That isn't a scientific poll. It was opt-in, it wasn't weighted and it had a low sample size. That means the results just reflect the opinions of the 250 who happened to vote, not physicians as a whole. Even if it did, you couldn't rule out political bias - for instance, they could be all higher-earners who favour the Republican Party for their lower tax rates.
As for the original question - if Clinton were to withdraw from the race, the Democrats would need to select another candidate which would be done by the national committee.