Table 3.

Evans et al. proposed a way to utilize composite endpoints: the desirability of outcome ranking (DOOR) and the response adjusted for the duration of antibiotic risk (RADAR) [33]

	DOOR/RADAR Evans et al. [33]	Win ratio Pocock et al. [32]
Relevance of the outcome is taken into account	Categorize patients based on overall clinical outcome as pre-specified in categories with increasing severity	Study the more serious event in matched pairs. Did the pair member with the treatment of interest experience it first: assign “loser”. Did the control member experience it first: assign “winner”
Outcome with lower priority is considered	Determine days of antibiotic use and rank patients within categories, whereby a shorter duration results in a lower rank (optional)	If no serious event occurred within a pair, study a less serious event. Who had it first? Determine whether the pair is a “winner” or “loser”
Patients are compared/ranked and effect measure is calculated	Rank all patients over all categories: Determine the number of control patients with a lower rank for each treated patient. Divide this sum by the total number of possible pairwise comparisons: Probability of a better rank for a random patient of the treatment group	Divide the number of “winners” by the number of “losers” to calculate the win ratio

Pocock et al. suggested the win ratio as a new effect measure that takes the different priorities of the components into account [32]. The generic version of DOOR/RADAR is very similar to the win ratio