Surrogate outcomes: experiences at the Common Drug Review

Table 3 Analysis of the drug recommendations with surrogate outcomes

Factor	Descriptive analysis		Univariate logistic regression analysis based on “not-accepted” surrogates
Factor	Accepted surrogates^¥N = 40	Non-accepted surrogates^§N = 28	Odds ratios^∞(95%CI) [p value]*
*Factors Associated with a DNL*
DNL recommendation	10/40 (25%)	18/28 (64%)	5.4 (1.9-15.5) [p = 0.002]
Clinical uncertainty	8/40 (20%)	18/28 (64%)	7.2 (2.4-21.5) [p < 0.001]
Price Only	28/40 (70%)	6/28 (21%)	0.1 (0.0-0.4) [p < 0.001]
Economic considered	11/40 (28%)	17/28 (61%)	4.1 (1.4-11.4) [p = 0.007]
Price greater than comparators	13/40 (33%)	16/28 (57%)	2.8 (1.0-7.5) [p = 0.046]
*Other Factors*
First in class	6/40 (15%)	14/28 (50%)	5.7 (1.8-17.7) [p = 0.003]
First in disease	2/40 (5%)	5/28 (18%)	4.1 (0.7-23.1) [p = 0.106]
Life threatening	3/40 (8%)	9/28 (32%)	5.8 (1.4-24.1) [p = 0.015]
Priority review requested	7/40 (18%)	9/28 (32%)	2.2 (0.7-7.0) [p = 0.166]

^∞; Odds Ratio > 1 is associated with higher odds of the surrogate not being accepted given the presence of a factor.
* Bold p values indicate statistical significance under univariate analysis.
^¥Accepted surrogates = all the recommendations with surrogate acceptability classified as: yes (e1) = implicit yes “evidence 1”; yes (e2) = implicit yes “evidence 2”; yes (used) = implicit yes “used before”; yes (ref) = implicit yes “reference”; yes (e) = explicit yes; N/S = no statement; N/A = not applicable.
^§Not-accepted surrogates = all the recommendations with surrogate acceptability classified as: no (e2) = implicit no “evidence 2”; no (ref) = implicit no “reference”; no (e) = explicit no “evidence 1”; no (e1 + e2) = explicit no “evidence 1” and implicit no “evidence 2”.

ISSN: 1478-7547