OBJECTIVES: Cancer is the second leading cause of death worldwide. Mechanisms involved in the relationship between type 2 Diabetes (T2D) and cancer is not completely understood, and even less is known about its association with Non-Small Cell Lung Cancer (NSCLC). While these systemic responses are reasonably well understood, an adequate explanation how health care utilization is impacted for patients with both T2D and NSCLC is lacking. Our objective was to investigate the cost of care for T2D patients with NSCLC.

METHODS:

We conducted a cohort study design utilizing Medicare claims from the SEER-Medicare linked database (2007 - 2014). Quantile regression was used to estimate the predicted mean at the 25^th, 50^th and 75^th percentile for total healthcare costs after incident NSCLC diagnosis. The quantile regression was adjusted for cancer stage, geographical region, gender, age, race and comorbidities. Total healthcare costs were then divided by the number of days within the 25^th, 50^th and 75^th quantiles to adjust for varying follow up. Cost data represented the actual paid (reimbursed by Medicare) amounts for health services. x

RESULTS:

Of 17,176 NSCLC patients, 5,096 patients had T2D in the pre-period. For NSCLC patients, the predicted mean value at the 25^th quantile was $2.36/day, at the 50^thquantile $5.69/day and at the 75^thquantile /$7.23/day. For, diabetic NSCLC patients, the predicted mean value at the 25^thquantile was $4.13/day, at the 50^thquantile $9.16/day and at the 75^th \quantile $48.64/day.

CONCLUSIONS: This study demonstrates the challenges in analyzing cost data due to variations in factors impacting healthcare costs. Care should be taken when developing interventions directing resources at those most likely to benefit as traditional regression analysis may be less useful than methods such as quantile regression, which provide robust insights into healthcare cost of specific patient sub-groups.