8
[ { "rec_id": "REC_RESPIRATORY_SUPPORT", "type": "INTERVENTION", "confidence": 0.95, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[ESCALATED] RISK_RESPIRATORY_DISTRESS: signals detected — dyspnea [Interaction: anxiety↔respiratory vicious circle detected]", "co...
VERIF-7665777-8
8674153-2
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8674153-2
7665777-5
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-7665777-5
7665777-4
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.96, "based_on_risks": [ "RISK_CLINICAL_DETERIORATION" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for...
VERIF-7665777-4
7665777-7
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.97, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-7665777-7
7665777-6
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.97, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-7665777-6
7665777-2
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS", "RISK_ANXIETY_EXACERBATION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ ...
VERIF-7665777-2
7665777-9
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.96, "based_on_risks": [ "RISK_CLINICAL_DETERIORATION" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for...
VERIF-7665777-9
8674458-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8674458-1
7665777-3
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-7665777-3
7665777-10
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.97, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-7665777-10
8674405-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8674405-1
8674685-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8674685-1
8674153-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_CLINICAL_DETERIORATION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", ...
VERIF-8674153-1
8675571-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8675571-1
8675572-1
[ { "rec_id": "REC_CARDIAC_MONITORING", "type": "MONITORING", "confidence": 0.91, "based_on_risks": [ "RISK_CARDIAC_EVENT" ], "justification": "[ACTIVE] RISK_CARDIAC_EVENT: signals detected — myocardial infarction", "constraints": [ "12-Lead ECG", "Troponin Serial", ...
VERIF-8675572-1
8675589-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8675589-1
8675574-1
[ { "rec_id": "REC_NEURO_ASSESSMENT", "type": "MONITORING", "confidence": 0.9, "based_on_risks": [ "RISK_NEUROLOGICAL_DETERIORATION" ], "justification": "[ACTIVE] RISK_NEUROLOGICAL_DETERIORATION: signals detected — confusion", "constraints": [ "GCS Assessment", "Neurology...
VERIF-8675574-1
8675578-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8675578-1
8675577-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8675577-1
8675592-2
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-8675592-2
8675592-1
[ { "rec_id": "REC_VTE_PROPHYLAXIS", "type": "PREVENTION", "confidence": 0.83, "based_on_risks": [ "RISK_VTE" ], "justification": "[ACTIVE] RISK_VTE: signals detected — d-dimer elevated", "constraints": [ "VTE Risk Score (Caprini)", "LMWH Prophylaxis if not contraindicate...
VERIF-8675592-1
8675596-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-8675596-1
8675596-2
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-8675596-2
8675591-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8675591-1
8675595-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_NEUROLOGICAL_DETERIORATION", "RISK_FALL_DECONDITIONING", "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring re...
VERIF-8675595-1
8675594-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.78, "based_on_risks": [ "RISK_FALL_DECONDITIONING" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for De...
VERIF-8675594-1
8675570-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_RENAL_FAILURE", "RISK_ANXIETY_EXACERBATION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP F...
VERIF-8675570-1
8675596-3
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-8675596-3
8675583-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8675583-1
8676056-2
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.68, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Diagnosis 'sapho syndrome' has no linked treatment in the clinical graph.", "constraints": [ "Verify Medication Orders"...
VERIF-8676056-2
8675801-1
[ { "rec_id": "REC_VTE_PROPHYLAXIS", "type": "PREVENTION", "confidence": 0.83, "based_on_risks": [ "RISK_VTE" ], "justification": "[ACTIVE] RISK_VTE: signals detected — swelling", "constraints": [ "VTE Risk Score (Caprini)", "LMWH Prophylaxis if not contraindicated", ...
VERIF-8675801-1
8675596-4
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-8675596-4
8675599-1
[ { "rec_id": "REC_CARDIAC_MONITORING", "type": "MONITORING", "confidence": 0.95, "based_on_risks": [ "RISK_CARDIAC_EVENT" ], "justification": "[ACTIVE] RISK_CARDIAC_EVENT: signals detected — chest pain, troponin", "constraints": [ "12-Lead ECG", "Troponin Serial", ...
VERIF-8675599-1
8676056-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_VTE", "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up wit...
VERIF-8676056-1
8677335-1
[ { "rec_id": "REC_INTENSIVE_MONITORING", "type": "MONITORING", "confidence": 0.88, "based_on_risks": [ "RISK_CLINICAL_DETERIORATION" ], "justification": "[ACTIVE] RISK_CLINICAL_DETERIORATION: signals detected — hypotension, altered mental status", "constraints": [ "Hourly Vita...
