Development of a colic severity score for predicting the outcome of equine colic.

The research aimed to develop a Colic Severity Score (CSS) based on certain physical and laboratory variables which could predict the survival outcome of horses suffering from acute abdominal disease. The study found high accuracy in this method with a score higher than 7 indicating a fatal outcome.

Research Methodology and Variables

• The researchers first identified 32 potential variables that could be indicators of a horse’s survival likelihood. These included physical examination and laboratory variables.
• Data was collected from 165 horses admitted for acute abdominal disease. The researchers then took measurements or observations for each of the chosen variables.

Analysis and Model Setup

• Univariate analyses were performed to determine the statistical significance of these variables in differentiating between horses that survived and those that didn’t.
• Following this, stepwise logistic regression was conducted to identify the variables with the highest predictive value.
• Of all the initial variables, only four (heart rate, peritoneal fluid total protein concentration, blood lactate concentration, and abnormal mucous membrane) remained statistically significant. These formed the model for the Colic Severity Score (CSS).
• The researchers then set specific thresholds or “cutting-points” for each of the four variables. These served to categorise the variables and assign specific values. These values would then be added up to calculate the CSS for each horse.

Validation of the Scoring Chart

• The scoring chart was then tested on 71 horses separate from the initial group. These horses were also suffering from acute abdominal disease, and their CSS score was calculated and compared to their survival outcome.
• The mortality rate in this group was similar to that of the initial group (21.1% versus 20.6%), which indicated a level of consistency in results across different sample groups.

Results and Relevance of Colic Severity Score

• All horses with a CSS above 7 were found to have died, while about 75% of horses with a CSS of 7 or below survived.
• The positive predictive value was 100% indicating that all horses predicted to have fatal outcomes based on their score did indeed die.
• The negative predictive value was 91.8% meaning that the score correctly identified 91.8% of horses that survived.
• Sensitivity stood at 66.7% representing the percentage of positive results correctly identified as such by the score.
• The score also had a specificity of 100% showing that all negative outcomes were correctly identified.
• Overall, the system was found to have an accuracy of 93% in predicting the outcome of equine acute abdominal disease, showing its potential as a useful tool for veterinary practitioners.