Development and Validation of a Natural Language Processing Tool to Identify Injuries in Infants Associated With Abuse |
| |
| Institution: | 1. Yale University School of Medicine (G Tiyyagura, AG Asnes, JM Leventhal, ED Shapiro, M Auerbach, W Teng, E Powers, A Thomas, AL Hsiao), New Haven, CT;2. University of Colorado Anschutz Medical Campus (DM Lindberg), Aurora Colo;3. 3M | M*Modal Health Information Systems (J McClelland, C Kutryb, T Polzin, K Daughtridge, V Sevin), Pittsburg, Pa;1. Department of Pediatrics (GG Guyol, MJ Corwin, LA Smith, and MG Parker), Boston Medical Center and Boston University School of Medicine, Boston, Mass;2. Department of Pediatrics, Division of Newborn Medicine (GG Guyol), Boston Children''s Hospital and Harvard Medical School, Boston, Mass;3. Slone Epidemiology Center (SM Kerr, MJ Corwin), Boston University, Boston, Mass;4. Department of Pediatrics (E Colson), Yale University, New Haven, Conn;5. Department of Pediatrics (E Colson), School of Medicine, Washington University in St. Louis, St. Louis, Mo;6. CDC Foundation (LA Smith), Atlanta, Ga;7. Department of Biostatistics (T Heeren), Boston University School of Public Health, Boston, Mass;8. Department of Health, Behavior & Society (MT Kiviniemi), College of Public Health, University of Kentucky, Lexington, Ky;1. Department of Student Affairs, Wake Forest School of Medicine (IL Tablazon and FC O''Brian), Winston-Salem, NC;2. Departments of Internal Medicine, Pediatrics and Epidemiology and Prevention, Wake Forest School of Medicine (D Palakshappa), Winston-Salem, NC;3. Departments of Pediatrics, Downtown Health Plaza Pediatrics Clinic, Wake Forest School of Medicine (B Ramirez), Winston-Salem, NC;4. Departments of Pediatrics and Epidemiology and Prevention, Wake Forest School of Medicine (JA Skelton), Winston-Salem, NC;5. Department of Pediatrics, Wake Forest School of Medicine (LW Albertini and KG Montez), Winston-Salem, NC;1. Division of Hospital Medicine (AM Jenkins, K Auger), Cincinnati Children''s Hospital Medical Center, Cincinnati, OH;2. Department of Internal Medicine (AM Jenkins), University of Cincinnati Medical Center, Cincinnati, OH;3. Division of General Pediatrics (JG Berry, P Dunbar, B Garrity), Boston Children''s Hospital, Harvard Medical School, Boston, Mass;4. Division of General Academic Pediatrics, Department of Pediatrics (JM Perrin, K Kuhlthau), MassGeneral Hospital for Children, Harvard Medical School, Boston, Mass;5. Children''s Hospital Association (M Hall), Lenexa, KS;6. Family Voices (C Hoover), Albuquerque, NM;7. Building Bright Futures (M Crossman), Williston, VT;8. Department of Pediatrics, University of Cincinnati College of Medicine (K Auger), Cincinnati, OH;9. James M. Anderson Center for Health Systems Excellence (K Auger), Cincinnati Children''s Hospital Medical Center, Cincinnati, OH;1. Department of Pediatrics, University of Washington School of Medicine and Seattle Children''s Hospital (EK Chung);2. Department of Health Services, University of Washington (I Painter and VD Souter), Seattle, Wash;3. The Foundation for Health Care Quality (I Painter, K Sitcov, and VD Souter), Seattle, Wash;1. Department of Pediatrics (R Linderer, N Srinivasan, and A Schwartz), University of Illinois at Chicago, Ill;2. Department of Medical Education (C Park, A Schwartz, and R Yudkowsky), University of Illinois at Chicago, Ill;3. Department of Anesthesiology (C Park), University of Illinois at Chicago, Ill;4. Simulation and Integrative Learning Institute (C Park), University of Illinois at Chicago, Ill |
| |
| Abstract: | ObjectivesMedically minor but clinically important findings associated with physical child abuse, such as bruises in pre-mobile infants, may be identified by frontline clinicians yet the association of these injuries with child abuse is often not recognized, potentially allowing the abuse to continue and even to escalate. An accurate natural language processing (NLP) algorithm to identify high-risk injuries in electronic health record notes could improve detection and awareness of abuse. The objectives were to: 1) develop an NLP algorithm that accurately identifies injuries in infants associated with abuse and 2) determine the accuracy of this algorithm.MethodsAn NLP algorithm was designed to identify ten specific injuries known to be associated with physical abuse in infants. Iterative cycles of review identified inaccurate triggers, and coding of the algorithm was adjusted. The optimized NLP algorithm was applied to emergency department (ED) providers’ notes on 1344 consecutive sample of infants seen in 9 EDs over 3.5 months. Results were compared with review of the same notes conducted by a trained reviewer blind to the NLP results with discrepancies adjudicated by a child abuse expert.ResultsAmong the 1344 encounters, 41 (3.1%) had one of the high-risk injuries. The NLP algorithm had a sensitivity and specificity of 92.7% (95% confidence interval CI]: 79.0%–98.1%) and 98.1% (95% CI: 97.1%–98.7%), respectively, and positive and negative predictive values were 60.3% and 99.8%, respectively, for identifying high-risk injuries.ConclusionsAn NLP algorithm to identify infants with high-risk injuries in EDs has good accuracy and may be useful to aid clinicians in the identification of infants with injuries associated with child abuse. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
