A joint project of the Graduate School, Peabody College, and the Jean & Alexander Heard Library

Title page for ETD etd-06012015-140322


Type of Document Master's Thesis
Author Temple, Michael William
Author's Email Address mtemple1@me.com
URN etd-06012015-140322
Title Using Daily Progress Note Data to Predict Discharge Date from the Neonatal Intensive Care Unit
Degree Master of Science
Department Biomedical Informatics
Advisory Committee
Advisor Name Title
Christoph U. Lehmann, M.D. Committee Chair
Daniel Fabbri, Ph.D. Committee Member
Kevin B. Johnson, M.D., M.S. Committee Member
William Gregg, M.D., M.S., M.P.H. Committee Member
Keywords
  • Intensive Care Units
  • Area Under Curve
  • Patient Discharge
  • Natural Language Processing
  • Random Forest
Date of Defense 2015-05-11
Availability unrestricted
Abstract
BIOMEDICAL INFORMATICS

Using Daily Progress Note Data to Predict Discharge Date from the Neonatal Intensive Care Unit

Michael William Temple

Thesis under the direction of Professor Christoph U. Lehmann

Discharging patients from the Neonatal Intensive Care Unit (NICU) may be delayed for non-medical reasons including the procurement of home medical equipment, parental education, and the need for children’s services. We describe a method to predict and identify patients that will be medically ready for discharge in the next 2-10 days – providing lead-time to address non-medical reasons for delayed discharge.

The study examined 26 clinical features (17 extracted, 9 engineered) from daily neonatology progress notes of 4,693 patients (103,206 patient-days) from the NICU at Vanderbilt University. A Random Forest model achieved an AUC ranging from 0.854-0.865 at 2 days-to-discharge (DTD) and 0.723-0.729 at 10 DTD. This model was also able to identify important features allowing for simplified model construction.

The second part of the study utilized Natural Language Processing (NLP) to analyze the Assessment and Plan section of the daily progress note in order to identify possible causes for patients who performed poorly in our original model and reasons that may have caused some patients to have delays in discharge. This analysis revealed that some patients with surgical diagnoses, pulmonary hypertension, retinopathy of prematurity and psychosocial issues did not perform well in our original model. Additionally, NLP identified that social issues including patients born to drug-addicted mothers and the involvement of children’s services may be a non-medical reasons causing delays in discharge.

Files
  Filename       Size       Approximate Download Time (Hours:Minutes:Seconds) 
 
 28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access 
  Temple.pdf 2.45 Mb 00:11:21 00:05:50 00:05:06 00:02:33 00:00:13

Browse All Available ETDs by ( Author | Department )

If you have more questions or technical problems, please Contact LITS.