Enrollment Trends and Disparity Among Patients With Lung Cancer in National Clinical Trials, 1990 to 2012

Abstract

Purpose Under-representation of elderly, women, and racial/ethnic minority patients with cancer in clinical trials is of national concern. The goal of this study was to characterize enrollment trends and disparities by age, sex, and race/ethnicity in lung cancer trials. Methods We analyzed data for 23,006 National Cancer Institute cooperative group lung cancer trial participants and 578,476 patients with lung cancer from the SEER registry from 1990 to 2012. The enrollment disparity difference (EDD) and enrollment disparity ratio (EDR) were calculated on the basis of the proportion of each subgroup in the trial population and the US lung cancer population. Annual percentage changes (APCs) in the subgroup proportions in each population were compared over time. Results Enrollment disparity for patients ≥ 70 years of age with non-small-cell lung cancer improved from 1990 to 2012 (test of parallelism, P = .020), with a remaining EDD of 0.22 (95% CI, 0.19 to 0.25) and EDR of 1.65 (95% CI, 1.51 to 1.82) in 2010 to 2012. No improvement was seen for elderly patients with small-cell lung cancer (SCLC), with an APC of 0.20 ( P = .714) among trial participants, despite a rising proportion of elderly patients with SCLC in the US population (APC, 0.32; P = .020). Enrollment disparity for women with lung cancer improved overall, with the gap closing by 2012 (EDD, 0.03 [95% CI, 0.00 to 0.06]; EDR, 1.07 [95% CI, 1.00 to 1.16]). Enrollment disparities persisted without significant improvement for elderly women, blacks, Asians/Pacific Islanders, and Hispanics. Conclusion Under-representation in lung cancer trials improved significantly from 1990 to 2012 for elderly patients with non-small-cell lung cancer and for women, but ongoing efforts to improve the enrollment of elderly patients with SCLC and minorities are needed. Our study highlights the importance of addressing enrollment disparities by demographic and disease subgroups to better target under-represented groups of patients with lung cancer.

Fig 1.

Proportion of elderly and women among patients with lung cancer: trial participants and US population, 1990 to 2012. (A) Elderly among patients with lung cancer. (B) Women among patients with lung cancer. (C) Elderly among patients with non–small-cell lung cancer (NSCLC). (D) Elderly among patients with small-cell lung cancer (SCLC). The annual percentage change (APC) P value corresponds to testing whether the APC is different from 0. The solid lines represent the fitted values of the joinpoint regression.

Fig 2.

Enrollment disparity difference with 95% CIs, 1990 to 2012. (A) Elderly among patients with lung cancer. (B) Elderly among patients with non–small-cell lung cancer (NSCLC). (C) Elderly among patients with small-cell lung cancer (SCLC). (D) Women among patients with lung cancer. The last year group consists of only 3 years because the study period was from 1990 to 2012. The definition of enrollment disparity difference is the absolute difference between the estimated subgroup proportion among the US lung cancer population and the subgroup proportion among trial participants. The solid lines are the bootstrapped 95% CIs.

Fig 3.

Proportion of black, Asian/Pacific Islander, and Hispanic among patients with lung cancer: trial participants and the US population, 1990 to 2012. (A) Black. (B) Asian/Pacific Islander. (C) Hispanic. The last year group consists of only 3 years because the study period was from 1990 to 2012. The solid lines are the bootstrapped 95% CIs. The P value corresponds to the test for equality of proportions between SEER and the trial for 1990 to 2012.

Fig 4.

Proportion of elderly and women patients: trial participants and US population, 1990 to 2012. (A) Elderly among patients with regional non–small-cell lung cancer (NSCLC). (B) Elderly among patients with distant NSCLC. (C) Women among patients with regional NSCLC. (D) Women among patients with distant NSCLC. The annual percentage change (APC) P value corresponds to testing whether the APC is different from 0. The solid lines represent the fitted values of the joinpoint regression. The year 1990 was excluded from analysis because of the small number of trial participants with regional or distant NSCLC.