Minimising the biases in the observational study of resuscitative endovascular balloon occlusion of the aorta: a research protocol for a prospective study analysed with propensity score matching with time-varying covariates