Negation scope and spelling variation for text-mining of Danish electronic patient records

What it is about

The paper describes a simple ICD-10 dictionary matcher for Danish psychiatric patient records, 5,543 records in all. Five add-on functions are tested: spelling variants, ending variants, gaps between words, shuffled words, and negation. The negation handling sets a scope window for each negation cue, and the scope is checked against a manually curated test set of 50 patients with 5,765 disease hits. The full tool reaches an incidence precision of 0.867 and an association precision of 0.888. Gaps and shuffles matter only for the roughly 12% of hits with more than one token. Different negation words have different useful scopes.

Ideas

• I1. en/src/expert.rs and the prag negation witness: give each English negation cue (not, no, never, nothing) its own scope window in tokens, and measure the precision of each cue on the English test set. Set the windows on a dev split, not on test.
• I2. en/src/morph.rs: a spelling and ending variant lookup for matching historical forms (thou, thee, hast, doth) to modern lemmas, reported with and without the lookup as the paper does. Measure precision on tools/gutenberg lessons.
• I3. Ablation copy: make a copy of the matcher without the negation step, run the same test, then remove the copy. The gain that negation gives should be stated with its date and the copy's result.

For Mova

The chapter is 21. The simple clause, which covers how negation sits in the clause. The code is en/src/expert.rs and the prag crate. The paper's data is Danish health records under NC terms and it must stay out of any build; only the method is used, on English data.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.3115/v1/w14-1109

@inproceedings{thomas2014negation,
  title = {{Negation scope and spelling variation for text-mining of Danish electronic patient records}},
  author = {Thomas, Cecilia Engel and Jensen, Peter Bjødstrup and Werge, Thomas and Brunak, Søren},
  booktitle = {Proceedings of the 5th International Workshop on Health Text Mining and Information Analysis (Louhi)},
  year = {2014},
  doi = {10.3115/v1/w14-1109},
  url = {https://doi.org/10.3115/v1/w14-1109}
}

References

13 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Agarwal, Shashank and Yu, Hong (2010) Biomedical negation scope detection with conditional random fields, J Am Med Inform Assoc., 17:696-701. find
2. Chapman, Wendy W., Bridewell, Will, Hanbury, Paul, Cooper, Gregory F., and Buchanan, Bruce G. (2001) A Simple Algorithm for Identifying Negated Findings and Diseases in Discharge Summaries, Journal of Biomedical Informatics 34, 301–310. find
3. Eriksson, Robert, Werge, Thomas, Jensen, Lars J., and Brunak, Søren (to appear 2014) Dose-specific adverse drug reaction identification in electronic patient records: temporal data mining in an inpatient psychiatric population. find
4. Friedman, Carol, Shagina, Lyudmila, Lussier, Yves, and Hripcsak, George (2004) Automated encoding of clinical documents based on natural language processing, J Am Med Inform Assoc., 11(5):392-402. find
5. Jensen, Peter B., Jensen, Lars J., and Brunak, Søren (2012) Mining electronic health records: towards better research applications and clinical care, Nature Rev. Genetics 13(6):395-405. find
6. Liao, Katherine P., Cai, Tianxi, Gainer, Vivian, Goryachev, Sergey, Zeng-Treitler, Qing, Raychaudhuri, Soumya, Szolovits, Peter, Churchill, Susanne, Murphy, Shawn, Kohane, Isaac, Karlson, Elizabeth W., and Plenge, Robert M. (2010) Electronic medical records for discovery research in rheumatoid arthritis, Arthritis Care Res (Hoboken) 62;1120-1127. find
7. Meystre, Stephane and Haug, Peter J. (2006) Natural language processing to extract medical problems from electronic clinical documents: performance evaluation, J Biomed Inform 39;589-599. find
8. Mutalik, Pradeep G., Deshpande, Aniruddha, and Nadkarni, Prakash M. (2001) Use of General-purpose Negation Detection to Augment Concept Indexing of Medical Documents: A Quantitative Study Using the UMLS, J Am Med Inform Assoc. 8:598–609. find
9. Park, Juyong, Lee, Deok-Sun, Christakis, Nicholas A., and Barabási, Albert-László (2009) The impact of cellular networks on disease comorbidity, Molecular Systems Biology 5:262. find
10. Prather, J.C., Lobach, D.F., Goodwin, L.K., Hales, J.W., Hage, M.L., and Hammond, W.E. (1997) Medical data mining: knowledge discovery in a clinical warehouse, Proc AMIA Annu Fall Symp; 101-105. find
11. Roque, Francisco S. Jensen, Peter B., Schmock, Henriette, Dalgaard, Marlene, Andreatta, Massimo, Hansen, Thomas, Søeby, Karen, Bredkjær, Søren, Juul, Anders, Werge, Thomas, Jensen, Lars J., and Brunak, Søren (2011) Using electronic patient records to discover disease correlations and stratify patient cohorts, PLOS computational biology. find
12. Suzuki, T., Yakoi, H., Fujita, S., and Takabayashi, K. (2008) Automatic DPC code selection from electronic medical records: text mining trial of discharge summary, Methods Inf Med 47;541-548. find
13. Tremblay, Monica C., Berndt, Donald J., Luther, Stephen L., Foulis, Philip R., and French, Dustin D. (2009) Identifying fall-related injuries: Text mining the electronic medical record, Inf Technol Manag 10;253-26. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/aclbulk-2014-thomas-negation-scope-spelling-variation-text/