scholarly journals Predicting 72-hour and 9-day return to the emergency department using machine learning

JAMIA Open ◽  
2019 ◽  
Vol 2 (3) ◽  
pp. 346-352 ◽  
Author(s):  
Woo Suk Hong ◽  
Adrian Daniel Haimovich ◽  
Richard Andrew Taylor
Keyword(s):  
Machine Learning ◽  
Emergency Department ◽  
Administrative Data ◽  
Trauma Center ◽  
Community Hospital ◽  
Gradient Boosting ◽  
Health Record ◽  
Test Sets ◽  
Level 1 ◽  
Level 1 Trauma Center

Abstract Objectives To predict 72-h and 9-day emergency department (ED) return by using gradient boosting on an expansive set of clinical variables from the electronic health record. Methods This retrospective study included all adult discharges from a level 1 trauma center ED and a community hospital ED covering the period of March 2013 to July 2017. A total of 1500 variables were extracted for each visit, and samples split randomly into training, validation, and test sets (80%, 10%, and 10%). Gradient boosting models were fit on 3 selections of the data: administrative data (demographics, prior hospital usage, and comorbidity categories), data available at triage, and the full set of data available at discharge. A logistic regression (LR) model built on administrative data was used for baseline comparison. Finally, the top 20 most informative variables identified from the full gradient boosting models were used to build a reduced model for each outcome. Results A total of 330 631 discharges were available for analysis, with 29 058 discharges (8.8%) resulting in 72-h return and 52 748 discharges (16.0%) resulting in 9-day return to either ED. LR models using administrative data yielded test AUCs of 0.69 (95% confidence interval [CI] 0.68–0.70) and 0.71(95% CI 0.70–0.72), while gradient boosting models using administrative data yielded test AUCs of 0.73 (95% CI 0.72–0.74) and 0.74 (95% CI 0.73–0.74) for 72-h and 9-day return, respectively. Gradient boosting models using variables available at triage yielded test AUCs of 0.75 (95% CI 0.74–0.76) and 0.75 (95% CI 0.74–0.75), while those using the full set of variables yielded test AUCs of 0.76 (95% CI 0.75–0.77) and 0.75 (95% CI 0.75–0.76). Reduced models using the top 20 variables yielded test AUCs of 0.73 (95% CI 0.71–0.74) and 0.73 (95% CI 0.72–0.74). Discussion and Conclusion Gradient boosting models leveraging clinical data are superior to LR models built on administrative data at predicting 72-h and 9-day returns.

scholarly journals (P2-74) Does Implementation of Protocols Improve the Red Area Disposition Time of Emergency Department at a Level-1 Trauma Center?

2011 ◽  
Vol 26 (S1) ◽  
pp. s160-s160
Author(s):  
R. Kumar ◽  
K. Shyamla ◽  
S. Bhoi ◽  
T.P. Sinha ◽  
S. Chauhan ◽  
...  
Keyword(s):  
Intensive Care Unit ◽  
Emergency Department ◽  
Trauma Center ◽  
Safety Issues ◽  
Ed Visits ◽  
Level 1 ◽  
The Impact ◽  
Level 1 Trauma Center ◽  
Final Disposition ◽  
Disposition Time

BackgroundAcute care addresses immediate resuscitation and early disposition to definitive care. Delay in final disposition from the emergency department (ED) affects outcomes in terms of morbidity and mortality. An audit was performed to assess the impact of protocols on red area disposition time.MethodsAn audit of red (resuscitation) area disposition time was performed among patients with compromised airway, breathing, and circulation. The red area disposition time was defined as the time from ED arrival to red area disposition. Pre-protocol data from nursing report books were reviewed for ED to operating room (OR), ED to intensive care unit (ICU), and overall disposition time between September 2007 and January 2008. Similar outcomes were documented after implementation of protocols during February to December 2008.ResultsIn the pre-protocol period, 992 red area patients were enrolled out of 10,000 ED visits. Out of which 527 (53.1%) were shifted to the OR and 222 (22.3%) to ICU. The average ED disposition time was 3.5 hours (range 2–5). Similarly, 1797 red area patients were enrolled in the post-protocol period out of 25,928. Of these, 453 (25.2%) patients were shifted to the OR, and 423 (23.7%) were shifted to the ICU. The average ED disposition time was 1.5 hours (range 10 minutes–3 hours).ConclusionsImplementation of protocols improves the red area disposition time of the ED. Auditing is an important tool to address patient safety issues.