VERIF-8677335-1
8692721-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_SEPSIS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for Decompensation"...
VERIF-8692721-1
8690961-1
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.724, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Consolidated 3 signals (secondary central precocious puberty, benign adrenal cortical adenoma, estrogen-secreting right adrenoco...
VERIF-8690961-1
8691080-1
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.786, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Consolidated 14 signals (stage ivb lung adenocarcinoma combing with bone metastases, neuroendocrine tumor, secondary mds) indica...
VERIF-8691080-1
8695686-1
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.68, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Diagnosis 'left superior carotid-ophthalmic aneurysm' has no linked treatment in the clinical graph.", "constraints": [ ...
VERIF-8695686-1
8693547-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.97, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-8693547-1
8695686-2
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.744, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Consolidated 5 signals (fisher i, subarachnoid hemorrhage, hunt, and hess ii) indicating MEDIUM RISK_TREATMENT_OMISSION.", "...
VERIF-8695686-2
8693548-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.98, "based_on_risks": [ "RISK_NEUROLOGICAL_DETERIORATION" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor...
VERIF-8693548-1
8693491-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8693491-1
8693538-1
[ { "rec_id": "REC_FALL_PREVENTION", "type": "PREVENTION", "confidence": 0.78, "based_on_risks": [ "RISK_FALL_DECONDITIONING" ], "justification": "[ACTIVE] RISK_FALL_DECONDITIONING: signals detected — balance", "constraints": [ "Falls Risk Assessment", "Physiotherapy Refe...
VERIF-8693538-1
8693492-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.91, "based_on_risks": [ "RISK_VTE" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for Decompensation" ...
VERIF-8693492-1
8695605-1
[ { "rec_id": "REC_CARDIAC_MONITORING", "type": "MONITORING", "confidence": 0.91, "based_on_risks": [ "RISK_CARDIAC_EVENT" ], "justification": "[ACTIVE] RISK_CARDIAC_EVENT: signals detected — chest tightness", "constraints": [ "12-Lead ECG", "Troponin Serial", "Card...
VERIF-8695605-1
8693539-1
[ { "rec_id": "REC_NEURO_ASSESSMENT", "type": "MONITORING", "confidence": 0.9, "based_on_risks": [ "RISK_NEUROLOGICAL_DETERIORATION" ], "justification": "[ACTIVE] RISK_NEUROLOGICAL_DETERIORATION: signals detected — disorientation", "constraints": [ "GCS Assessment", "Neur...
VERIF-8693539-1
8693546-1
[ { "rec_id": "REC_SEPSIS_BUNDLE", "type": "IMMEDIATE_ACTION", "confidence": 0.92, "based_on_risks": [ "RISK_SEPSIS" ], "justification": "[ESCALATED] RISK_SEPSIS: signals detected — septic shock", "constraints": [ "Blood Cultures x2", "IV Antibiotics within 1hr", "I...
VERIF-8693546-1
8693699-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", ...
VERIF-8693699-1
8694018-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8694018-1
8693695-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.78, "based_on_risks": [ "RISK_FALL_DECONDITIONING" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for De...
VERIF-8693695-1
8695790-2
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.922, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for De...
VERIF-8695790-2
8695790-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8695790-1
8695756-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_CARDIAC_EVENT", "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Fol...
VERIF-8695756-1
8695792-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8695792-1
8695792-3
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8695792-3
8695792-2
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8695792-2
8695924-1
[ { "rec_id": "REC_SEPSIS_BUNDLE", "type": "IMMEDIATE_ACTION", "confidence": 0.92, "based_on_risks": [ "RISK_SEPSIS" ], "justification": "[ESCALATED] RISK_SEPSIS: signals detected — septic shock", "constraints": [ "Blood Cultures x2", "IV Antibiotics within 1hr", "I...
VERIF-8695924-1
8695924-2
[ { "rec_id": "REC_SEPSIS_BUNDLE", "type": "IMMEDIATE_ACTION", "confidence": 0.92, "based_on_risks": [ "RISK_SEPSIS" ], "justification": "[ESCALATED] RISK_SEPSIS: signals detected — septic shock", "constraints": [ "Blood Cultures x2", "IV Antibiotics within 1hr", "I...
VERIF-8695924-2
8695846-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8695846-1
8697958-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.856, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for De...
VERIF-8697958-1
8696182-2
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-8696182-2
8696182-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-8696182-1
8696182-3
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for D...
VERIF-8696182-3
8698474-1
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.724, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Consolidated 3 signals (pigmented lesion, melanocytomas, gna11 c.626a>t p.q209l oncogenic mutation) indicating MEDIUM RISK_TREAT...
VERIF-8698474-1
8696155-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8696155-1
8698877-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", ...
