Intervention treatment distributions that depend on the observed treatment process and model double robustness in causal survival analysis