J. CHEETHAM, A. S. RIORDAN, H. O. MOHAMMED † , C. W. McILWRAITH ‡ and L. A. FORTIER Departments of Clinical Sciences, and †Population Medicine and Diagnostic Sciences, Cornell University, Ithaca, New York 14853; and ‡Orthopaedic Research Center, Colorado State University, Fort Collins, Colorado 80523, USA. Equine Veterinary Journal May 2010 42 4 346 - 350 Reasons for performing study: There is no consensus on objective outcome measures that can be used to determine if a medical or surgical treatment affects race performance. Objective: To determine the association between 2 commonly used outcome measures (total starts and total earnings) and age, sex, gait and race surface. Methods: A cross-sectional study was performed using the race performance data for all Thoroughbred horses age 2, 3, 4 and 5 years racing in the United States, and Standardbred horses of the same ages racing in the United States and Canada during the year 2006. Median earnings and starts were determined for each combination of age, sex and track surface (for Thoroughbred) or gait (for Standardbred). The effect these variables had on starts on race earnings ($) was determined using linear regression. Results: Race records for 68,649 Thoroughbreds and 25,830 Standardbreds were obtained. All independent variables (age, breed, sex, gait, track surface and total number of starts) had a significant impact on total earnings (P<0.0001). Conclusions: The data show considerable variation across age groups and track surfaces for Thoroughbreds and across age groups for Standardbreds. They also show that the decision to use earnings or starts as outcome measures could have a marked effect on reported success for a particular treatment. Potential relevance: Both earning and start data should be reported in studies evaluating outcome following surgery or other intervention. Considerations of age, breed, sex, track surface and gait should be included in the design of these studies.

• Epidemiology
• Equine