Figure 3. Representativeness of the Sutter Health and Kaiser Permanente Hawai’i Lung Cancer Cohort, 2000–2013.

(a) Distribution of females and males in the pooled cohort by age and race/ethnicity are compared to the 2010 US Census. (b) Proportion of never smoking females and males by age and race/ethnicityare compared to U.S. Behavioral Risk Factors Surveillance System (BRFSS) data.. (c) Demographic (age, race/ethnicity) and tumor (stage, histology) characteristics for female and male lung cancer cases in the pooled cohort were compared to SEER-18. Standardized mean differences (SMD), which may not be representative of the reference population when they are beyond +/− 15%, were calculated as $SMD=\frac{P_s-P_r}{\sqrt{\frac{\left(P_S^{\ast }\left(1-P_S\right)\right)+\left(P_Y^{\ast }\left(1-P_r\right)\right)}{2}}}$ where $P_S$ is the study proportion and $P_r$ is the reference population proportion for each variable category.