. 2022 Dec 12;11:265. doi: 10.1186/s13643-022-02118-9

Table 3.

Risk of bias of studies included

Study	Parallelism of groups	Comparability of groups	Blinding		Selective reporting improbable	Absence of other factors potentially causing bias	Risk of bias: study level	Risk of bias: outcome level
Study	Parallelism of groups	Comparability of groups	Patients	Treating staff	Selective reporting improbable	Absence of other factors potentially causing bias	Risk of bias: study level	Risk of bias: outcome level
Carlson 2021 [18, 19]	Yes	Yes	No	No	Unclear	No^a	High	High
Myrseth 2009 [20]	Yes	No^b	No	No	Unclear	Yes	High	High
Pollock 2006 [21]	Yes	No^c	No	No	Unclear	Yes	High	High

^apatients were screened for inclusion criteria between 2005 and 2019, although the study start date was reported as 2014. Thus, it remains unclear for what proportion of patients the survey was at least partially retrospective (“obtained by medical record review”). Based on all publications, it cannot be assumed that this is a retrospective study

^balthough the essential data are available at baseline (age, sex, symptom severity, and tumour size), the groups differ in age by an average of 5.0 years (p = 0.06)

^calthough the essential data are available at baseline (age, sex, symptom severity, and tumour size), the groups differ statistically significantly in age by an average of 5.7 years (p = 0.03)