scholarly journals (A138) Duration of Emergency Department Stay and Injury Profile of Homeless Patients in a Level-1 Trauma Center

2011 ◽  
Vol 26 (S1) ◽  
pp. s39-s39
Author(s):  
S. Chauhan ◽  
S. Bhoi ◽  
D.T. Sinha ◽  
M. Rodha ◽  
L. Alexender ◽  
...  
Keyword(s):  
Emergency Department ◽  
Head Injury ◽  
Trauma Center ◽  
Injury Severity ◽  
Road Traffic ◽  
Road Traffic Crash ◽  
Emergency Department Stay ◽  
Level 1 ◽  
Gcs Score ◽  
Level 1 Trauma Center

Background and ObjectiveImmediate resuscitation and early disposition to definitive care improves outcomes. Homeless patients are neglected in emergency department (ED). The duration of ED stay and profile of injury of homeless patients at a Level-1 Trauma center were measured.MethodsThe study was performed from October 2008 to September 2009. Homeless patients were defined as patients who had no attendant and did not have any shelter. Duration of ED stay was noted from the ED arrival time to entry time at the definitive care (intensive care unit/ward). Clinical and demographic details were recorded. Subjects who had: (1) an attendant; (2) were discharged from the ED; or (3) expired in the ED were excluded.ResultsForty-one homeless patients were admitted. The mode of injury was road traffic crash in 73.2%; assault in 7.3%; fall from height in 7.3%; and in 12.2%, the mode of injury unknown. The average Injury Severity Score (ISS) was 6.76, with a maximum of 34 and minimum of 1. A total of 24 subjects (59%) had a Glasgow Coma Scale (GCS) score of ≤ 8 (severe head injury), 10 patients (24%) had GCS score 9–12 (moderate head injury), and seven subjects (17%) had GCS score 13–15 (minor head injury). Breath alcohol test was positive in 13%. The average duration of ED stay was 35 (3–173) hours in the homeless group and 12 (0.5–18) hours for patients with an attendant. Twenty-one subjects were admitted to neurosurgery (51.2%) with an average ED stay of 22.4 hours, five to surgery (12.20%) with average ED stay of 56.6 hours, and 15 to orthopedics (36.6%) with average ED stay of 45.3 hours.ConclusionsThe emergency department stay of homeless patients was 35 hours. Orthopedic trauma subjects had a prolonged disposal time. This addresses serious patient safety concerns and immediate remedial measures.

scholarly journals (P2-65) Perception of Emergency Care Providers Toward the Implementation of an Electronic Medical Record System in the Emergency Department of a Level-1 Trauma Center

2011 ◽  
Vol 26 (S1) ◽  
pp. s156-s157
Author(s):  
S. Chauhan ◽  
S. Bhoi ◽  
D.T. Sinha ◽  
M. Rodha ◽  
G. Adhikari ◽  
...  
Keyword(s):  
Emergency Department ◽  
Medical Record ◽  
Electronic Medical Record ◽  
Emergency Care ◽  
Trauma Center ◽  
Positive Reinforcement ◽  
Care Providers ◽  
Record System ◽  
Level 1 ◽  
Level 1 Trauma Center