VERIF-8698877-1
8696256-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8696256-1
8698328-2
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_CARDIAC_EVENT" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Pa...
VERIF-8698328-2
8698328-1
[ { "rec_id": "REC_RESPIRATORY_SUPPORT", "type": "INTERVENTION", "confidence": 0.93, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[ACTIVE] RISK_RESPIRATORY_DISTRESS: signals detected — shortness of breath", "constraints": [ "Titrate Oxygen", "Posit...
VERIF-8698328-1
8699387-1
[ { "rec_id": "REC_FALL_PREVENTION", "type": "PREVENTION", "confidence": 0.78, "based_on_risks": [ "RISK_FALL_DECONDITIONING" ], "justification": "[ACTIVE] RISK_FALL_DECONDITIONING: signals detected — weakness", "constraints": [ "Falls Risk Assessment", "Physiotherapy Ref...
VERIF-8699387-1
8697878-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", ...
VERIF-8697878-1
8698542-1
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.91, "based_on_risks": [ "RISK_CARDIAC_EVENT" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for Decompen...
VERIF-8698542-1
8699357-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", ...
VERIF-8699357-1
8699499-2
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.708, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Consolidated 2 signals (multilocular schwannoma, intracochlear schwannoma) indicating MEDIUM RISK_TREATMENT_OMISSION.", "con...
VERIF-8699499-2
8699499-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8699499-1
8699499-3
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.68, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Diagnosis 'ic vs' has no linked treatment in the clinical graph.", "constraints": [ "Verify Medication Orders", "...
VERIF-8699499-3
8699863-1
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.68, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Diagnosis 'sars-cov-2' has no linked treatment in the clinical graph.", "constraints": [ "Verify Medication Orders", ...
VERIF-8699863-1
8699863-3
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.708, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Consolidated 2 signals (acute complicated appendicitis, acute uncomplicated appendicitis) indicating MEDIUM RISK_TREATMENT_OMISS...
VERIF-8699863-3
8699863-2
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.752, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Consolidated 6 signals (broad intra-abdominal inflammation, peritonitis, acute covid-19 infection) indicating MEDIUM RISK_TREATM...
VERIF-8699863-2
8699863-4
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8699863-4
8698451-1
[ { "rec_id": "REC_RESPIRATORY_SUPPORT", "type": "INTERVENTION", "confidence": 0.97, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[ACTIVE] RISK_RESPIRATORY_DISTRESS: signals detected — dyspnea", "constraints": [ "Titrate Oxygen", "Positioning Maneu...
VERIF-8698451-1
8699836-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8699836-1
8699610-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_NEUROLOGICAL_DETERIORATION", "RISK_ANXIETY_EXACERBATION", "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring r...
VERIF-8699610-1
8699863-6
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8699863-6
8699863-5
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8699863-5
8699977-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8699977-1
8699918-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", ...
VERIF-8699918-1
8699977-2
[ { "rec_id": "REC_CONTINUE_CURRENT_PLAN", "type": "MAINTENANCE", "confidence": 0.99, "based_on_risks": [ "RISK_GI_BLEED" ], "justification": "[MANAGED] Risk controlled — maintain current care plan.", "constraints": [ "Continue Current Therapy", "Monitor for Decompensatio...
VERIF-8699977-2
8699826-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8699826-1
8699901-1
[ { "rec_id": "REC_PLAN_REVIEW", "type": "INTERVENTION_CHECK", "confidence": 0.791, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "[ACTIVE] Consolidated 16 signals (coarctation of the aorta, ventricular septal defect, atrioventricular and ventriculoarterial concordanc...
VERIF-8699901-1
8699977-3
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8699977-3
8699959-1
[ { "rec_id": "REC_VTE_PROPHYLAXIS", "type": "PREVENTION", "confidence": 0.83, "based_on_risks": [ "RISK_VTE" ], "justification": "[ACTIVE] RISK_VTE: signals detected — swelling", "constraints": [ "VTE Risk Score (Caprini)", "LMWH Prophylaxis if not contraindicated", ...
VERIF-8699959-1
8699933-1
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8699933-1
8700032-3
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", "Patient Education on Warning Signs...
VERIF-8700032-3
8699839-1
[ { "rec_id": "REC_RESPIRATORY_SUPPORT", "type": "INTERVENTION", "confidence": 0.97, "based_on_risks": [ "RISK_RESPIRATORY_DISTRESS" ], "justification": "[ACTIVE] RISK_RESPIRATORY_DISTRESS: signals detected — dyspnea", "constraints": [ "Titrate Oxygen", "Positioning Maneu...
VERIF-8699839-1
8700032-4
[ { "rec_id": "REC_OUTPATIENT_FOLLOW_UP", "type": "DISCHARGE_PLANNING", "confidence": 0.9, "based_on_risks": [ "RISK_TREATMENT_OMISSION" ], "justification": "All identified risks resolved. Outpatient monitoring recommended.", "constraints": [ "GP Follow-up within 2 weeks", ...
VERIF-8700032-4
End of preview.

