Identifying Patients with Depression Using Free-text Clinical Documents

被引:42
作者
Zhou, Li [1 ,2 ]
Baughman, Amy W. [2 ]
Lei, Victor J. [1 ]
Lai, Kenneth H. [2 ]
Navathe, Amol S. [3 ]
Chang, Frank [1 ]
Sordo, Margarita [1 ,2 ]
Topaz, Maxim [2 ]
Zhong, Feiran [2 ]
Murrali, Madhavan [4 ]
Navathe, Shamkant [4 ]
Rocha, Roberto A. [1 ,2 ]
机构
[1] Partners Healthcare Inc, Partners eCare, Clin Informat, Boston, MA USA
[2] Harvard Med Sch, Brigham & Womens Hosp, Div Gen Internal Med & Primary Care, Boston, MA USA
[3] Univ Penn, Perelman Sch Med, Philadelphia, PA 19104 USA
[4] Georgia Inst Technol, Coll Comp, Sch Comp Sci, Atlanta, GA 30332 USA
来源
MEDINFO 2015: EHEALTH-ENABLED HEALTH | 2015年 / 216卷
关键词
Depression; Natural Language Processing; Text Classification; Machine Learning; CARE; DISPARITIES; HEALTH;
D O I
10.3233/978-1-61499-564-7-629
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
About 1 in 10 adults are reported to exhibit clinical depression and the associated personal, societal, and economic costs are significant. In this study, we applied the MTERMS NLP system and machine learning classification algorithms to identify patients with depression using discharge summaries. Domain experts reviewed both the training and test cases, and classified these cases as depression with a high, intermediate, and low confidence. For depression cases with high confidence, all of the algorithms we tested performed similarly, with MTERMS' knowledge-based decision tree slightly better than the machine learning classifiers, achieving an F-measure of 89.6%. MTERMS also achieved the highest F-measure (70.6%) on intermediate confidence cases. The RIPPER rule learner was the best performing machine learning method, with an F-measure of 70.0%, and a higher precision but lower recall than MTERMS. The proposed NLP-based approach was able to identify a significant portion of the depression cases (about 20%) that were not on the coded diagnosis list.
引用
收藏
页码:629 / 633
页数:5
相关论文
共 24 条
[1]   The epidemiology of major depressive episodes:: results from the International Consortium of Psychiatric Epidemiology (ICPE) Surveys [J].
Andrade, L ;
Caraveo-Anduaga, JJ ;
Berglund, P ;
Bijl, RV ;
De Graaf, R ;
Vollebergh, W ;
Dragomirecka, E ;
Kohn, R ;
Keller, M ;
Kessler, RC ;
Kawakami, N ;
Kiliç, C ;
Offord, D ;
Ustun, TB ;
Wittchen, HU .
INTERNATIONAL JOURNAL OF METHODS IN PSYCHIATRIC RESEARCH, 2003, 12 (01) :3-21
[2]  
[Anonymous], 2014, C4. 5: programs for machine learning
[3]  
[Anonymous], 2010, MMWR, V59, P1229
[4]  
[Anonymous], 2000, DIAGN STAT MAN MENT, DOI DOI 10.1176/APPI.BOOKS.9780890425787
[5]  
Campbell KP, PURCHASERS GUIDE CLI
[6]   Dose-response relationship between depressive symptoms and hospital readmission [J].
Cancino, Ramon S. ;
Culpepper, Larry ;
Sadikova, Ekaterina ;
Martin, Jessica ;
Jack, Brian W. ;
Mitchell, Suzanne E. .
JOURNAL OF HOSPITAL MEDICINE, 2014, 9 (06) :358-364
[7]  
Cohen WW, 1995, MACHINE LEARNING
[8]   Reducing the societal burden of depression - A review of economic costs, quality of care and effects of treatment [J].
Donohue, Julie M. ;
Pincus, Harold Alan .
PHARMACOECONOMICS, 2007, 25 (01) :7-24
[9]   Clinician screening and treatment of alcohol, drug, and mental problems in primary care -: Results from healthcare for communities [J].
Edlund, MJ ;
Unützer, A ;
Wells, KB .
MEDICAL CARE, 2004, 42 (12) :1158-1166
[10]  
Fischer LR., 2008, Clin Med Res, V6, P125