Facebook Twitter YouTube



Comparison of Clinician Suspicion Versus Prediction Rule In Detecting Children At Low Risk for Intraabdominal Injury After Blunt Torso Trauma

Friday, October 19, 2012: 2:15 PM
Room 278-280 (Morial Convention Center)
Prashant Mahajan, MD, Pediatrics, Children's Hospital of Michigan, Detroit, MI, M. Tunik, MD, Pediatrics and Emergency Medicine, NYU Langone School of Medicine / Bellevue Hospital Center, New York, NY, Kenneth Yen, M.D, Department of Pediatrics, Section of Emergency Medicine, Medical College of Wisconsin, Milwaukee, WI, Shireen M. Atabaki, MD, Division of Pediatric Emergency Medicine, Children's National Medical Center, Washington, DC, Lois K. Lee, M.D, Emergency Medicine, Children's Hospital Boston, Boston, MA, Angela Ellison, MD, Children's Hospital of Philadelphia, Bema K. Bonsu, MD, Division of Emergency Medicine, Children's Hospital, Columbus, Cody S. Olsen, MS, Department of Pediatrics, University of Utah School of Medicine, Salt Lake City, UT, Maria Y. Kwok, MD, Pediatrics, Columbia University Medical Center, New York, NY, Kathleen Lillis, MD, Pediatric Emergency Medicine, Women and Children's Hospital of Buffalo, Buffalo, NY, Nathan Kuppermann, MD, MPH, FAAP, Emergency Medicine, University of California, Davis, Sacramento, CA, James F. Holmes, MD, Department of Emergency Medicine, University of California, Davis, Sacramento, CA and Pediatric Emergency Care Applied Research Network, IAI Study Group, Sacramento, CA

Purpose: Identification of children with intra-abdominal injuries undergoing acute intervention (IAIAI) is challenging and it is unclear if a clinical prediction rule is superior to clinician judgment. The main objective of this study was to compare the test characteristics of clinician suspicion to detect children with IAI undergoing acute intervention (IAIAI) with those of a clinical prediction rule for IAIAI among children with blunt torso trauma.

Methods: We performed a prospective, multicenter observational study of children (<18 years old) with blunt torso trauma. Clinicians documented patient history, physical findings and suspicion for the presence of IAIAI (defined as a child with an IAI undergoing therapeutic laparotomy, angiographic embolization, blood transfusion for abdominal hemorrhage, or IV fluid administration for ≥ 2 days in those with pancreatic/duodenal injuries). Suspicion for IAIAI was documented in risk increments as <1%, 1-5%, 6-10%, 11-50%, and >50%. Patients were considered to be positive by clinician suspicion if suspicion was ≥ 1%. We compared the performance of clinicians suspicions to predict IAIAI with the performance of a derived prediction rule (which was derived with a misclassification cost of type II error set at 500:1).

Results: 12,044 eligible patients were enrolled. Clinicians recorded suspicion of IAIAI in 11,919 (99%) patients while IAIAI was diagnosed in 203 (2%). Clinicians had a suspicion ≥1% for IAIAI in 2,667 (22%) of patients while 168 (6%, 95% CI 5-7%) had IAIAI. Among patients with < 1% clinician suspicion for IAIAI, 35 (0.4%, 95% CI 0.3-0.5%) had IAIAI. Abdominal CT scans were obtained in the ED for 2,302 or 86% of patients (86%, 95% CI 85-88%) with clinician suspicion ≥1% for IAIAI and in 3,016 patients or 33% (33%, 95% CI 32-34%) with clinician suspicion < 1% for IAIAI.

 

IAIAI

Prediction rule

Clinician suspicion≥1%

Sensitivity (95% CI)

97.0 (93.7, 98.6)

82.8 (77.0, 87.3)

Specificity (95% CI)

42.5 (41.6, 43.4)

78.7 (77.9, 79.4)

NPV (95% CI)

99.9 (99.7, 99.9)

99.6 (99.5, 99.7)

PPV (95% CI)

2.8 (2.5, 3.3)

6.3 (5.4, 7.3)

LR+ (95% CI)

1.69 (1.62, 1.76)

3.88 (3.61, 4.17)

LR- (95% CI)

0.07 (0.03, 0.15)

0.22 (0.16, 0.30)

Conclusions: The prediction rule had a significantly better sensitivity than clinician suspicion for identifying IAIAI. Furthermore, the higher specificity of clinician suspicion for IAIAI did not translate into clinical practice. Clinicians frequently obtained CT scans in patients considered at low risk for IAIAI. This prediction rule can assist in clinical decision making around CT use for IAIAI after blunt abdominal trauma in children.