Jim Dandy:  To my mind, that potentially variable meaning of "starts" won't lead to informative statistics.  FYI, I tended to take numbers presented to the Welfare and Safety of the Racehorse Summit(s).  Those strike me as the most reliable I've seen (even though they wound up having to revise some of the data).  Otherwise, I think there's just too much state-to-state variation, if the numbers aren't comparable.  I wouldn't even dare to look at data outside the U.S., other than to say that jumping races are far more deadly.

I miss chatting with one of my curling mates.  She's quite the horsewoman and loves hers dearly.  I'd love to hear her opinions on some of these matters.