BackgroundManual documentation has an inherent problem of improper communication, manipulation, and validity. An electronic medical record (EMR) is a computerized medical record created in an organization that delivers care, such as a hospital. EMRs tend to be a part of a local, stand-alone, health information system that allows for storage and retrieval.ObjectivesThe objective of this study was to assess the perception of emergency care providers toward the implementation of an EMR System in the emergency department of a Level-1 Trauma Center.MethodsA qualitative survey was conducted among consenting doctors and nurses in the emergency department of the All India Institute of Medical Sciences February to October 2010. Data were collected from a sample of 22—eight doctors and 14 nurses. The collection tool was a structured, closed-ended questionnaire of 12 questions based on usability, applicability, and security, of EMR. A Likert scale (LS) was used (1 = worst, 4 = best). Surveys were done on Day 20, Day 45, and after nine months of implementation of. Responses of emergency care providers were compiled and analyzed using SPSS version 16.ResultsThree surveys consisted of 22 participants in each survey. The survey domain of usability improved on Survey 3 (LS = 2.57), Survey 2 (LS = 2.46), Survey 1 (LS = 2.24). Application of EMR improved from Survey 1 to Survey 3. The data regarding perception of security concerns such as manipulation of data, transparency, and accountability were comparable among Survey 1, Survey 2, and Survey 3. Initial satisfaction was strongly associated with perception of usefulness of data mining for research purposes.ConclusionsSatisfaction with an EMR system at its implementation generally persisted through the first year of use. Implementation plans must include positive reinforcement regarding EMR among emergency care providers.

scholarly journals Level 1 Trauma Centers and OEF/OIF Emergency Departments: Comparison of Trauma Patient Populations

2020 ◽  
Vol 185 (9-10) ◽  
pp. e1569-e1575
Author(s):  
Shelia C Savell ◽  
Alexis Blessing ◽  
Nicole M Shults ◽  
Alejandra G Mora ◽  
Kimberly L Medellin ◽  
...  
Keyword(s):  
Emergency Department ◽  
Trauma Patient ◽  
Emergency Departments ◽  
Trauma Center ◽  
San Antonio ◽  
Trauma Patients ◽  
Trauma Centers ◽  
Level 1 ◽  
Hospital Days ◽  
Level 1 Trauma Center

Abstract Introduction Brooke Army Medical Center (BAMC), the largest military hospital and the only level 1 trauma center in the DoD, cares for active duty, retired uniformed services personnel, and beneficiaries. In addition, BAMC works in collaboration with the Southwest Texas Regional Advisory Council (STRAC) and University Hospital (UH), San Antonio’s other level 1 trauma center, to provide trauma care to residents of the city and 22 counties in southwest Texas from San Antonio to Mexico (26,000 square mile area). Civilian-military partnerships are shown to benefit the training of military medical personnel; however, to date, there are no published reports specific to military personnel experiences within emergency care. The purpose of the current study was to describe and compare the emergency department trauma patient populations of two level 1 trauma centers in one metropolitan city (BAMC and UH) as well as determine if DoD level 1 trauma cases were representative of patients treated in OEF/OIF emergency department settings. Materials and Methods We obtained a nonhuman subjects research determination for de-identified data from the US Air Force 59th Medical Wing and the University of Texas Health Science Center at San Antonio Institutional Review Boards. Data on emergency department patients treated between the years 2015 and 2017 were obtained from the two level 1 trauma centers (BAMC and UH, located in San Antonio, Texas); data included injury descriptors, ICU and hospital days, and department procedures. Results Two-proportion Z-tests indicated that trauma patients were similar across trauma centers on injury type, injury severity, and discharge status; yet trauma patients differed significantly in terms of mechanism of injury and regions of injury. BAMC received significantly greater proportions of patients injured from falls, firearms and with facial and head injuries than UH, which received significantly greater proportion of patients with thorax and abdominal injuries. In addition, a significantly greater proportion of patients spent more than 2 days in the ICU and greater than two total hospital days at BAMC than in UH. In comparison to military emergency departments in combat zones, BAMC had significantly lower rates of blood product administration and endotracheal intubations. Conclusions The trauma patients treated at a military level 1 trauma center were similar to those treated in the civilian level 1 trauma center in the same city, indicating the effectiveness of the only DoD Level 1 trauma center to provide experience comparable to that provided in civilian trauma centers. However, further research is needed to determine if the exposure rates to specific procedures are adequate to meet predeployment readiness requirements.

Sign in / Sign up

Export Citation Format

Share Document