Table 1

Pooled risk estimates for prostate cancer by exposure assessment method, study design, publication year period and geographic region, based on meta-analysis of articles on occupational pesticide exposure published between 1995 and 2019.

	Number of risk estimates	sRR	95% CI	Heterogeneity measures
	Number of risk estimates	sRR	95% CI	I² (%)	P value	Q	P value
Exposure assessment method						3.28	0.35
Group-level	13	1.09	1.00 to 1.20	92	<0.01
Self-reported exposure	5	1.35	0.95 to 1.94	76	<0.01
Expert-level	8	1.41	0.99 to 2.01	79	<0.01
Biomonitoring	1	1.32	0.75 to 2.33	.	.
Study design						7.59	<0.02
Cohort (prospective)	5	1.08	1.03 to 1.14	64	<0.01
Cohort (retrospective)	8	1.09	0.90 to 1.31	95	<0.01
Case–control	12	1.63	1.22 to 2.18	79	<0.01
Publication year period						0.01	0.93
1995–2006	14	1.12	0.94 to 1.35	92	<0.01
2007–2019	13	1.11	1.04 to 1.19	77	<0.01
Geographic region						9.15	<0.01
Europe	12	1.03	0.96 to 1.11	66	<0.01
North America	13	1.28	1.13 to 1.45	92	<0.01
Other	2	2.17	0.42 to 11.4	86	<0.01
Case–control studies only
Exposure assessment method						1.15	0.56
Group-level	1	2.37	1.22 to 4.61	.	.
Self-reported exposure	4	1.53	0.89 to 2.62	68	0.02
Expert-level	7	1.63	1.11 to 2.40	79	<0.01
Exposure assessment method during publication year periods
1995–2006
Exposure assessment method						0.53	0.91
Group-level	7	1.16	0.87 to 1.55	95	<0.01
Self-reported exposure	2	1.02	0.45 to 2.33	67	0.08
Expert-level	4	1.04	0.69 to 1.57	70	0.02
Biomonitoring	1	1.32	0.75 to 2.33	.	.
2007–2019
Exposure assessment method						5.8	0.05
Group-level	6	1.08	1.02 to 1.14	73	<0.01
Self-reported exposure	3	1.57	0.96 to 2.56	85	<0.01
Expert-level	4	2.00	1.07 to 3.75	76	<0.01

I²=percentage of variation across studies due to heterogeneity
Q=Cochran’s Q.
sRR, summary risk ratio.