First reply is no, but I’d not rule it out nor would I make it the first thing I do. Do you know if \ you have a large number of concurrent requests that are timing out where splitting them to multiple origins would fix the timeouts?
Next, I’d check the origin server to see if you can find why it’s timing out. Once you can rule out the suggestions in the tip
you could figure out if is a better than load balancing.