PRD Clinical Multi-Task Corpus

A production-scale, LLM-generated clinical multi-task corpus derived from the PMC-Patients patient narratives. Each of the 167,034 source narratives is processed into a structured, multi-task record containing named entities, relations, question answering with reasoning traces, medications, temporal events, a grounded summary, and downstream deterministic risk-reasoning artefacts.

This corpus accompanies the paper "How much of an LLM-generated clinical corpus is actually new? A production-scale measurement of content redundancy for provenance classification" Paper, Github and is released to enable reproduction of the provenance-based redundancy decomposition (PRD) reported there.

Important. This corpus is LLM-generated and has not been clinically validated. It is released for research on corpus structure, content provenance, and NLP method development. It is not a gold-standard annotation resource and must not be used for clinical decision-making or patient care. Generated fields (entities, relations, answers, risk statements) are model outputs and may contain errors.

Dataset at a glance

Property Value
Records (patients) 167,034 (163,400 human, 3,634 veterinary)
Source PMC-Patients (Zhao et al., 2023), CC-BY-NC-SA-4.0
Generator model Llama-3.3-70B (instruction-tuned)
Verifier model MMed-Llama-3.1-70B
Tasks per record 7 in-record channels + 1 visualization payload + 3 downstream risk channels
Total generated tokens ~2.51B
Format JSON Lines (one patient per line)
License CC-BY-NC-SA-4.0

Data Generation and Verification Pipeline

The PRD Clinical Corpus was generated through a large-scale generator–verifier pipeline. Clinical narratives from the PMC-Patients corpus were processed by a general-purpose large language model to generate candidate clinical information across multiple NLP tasks. A separate medical-domain model then assessed the generated outputs against the source clinical narrative and task-specific verification criteria.

The resulting verifier-assessed outputs were assembled into the PRD Clinical Corpus together with provenance and redundancy information, enabling downstream research in clinical NLP, synthetic clinical data auditing, provenance analysis, and model training.

Corpus statistics

Aggregate item counts across all 167,034 records:

Task channel Total items Records with the task
Named entities (ner) 4,471,110 167,034
Relations (re) 4,129,537 166,091
Temporal events (temporal_events) 2,614,040 166,744
Question answering (qa) 1,035,199 165,082
Risk-oriented QA (risk_qa) 760,483 123,724
Medications (medications) 372,969 100,556
Summaries (summary) 167,034 167,034

Downstream deterministic risk channels:

Channel Total items Records
Active risks (active_v6_risks) 193,364 167,034
Risk states (active_v6_risk_states) 193,364 167,034
Recommendations (active_v6_recommendations) 215,435 167,034

Named-entity categories (4,471,110 entities, 11 categories)

Category Count
Tests 1,356,195
Treatment 914,530
Symptoms 826,180
Diagnosis 798,304
Observation 236,837
Unknown 179,855
History 106,569
Risk_Factor 43,000
Outcome_ClinicalStatus 6,557
Outcome_Disposition 3,068
Adverse_Event 15

Relation labels (4,129,537 relations, 8 labels)

Label Count
ASSOCIATED_WITH 2,527,000
REVEALS 604,305
MANIFESTS 517,165
TREATS 265,732
PREVENTS 100,613
GUIDED_BY 74,186
COMPLICATED_BY 23,709
CAUSES 16,827

Question-answering categories (1,035,199 pairs, 11 categories)

Category Count
Diagnosis 158,716
Treatment 138,488
Tests 137,969
Symptoms 131,205
Outcome_ClinicalStatus 104,455
Observation 99,619
Adverse_Event 82,637
Outcome_Disposition 72,550
History 66,891
Outcome_Mortality 23,185
Risk_Factor 19,484

Note that these counts are of generated items; the confidence and verifier_status fields record the pipeline's own assessment, not verified ground truth.

