In the field of neonatal patient safety, the paucity of systematic research is a critical barrier to progress. Notably missing are studies that meticulously investigate Electronic Health Records (EHR) and information technology in detecting neonatal intensive care-related errors. The expert panel at the National Institute of Child Health and Human Development (NICHHD) identified multiple gaps in the current knowledge of neonatal patient safety research. The proposed work is a well focused response to three dimensions of the Funding Opportunity Announcement: 1.Develop prospective and retrospective study designs to collect data on patient safety and adverse events. 2.Study the strength and limitations of current methods of error reporting systems. 3.Study the usefulness of commercial IT systems and EHRs in reducing medical errors. In our study we seek to shift patient safety research toward an automated and computerized approach to achieve a more comprehensive patient safety paradigm. We will develop novel Electronic Health Record (EHR) content-based automated algorithms that are new to patient safety research to 1) detect errors (Aim 1) and 2) categorize subsequent harm (Aim 2). State of the art information extraction and statistical classification techniques from the field of clinical Natural Language Processing (NLP) will be adapted to the patient safety research tasks.
In Aim 1 we will fill the gap in the literatre by implementing a focused manual review of 700 charts (one full year of patient admissions at our institution) in one of the largest Neonatal Intensive Care Units (NICU) in the nation. Using a trigger tool, we will identify errors occurring in three specified categories - laboratory test errrs, medication/fluid errors, and airway management errors. We will develop novel algorithms for automated EHR-based detection of the errors and evaluate the performance of the new algorithms against the performance of both trigger tool review by human chart reviewers (current gold standard) and the voluntary incident reporting system (accepted standard).
In Aim 2, we will study the utility of novel EHR-based information extraction and statistical algorithms for the automated categorization of errors according to the resulting level of harm. Our proposed work has the potential to accomplish a paradigm shift in the methods of neonatal patient safety research and practice. The study is a fundamental step to automating patient safety monitoring on a large scale and improving error identification and patient safety in NICUs for millions of children every year.

Public Health Relevance

We are developing an automated error detection technique to improve the safety of newborn babies during hospital care. Our work is the first known attempt to use text analysis in the electronic health records on a large scale to reduce the cost while at the same time increase the speed and comprehensiveness of error detection in the clinical care of newborns.

National Institute of Health (NIH)
Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD)
Exploratory/Developmental Grants (R21)
Project #
Application #
Study Section
Health Services Organization and Delivery Study Section (HSOD)
Program Officer
Raju, Tonse N
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Cincinnati Children's Hospital Medical Center
United States
Zip Code
Lingren, Todd; Deleger, Louise; Molnar, Katalin et al. (2014) Evaluating the impact of pre-annotation on annotation speed and potential bias: natural language processing gold standard development for clinical named entity recognition in clinical trial announcements. J Am Med Inform Assoc 21:406-13
Li, Qi; Melton, Kristin; Lingren, Todd et al. (2014) Phenotyping for patient safety: algorithm development for electronic health record based automated adverse event and medical error detection in neonatal intensive care. J Am Med Inform Assoc 21:776-84
Zhai, Haijun; Brady, Patrick; Li, Qi et al. (2014) Developing and evaluating a machine learning based algorithm to predict the need of pediatric intensive care unit transfer for newly hospitalized children. Resuscitation 85:1065-71
Deleger, Louise; Lingren, Todd; Ni, Yizhao et al. (2014) Preparing an annotated gold standard corpus to share with extramural investigators for de-identification research. J Biomed Inform 50:173-83
Zhai, Haijun; Lingren, Todd; Deleger, Louise et al. (2013) Web 2.0-based crowdsourcing for high-quality gold standard development in clinical natural language processing. J Med Internet Res 15:e73
Li, Qi; Zhai, Haijun; Deleger, Louise et al. (2013) A sequence labeling approach to link medications and their attributes in clinical notes and clinical trial announcements for information extraction. J Am Med Inform Assoc 20:915-21
Li, Qi; Deleger, Louise; Lingren, Todd et al. (2013) Mining FDA drug labels for medical conditions. BMC Med Inform Decis Mak 13:53