Related Experiment Videos
Multivariable prediction model for the need for surgery in horses with colic
M J Reeves1, C R Curtis, M D Salman
1Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins 80523.
American Journal of Veterinary Research
|November 1, 1991
Summary
A multivariable model was developed to predict the need for surgery in equine colic cases. However, the model demonstrated a poor fit when validated, indicating limitations in its predictive accuracy for surgical intervention in horses.
Area of Science:
- Veterinary Medicine
- Epidemiology
- Biostatistics
Background:
- Equine colic presents a significant challenge in veterinary practice, necessitating accurate diagnostic tools.
- Predicting the need for surgical intervention in colic cases is crucial for timely and effective treatment.
Purpose of the Study:
- To develop and validate a multivariable predictive model for surgical intervention in equine colic cases.
- To identify key clinical variables associated with the need for surgery.
Main Methods:
- A multivariable logistic regression model was constructed using a stepwise algorithm.
- Data from 1,965 equine colic cases across 10 referral centers were utilized.
- Variables included rectal findings, abdominal pain signs, pulse strength, and abdominal sounds.
Main Results:
- The developed model included rectal findings, abdominal pain signs, peripheral pulse strength, and abdominal sounds.
- Validation of the model using a separate dataset revealed a poor fit, indicated by a high Hosmer-Lemeshow chi-squared statistic (26.7, P < 0.001).
Conclusions:
- The developed multivariable model showed limitations in accurately predicting the need for surgery in equine colic cases upon validation.
- Further refinement and validation are necessary for a reliable surgical prediction model in equine colic.