How the corpus was generated (summary)

Each PMC-Patients narrative was processed by a general-purpose generator (Llama-3.3-70B) proposing candidate extractions, with a medical-domain verifier (MMed-Llama-3.1-70B) attesting each candidate; a deterministic clinical-policy layer then derives downstream risk-reasoning artefacts from the verified evidence. Every task output is emitted as a structured JSON object. For each extraction the pipeline records the model confidence and a verifier status, enabling downstream filtering. Full generation configuration (decoding parameters, serving setup) is provided in the accompanying paper's Supplementary Information.

The corpus exhibits substantial content redundancy by construction: source context is re-serialized into per-item fields across tasks, so raw token count greatly overstates the amount of lexically novel content. Measuring and decomposing this redundancy is the subject of the accompanying paper.

Files

File Description
multitask_data_enriched.jsonl Main corpus: one patient per line, all in-record task channels
visualization_payload.jsonl Per-patient knowledge-graph visualization payload
active_v6_risks.jsonl Downstream: active risks per patient (deterministic)
active_v6_risk_states.jsonl Downstream: temporal risk-state transitions
active_v6_recommendations.jsonl Downstream: recommendations per patient

Records are joined across files by the patient identifier (uid).

Record structure

Each line of multitask_data_enriched.jsonl is a JSON object:

{
  "uid": "7665777-11",
  "text": "This 77-year-old male patient was transferred to our ICU ...",
  "meta_species": "human",
  "tasks": {
    "ner": [ ... ],
    "qa": [ ... ],
    "re": [ ... ],
    "medications": [ ... ],
    "temporal_events": [ ... ],
    "risk_qa": [ ... ],
    "summary": "..."
  }
}
  • uid — patient identifier of the form PMID-x, where PMID is the PubMed identifier of the source article and x is the patient index within that article (inherited from PMC-Patients).
  • text — the source patient narrative (from PMC-Patients).
  • meta_specieshuman or veterinary.
  • tasks — the seven in-record task channels, detailed below.

Task channels (in depth)

1. ner — Named Entity Recognition

List of extracted clinical entities. Each entity:

Field Type Description
entity string The entity span text
category string Clinical category (e.g. Diagnosis, Treatment, Symptoms, Unknown)
start_char, end_char int Character offsets into text
llm_confidence float Generator's self-reported confidence
assertion_status string e.g. Present, Absent
is_negated_source bool Whether the mention is negated in the source
confidence float Combined confidence score
verifier_status string Verifier verdict (e.g. PASS)

2. qa — Question Answering with reasoning

List of question-answer items with structured reasoning traces. Each item:

Field Type Description
qa_id string Unique QA identifier
question string The generated question
answer string The extracted answer
category string Question category (e.g. Diagnosis)
reason object Structured reasoning trace (see below)
context string Source context the answer is grounded in
meta object Additional metadata
verifier_status string Verifier verdict
verdict_path string/array Trace of the verification decision

The reason object contains a multi-level reasoning trace, including an interpretation, a decision rule with a confidence score, a conclusion, the extraction method, and a structured reasoning_schema (decision type, input signal, severity marker, causal links). This is the richest channel and a primary contributor to copied-context redundancy, since context re-serializes source text per item.

3. re — Relation Extraction

List of typed relations between entities. Each relation:

Field Type Description
rel_id string Unique relation identifier
head, tail string The two related entities
head_type, tail_type string Entity types
label string Relation type
direction string Relation direction
confidence float Confidence score
context string Source context grounding the relation
verifier_status, status string Verifier verdict and final status

4. medications — Medication Extraction

List of extracted medications with associated attributes (drug, and where present route, frequency, and related fields).

5. temporal_events — Temporal Event Extraction

List of clinical events with temporal anchoring. Each event:

Field Type Description
event_id string Unique event identifier
sentence string Source sentence containing the event
timepoint string Extracted timepoint
timepoint_type string Type of timepoint
event_type string Type of event
entities_involved list Entities participating in the event
temporal_order int/string Ordinal position in the timeline
is_anchored bool Whether the event is anchored to a concrete timepoint
confidence float Confidence score

6. risk_qa — Risk-oriented Question Answering

List of risk-focused question-answer items (structure parallels qa).

