modeling human comprehension
play

Modeling human comprehension of Swedish medical records for - PowerPoint PPT Presentation

Modeling human comprehension of Swedish medical records for intelligent access and summarization systems Future vision, a physicians perspective Karolinska Universitetssjukhuset Maria Kvist Maria Skeppstedt Inst fr data- och


  1. Modeling human comprehension of Swedish medical records for intelligent access and summarization systems – Future vision, a physician’s perspective Karolinska Universitetssjukhuset Maria Kvist Maria Skeppstedt Inst för data- och Sumithra Velupillai systemvetenskap(DSV), Stockholms universitet Hercules Dalianis

  2. Clinical documentation Mia Kvist

  3. Mia Kvist

  4. Clinical documentation Mia Kvist

  5. We need: Overview Search tool Text summarization Knowledge extraction Mia Kvist

  6. Din journal på nätet electronic access on Internet to my own patient record Mia Kvist

  7. Din journal på nätet electronic access on Internet to my own patient record Patients need: Overview Search tool Summarization tool Translation tool Mia Kvist

  8. We need: Overview Search tool Text summarization Knowledge extraction Patients need: Overview Search tool Summarization tool Translation tool Mia Kvist

  9. Talented group of people Inst. för data- och systemvetenskap Hercules Dalianis Martin Hassel Sumithra Velupillai Maria Skeppstedt Aron Henriksson Gunnar Nilsson, KI Mia Kvist

  10. Text summarization • Domain specific language • How do physicians read, comprehend, summarize and reach an understanding for a patients situation and needs? Mia Kvist

  11. Text summarization • Domain specific language Need better understanding • How do physicians read, comprehend, summarize and reach an understanding for a patients situation and needs? Need models to spot what is important Need grading of the certainty of knowledge Mia Kvist

  12. Stockholm EPR Corpus • Karolinska University Hospital • 2 milj deidentified patient records • 2006-2010 • Various medical, surgical specialities + childrens hospital • No psychiatry Mia Kvist

  13. Finding important facts • Entity recognition Patient’s twin probably had a myocardial infarction two years ago Mia Kvist

  14. Finding important facts • Entity recognition • Certainty Patient’s twin probably had a myocardial infarction two years ago Mia Kvist

  15. Finding important facts • Entity recognition • Certainty • Temporality Patient’s twin probably had a myocardial infarction two years ago Mia Kvist

  16. Finding important facts • Entity recognition • Certainty • Temporality • Subject identification Patient’s twin probably had a myocardial infarction two years ago Mia Kvist

  17. Finding important facts • Entity recognition • Certainty • Temporality • Subject identification Patient’s twin probably had a myocardial infarction two years ago P’s twin prob had MI 2y ago Mia Kvist

  18. Electronic patient records - language • Unstructured free text incompleate sentences, few subjects, passive verbs • Medical terminology • Medical jargong • Abbreviations and acronyms • Latin, Greek and English Mia Kvist

  19. MIE 2011: Using SNOMED CT for High Precision Entity Recognition in Swedish Clinical Text Maria Skeppstedt Hercules Dalianis Mia Kvist

  20. Entity recognition - Clinical findings Spot all clinical findings in free text. Traditional way: Match to list of diagnoses, symptoms. (ICD-10, SNOMED CT etc) But do we express diagnoses in that way in free text? Mia Kvist

  21. Entity recognition - Implicit disorders Dialysbehandlad kärlsjuk diabetiker med huvudvärksproblematik. Dialysistreated vascular disease-sick diabetic with headache problems. Mia Kvist

  22. Entity recognition - annotation of clinical findings Methods Annotation Machine learning systems • Train on the annotated data • Automatically detect symptoms and disorders through contextual or other markers Mia Kvist

  23. Entity recognition - annotation of clinical findings Qualitative results The same clinical expression can be a finding or a disorder depending on context. Ex) tachycardia Some clinical findings are not recognized as the same. Different expressions, abbreviations, misspellings. Ex) Troponin-T normal Mia Kvist

  24. Swedish is rich in compound words Chestpain symptom Pain in the chest symptom and body part Lungclinic a place or institution, no body part Mia Kvist

  25. MIE 2011: Factuality Levels of Diagnoses in Swedish Clinical Text Sumithra Velupillai (presenter) Hercules Dalianis Maria Kvist

  26. Certainty of diagnosis - annotation Machine learning process Model subtelties expressed in natural language Learn to determine the certainty of a diagnosis by • cue words • patterns inherent in the diagnosis expression Mia Kvist

  27. - + Certainly Probably Possibly Possibly Probably Certainly Positive Positive Positive Negative Negative Negative Patient has Parkinsons disease. Physical examination strongly suggests Parkinson. Patient possibly has Parkinson. Parkinson cannot yet be outruled. No support for Parkinson. Parkinsson can be excluded. Mia Kvist

  28. Certainly Probably Possibly Possibly Probably Certainly Positive Positive Positive Negative Negative Negative 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Annotation classes for 15 diagnoses 28

  29. Certainty of diagnosis - annotation Qualitative results There is more to it than just cue phrases. • Overt findings - high certainty • Diagnoses that lack negative classes • Some diagnoses do not need to be certain • Comlementary diagnoses Mia Kvist

  30. 50 ischemia heart attack angina pectoris 40 Annotations 30 20 10 0 Certainly Probably Possibly Probably Possibly Certainly Pos Pos Pos Neg Neg Neg

  31. Conclusions It may be possible to, bit by bit, put together system that can • find important facts • determine their certainty • summarize Many obstacles are yet to be overcome Mia Kvist

Recommend


More recommend