Table 3

Changes in outcomes between review protocol and publication

Primary outcomesSecondary outcomes
Total number of outcomes included in the review (Median, IQR, range)3 (IQR 2, 3 and range 1, 8)7 (IQR 5, 9 and range 2, 13)
Reviews with any discrepancy in outcomes between protocol and full reviewProtocol distinguished outcomes (n=37)*14 (38%)
Protocol did not distinguish outcomes (n=9)†4 (44%)
Reviews which have upgraded at least one outcome from secondary in the protocol to primary in the full review (number of outcomes; minimum per review; maximum per review)Protocol distinguished outcomes (n=37)*3 (8%) (3 outcomes)
Protocol did not distinguish outcomes (n=9)†0
Reviews which have downgraded at least one outcome from primary in the protocol to secondary in the full review (number of outcomes; minimum per review; maximum per review)Protocol distinguished outcomes (n=37)*9 (24%)
(16 outcomes; min 1, max 5)
Protocol did not distinguish outcomes (n=9)†1 (11%)
(2 outcomes)
Reviews which have added a new outcome in the full review which was not included in the protocol (number of outcomes; minimum per review; maximum per review)Protocol distinguished outcomes (n=37)*2 (5%)
(3 outcomes)
2 (5%)
(4 outcomes; min 1, max 3)
Protocol did not distinguish outcomes (n=9)†1(11%)
(1 outcome)
2 (22%)
(2 outcomes)
Reviews which have excluded an outcome from the full review which was included in the protocol (number of outcomes; minimum per review; maximum per review)Protocol distinguished outcomes (n=37)*2 (5%)
(10 outcomes; min 1, max 9)
3 (8%)
(5 outcomes; min 1; max 2)
Protocol did not distinguish outcomes (n=9)†00
  • *Protocol distinguished primary from secondary outcomes.

  • †Protocol did not distinguish primary from secondary outcomes.