Standard 9b: a consensus definition was described a priori
Eleven studies (42%) met the standard, providing clear evidence that
consensus methods were defined a priori. In eight studies (31%) it was
unclear and seven (27%) did not use consensus methods. Those that met
this criterion provided clear evidence in published protocols, COMET
registry entries or authors stated ‘a priori ’ consensus
definition.
Standard 10: Criteria for including/dropping/adding
outcomes was described a priori.
Eight studies (31%) met this standard providing clear evidence that all
three elements were defined a priori through published protocols,
registry entry or stated ‘a priori ’ within the body of the text
specific to each element. In 18 (69%) studies it was unclear if the
standard was met, commonly because all three elements were not clearly
described.
Standard 11: Care was taken to avoid ambiguity of
language used in the list of
outcomes
Ten studies (38%) met this standard if evidence was described in either
the protocol or main study paper. Perry et al. for example developed
‘lay definitions for individual outcomes’ which were reviewed by
consumer group representatives,22 while the study
protocol of Bogdanet et al. described ‘the questionnaire will
contain lay terminology… ’.23 One study (4%)
did not meet the standard, describing as a limitation, ‘illegible
translated outcomes that were not included in the
list ’.24 In 15 studies (58%) it was unclear if
language ambiguity had been considered.
Outcomes and measurement
considerations
Core outcomes, definitions, and measurement considerations described in
26 included COS are outlined in Table S7. The number of outcomes
included in each COS ranged from six to 56. Maternal COS included both
maternal and fetal/neonatal outcomes (Mdn = 17, range = 50), while
neonatal specific COS generally included only neonatal outcomes (Mdn =
8, range = 20). To aid analysis, outcomes were organised into grouping
domains (i.e. survival, maternal morbidity, neonatal morbidity, resource
utilisation). Survival was common across 16 separate COS, related to
maternal death, fetal and neonatal loss but only clearly defined in
three COS.19,25,26 Similarly, resource utilisation was
shared across 12 COS relating mainly to maternal/neonatal admission to
intensive care, but definitions were only clearly defined in one
COS.22
Significant overlap of outcomes between similar studies was evident. For
example, eclampsia and pre-eclampsia are core outcomes outlined in six
separate COS,25,27-31 but only defined in
one.31 Similarly, while maternal haemorrhage is a
domain shared across three separate COS, 25,32,33 a
definition is only offered in one.33 Two COS (8%),
related to maternity care and gastroschisis also addressed ‘how’and ‘when’ to measure outcomes.19,26 While how
to measure outcomes were considered in four additional
papers,33-36 clear recommendations were not reported.
Although future work is planned by five COS developers to outline
recommendations for how and when to measure
outcomes,22-24,37,38 and is acknowledged as needed by
two,30,39 17 COS offer no guidance on how or when to
measure outcomes, with no reported future plans to do so.