Online supplementary table E3: Risk of bias in primary studies

First author, yr / Risk of bias assessment / Total score
Case series / Newcastle–Ottawa Quality Assessment Scale[16] / (/8)
Selection (0-3)# / Comparability (0-2) / Outcome (0-3)
Andersen 2013[22] / 1 / - / 2 / 3
Hoffman 2013[24] / 2 / - / 2 / 4
Peddle-McIntyre 2013[21] / 2 / - / 3 / 5
Peddle 2009[23] / 2 / - / 3 / 5
Mean (SD) / 1.7 ± 0.5 / 0 / 2.5 ± 0.6 / 4.2 ± 1.0
Cross-sectional / Agency for Healthcare Research and Quality Methodology Checklist[17] / (/11)
Kartolo 2016[25] / 7
Leach 2015[27] / 8
Phillip 2015[26] / 6
Granger 2014[29] / 11
Phillip 2014[28] / 7
Cavalheri 2013[32] / 8
Lebel 2013[31] / 8
Lin 2013[30] / 8
Nwosu 2012[33] / 10
Coups 2009[34] / 8
Clark 2008[35] / 9
Mean (SD) / 8.2 ± 1.4
Qualitative / COREQ checklist[18]
Team & reflexivity (/8) / Study design (/15) / Analysis & findings(/9) / (/32)
Brown 2015[39] / 0 / 10 / 7 / 17
Kujpers 2015[38] / 1 / 7 / 7 / 15
Mas 2015[37] / 3 / 11 / 6 / 20
Missel 2015[36] / 0 / 11 / 5 / 16
Hoffman 2014[40] / 1 / 10 / 8 / 19
Lehto 2013[42] / 1 / 7 / 7 / 15
Maddocks 2013[41] / 0 / 9 / 2 / 11
Adamsen 2012[45] / 2 / 8 / 6 / 16
Cheville 2012[44] / 2 / 12 / 6 / 20
Ellis 2012[43] / 1 / 9 / 7 / 17
Jones 2009[46] / 0 / 10 / 5 / 15
Mean (SD) / 1.0 ± 1.0 / 9.4 ± 1.6 / 6.0 ± 1.6 / 16.4 ± 2.6

Abbreviations: COREQ, Consolidated criteria for reporting qualitative research; SD, standard deviation; yr, year.

* Criteria satisfied - Criteria not satisfied, high risk of bias

# Domain 1 (selection) scored out of 3 not the traditional 4 points because one item from this domain was deemed not appropriate for this review (demonstration that outcome of interest was not present at start of study)