Fig. 2

Forrest plot showing unadjusted odds ratios of appropriateness of level three care associated with demographic and clinical variables. There were associations between treatment escalation decisions and age, frailty and burden of co-morbidities