Table 2:

Summary of risk of bias and quality of study reporting

	Study risk of bias		Study reporting quality
	Entire cohort		Entire cohorta			Cocaine sub-cohorta
	Representativeness	Outcome measurement	Cohort description (age/sex)	Completeness of data (numerator/denominator)	Cause of death definitionb	Cohort description (age/sex)	Completeness of data (numerator/denominator)
Accurso, 2015(39)	↑	↓	↑	↓	NA	↓	↓
Arendt, 2011(40)	↓	↓	↑	↑	NA	↓	↓
Barrio, 2013(41)	↓	↓	↑	↑	↓	↑	↑
Bohnert, 2017(23)	↑	↓	↑	↓	↑	↓	↓
Callaghan, 2013(22)	↓	↓	↑	↑	↑	↑	↑
Callaghan, 2012(42)	↓	↓	↑	↓	NA	↑	↓
de la Fuente, 2016(43)	↓	↓	↑	↑	NA	↑	↓
Brugal, 2016(44)	↓	↓	↑	↑	↑	↑	↑
Colell, 2018(45)	↑	↓	↑	↑	NA	↑	↑
Molist, 2018(21)	↓	↓	↑	↑	↑	↑	↑
Dias, 2011(46)	↑	↑	↑	↓	↓	↑	↓
Gossop, 2002(47)	↑	↓	↓	↓	↑	↓	↓
Hayashi, 2016(48)	↓	↓	↑	↑	↑	↓	↓
Tyndall, 2001(49)	↓	↓	↑	↓	↑	↓	↓
Hser, 2012(50)	↓	↓	↑	↓	↓	↓	↓
Lopez, 2004 (OFDT)(51)	↑	↓	↑	↓	↓	↓	↓
Markota, 2016(52)	↓	↑	↓	↓	↓	↓	↓
Martell, 2009^(53)	↑	↑	↑	↓	NA	↑	↓
Nielsen, 2011(54)	↓	↓	↑	↓	↑	↓	↓
O’Driscoll, 2001(55)	↓	↓	↑	↑	↓	↓	↑
Pavarin, 2017(56)	↑	↓	↑	↓	↓	↑	↓
Pavarin, 2008(57)	↑	↓	↑	↑	↑	↑	↑
Pavarin, 2013(58)	↑	↓	↑	↓	↓	↑	↓
Ryb, 2009(59)	↑	↓	↑	↓	↓	↑	↓
Sanvisens, 2014(60)	↑	↓	↑	↑	↑	↑	↑
van Haastrecht, 1996(61)	↓	↓	↓	↑	↓	↓	↑
Vlahov, 2008(62)	↓	↓	↑	↑	NA	↓	↓
Wang, 2005(63)	↓	↓	↑	↑	↓	↓	↓
Higher risk of bias/ Lower quality of reportingc	9/21 (42.9%)	3/21 (14.9%)	3/21 (14.9%)	11/21 (52.4%)	10/17 (58.8%)	13/21 (61.9%)	16/21 (76.2%)

Note. Full details of the risk of bias and quality of study reporting assessment are available in Appendix H, noting that studies with insufficient information to evaluate each criteria were deemed ‘higher risk of bias’/’lower quality of reporting’; ↓ denotes lower risk of bias or lower quality of reporting; ↑ denotes higher risk of bias or higher quality of reporting

^a‘Entire cohort’ refers to reporting for sample of interest in the original study which may include people who do not consume cocaine; ‘cocaine sub-cohort’ refers to reporting where mortality outcomes were reported for the subsample of interest for this paper (i.e., people with regular/problematic cocaine use) but whom may not be the primary focus of the paper.

^bIncludes 17 cohorts with cause-specific mortality data; ‘NA’ denotes cohorts where cause-specific mortality data were not collected;

^cTotals only include the primary publication if there are multiple publications reporting on the same cohort.