Lung cancer is predominantly a disease of older adults. Elderly patients however, are less aggressively treated and under-represented in clinical trials. As a consequence, there are several areas of controversy regarding the treatment of these patients. The goal of this proposal is to use national, population-based data to assess the effectiveness of several key therapies for elderly patients with stage I, II, and IMA non-small cell lung cancer (NSCLC).
The specific aims are to: 1) determine whether survival after limited resection (wedge resection or segmentectomy) is equivalent to that after lobectomy in patients with small (5 2cm) stage I NSCLC, 2) assess the real world effectiveness of adjuvant chemotherapy in elderly patients with resected stage I to IMA NSCLC, 3) evaluate the effectiveness of RT alone for the treatment of unresected stage I and II NSCLC, and 4) assess the effectiveness of chemotherapy for the primary treatment of elderly patients with unresected stage I and II NSCLC. To address these aims, we will use data from the Surveillance, Epidemiology, and End Results (SEER) registry linked to Medicare claims. The study cohort will be comprised of all patients >65 years with histologically confirmed stage I, II, and IMA NSCLC diagnosed between 1991 and 2002. Sociodemographic data and information on the extent of disease at diagnosis will be obtained from SEER. Data on the initial course of treatment will be ascertained from both SEER and Medicare. Linked data from the Area Resource File will be used to obtain information on community level factors such as physician availability. The primary outcome will be overall survival and the secondary outcome will be lung cancer-specific survival, as assessed by Medicare and SEER data, respectively.
Each Specific Aim will be evaluated by comparing unadjusted survival rates of patients who received and did not receive the therapy under study, by conducting stratified analyses within relevant subgroups, and by using multivariate analyses to control for important prognostic factors such as socioeconomic characteristics, physician availability, tumor histology, stage, and comorbidities. Additionally, we will use propensity score methods, sensitivity analysis, and instrumental variable techniques to minimize bias inherent to observational studies. The findings from this project will provide information that can directly impact the care of elderly patients with NSCLC and inform the design of future trials of lung cancer treatment strategies.