7. summary — Grounded Summary

A single free-text grounded summary string for the patient. Unlike the per-item channels, the summary is a single generation and carries no copied-context redundancy.

Patient knowledge-graph visualization (visualization_payload.jsonl)

A separate file with one JSON object per patient (167,034 lines), encoding the patient-level knowledge graph derived from the verified entities and relations, in a node-link format suitable for graph visualization. Joined to the main corpus by meta.uid.

Top-level fields: nodes, links, groups, meta, meta_species.

  • meta — graph-level metadata: uid (join key), meta_species, node_count, entity_count, risk_count, med_count, link_count, graph_density, has_temporal, category_breakdown (counts per clinical category), summary_coverage.
  • nodes — clinical-entity nodes. Each node carries an id ({uid}_{index}), label, group (clinical category), level, display attributes (color, size, opacity, border_style), confidence, llm_confidence, assertion_status, is_negated, char_span (offsets into the source text), node_type, verifier_status, a temporal object (temporal_order, timepoint, timepoint_type, event_type), and in_summary.
  • links — edges corresponding to extracted relations, each with source and target node ids, label (relation type), status, confidence, verifier_votes, direction, head_type, tail_type, display attributes (width, dashed), and link_type.
  • groups — per-category summaries (count, avg_conf, color, level) used for layout and colouring.

This payload is a downstream rendering artefact derived from the verified extraction channels; it is included for completeness and is not part of the text-provenance analysis reported in the accompanying paper.

Downstream risk channels (separate files, joined by uid)

These are deterministic functions of the verified upstream evidence:

  • active_v6_risks.jsonl{uid, risks, provenance}: active clinical risks derived from verified entities and relations.
  • active_v6_risk_states.jsonl{uid, states, provenance, agentic_check}: temporal risk-state transitions (e.g. resolved, escalated).
  • active_v6_recommendations.jsonl{uid, recommendations, provenance}: recommendations, each traceable through a fixed policy to the specific verified evidence that triggered it.

Intended uses

  • Research on content provenance and redundancy in LLM-generated corpora.
  • Benchmarking de-duplication and corpus-reporting methods.
  • Studying multi-task clinical extraction output structure.
  • Method development in clinical NLP where the model-generated nature is explicitly accounted for.

Out-of-scope / prohibited uses

  • Any clinical or patient-care use. The corpus is not clinically validated.
  • Training systems presented as clinically reliable without independent validation.
  • Commercial use (prohibited by the CC-BY-NC-SA-4.0 license).

Limitations

  • Outputs are model-generated and contain errors; confidence and verifier fields indicate the pipeline's own assessment, not ground truth.
  • Coverage and phrasing reflect the generation pipeline's design; redundancy and task mixture are properties of that design (see the accompanying paper).
  • Derived from published case reports, which differ in distribution from real-world EHR notes.

Licensing and attribution

Released under CC-BY-NC-SA-4.0, inherited from the source dataset. You must attribute both this corpus and PMC-Patients, share derivatives under the same license, and use it non-commercially.

Dataset versions

Current version: v1.0

The dataset is versioned so that analyses can be reproduced against a fixed release. Future releases may contain corrections, additional metadata, or regenerated records and will receive a new version identifier.

Source PMC-Patients dataset:

@article{zhao2023large,
  title={A large-scale dataset of patient summaries for retrieval-based clinical decision support systems},
  author={Zhao, Zhengyun and Jin, Qiao and Chen, Fangyuan and Peng, Tuorui and Yu, Sheng},
  journal={Scientific Data},
  volume={10}, number={1}, pages={909}, year={2023},
  publisher={Nature Publishing Group UK London}
}

Citation

If you use this corpus, please cite:

@article{lazem2026prd,
  title={How much of an LLM-generated clinical corpus is actually new? A production-scale measurement of content redundancy for provenance classification},
  author={Lazem, Ali H. and Teahan, William J.},
  year={2026},
  note={arXiv:2606.29605}
}

Contact

  • Ali H. Lazemlhl23prg@bangor.ac.uk
    • School of Computer Science and Engineering, Bangor University, UK.
    • University of Thi-Qar, Nasiriyah, 64001, Iraq.
  • William J. Teahanw.j.teahan@bangor.ac.uk
    • School of Computer Science and Engineering, Bangor University, UK.
Downloads last month
78

Paper for lazem/prd-clinical-corpus