New Drugs Not Major Cost Drivers for Lung Cancer Care

We characterized the direct cost of stage 4 non-small cell lung cancer from 2000 to 2009, with a focus on drug treatments.
Published Online: June 20, 2014
Allison M. Bell, PharmD, MSc, BCPS; Mitra Corral, MS, MPH; John R. Penrod, PhD; Ying Zhang, MD, MSPH; Christopher R. Frei, PharmD, MSc, BCPS; Laurajo Ryan, PharmD, MSc, BCPS, CDE; Mark E. Salvati, PhD; and Jim M. Koeller, MS
Objectives: Our objective was to characterize the direct cost of stage 4 non-small cell lung cancer (NSCLC) from 2000 to 2009, cost of disease segments (diagnosis, active treatment, end-of-life), cost impact of new therapies, and cost trend from 2003 to 2009.

Study Design: We used the PharMetrics claims database to con- duct a retrospective cohort study using International Classification of Diseases, Ninth Revision, Clinical Modification codes.

Methods: Patients 20 years and older with newly diagnosed metastatic NSCLC were included. The study period was divided into the pre (2000-2002), transition (2003-2005), and current (2006-2009) periods to account for incorporation of newer drugs (biologics, tyrosine kinase inhibitors, pemetrexed) into guidelines. Descriptive statistics (median), χ2 test, and Wilcoxan rank sum tests were used. We created multiple least squares regression models to adjust for baseline confounders.

Results: Overall cost (n = 969) from 2000 to 2009 was $10,281 per patient per month (PPPM), with no difference between transi- tion ($10,525 PPPM; n = 439) and current ($10,226 PPPM; n = 503) periods. Overall diagnosis cost was $6601 PPPM. Active treatment cost was $9287 PPPM. End-of-life cost was $12,215 PPPM. Drug cost was 23% of overall cost in transition and 21% in current. Gold standard cost (≥5 months active treatment; n = 316) was $144,147 per patient and $9371 PPPM. Comorbidities had no effect on cost.

Conclusions: This study showed no significant change in cost from 2003 to 2009 in the treatment of stage 4 NSCLC. Individual newer agent costs did not appear to be primary drivers of treatment cost, contradicting the belief that newer cancer treatments are driving cost increases. However, drug cost still represented roughly 20% of cost for stage 4 NSCLC.

Am J Pharm Benefits. 2014;6(3):e50-e59
New drug products are often thought to be significant drivers of cost. Pub- lished literature characterizing the cost of stage 4 non-small cell lung can- cer (NSCLC) since new drug treatments have come to market is lacking. In this unbranded study, we sought to characterize changes in cost of stage 4 NSCLC from 2000 to 2009, a time when many new cancer treatments appeared on the market.
  • No change in cost of treating stage 4 NSCLC.

  • Individual newer agents did not appear to be primary drivers of cost.

  • Characterized cost of 5 or more months of active treatment.

  • Comorbidities had no effect on cost.
Lung and bronchus cancer combined were the lead- ing cause of cancer-related deaths in 2010,1 with non-small cell lung cancer (NSCLC) constituting ap- proximately 85% of these cases.2,3 The median overall survival (OS) for locally advanced (stage 3b with pleural effusions) or metastatic (stage 4) NSCLC is less than 1 year.2,3

The National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines recommend platinum doublet chemotherapy (PDC) over best supportive care for NSCLC treatment.4 First-line recommendations include combining PDC with biologic agents (either bevacizumab or cetuximab). Cetuximab has compendia listing for reimbursement but is not FDA approved. Pemetrexed, a newer antimetabolite, is another option for first-line use with platins. The small-molecule tyrosine kinase inhibitor (TKI) erlotinib is a first-line monotherapy option in patients with epidermal growth factor receptor-positive (EGFR+) mutation tumors. Second-line options are single-agent: docetaxel, pemetrexed, or erlotinib.4 Erlotinib is also compendia listed for third-line therapy.4 Some treatments are indicated for specific cell histology, narrowing use to particular patient subsets.

Newer treatments (biologics, TKIs, pemetrexed) are perceived to contribute to a substantial increase in the cost of treating NSCLC. Recent literature describing overall cost of NSCLC uses data from earlier this decade,5 not including newer treatments. No studies evaluating direct medical cost of care with current treatment standards have been published.

Our primary objective was to characterize total and segmented (diagnosis, active treatment, end-of-life) direct cost of care for stage 4 NSCLC treatment. We analyzed cost of biologics, TKIs, pemetrexed, PDC, and other prescription medications during active treatment of stage 4 NSCLC across all lines of therapy, as well as cost of drugs relative to overall NSCLC cost and cost trend from 2003 to 2009.


This private payer claims (PharMetrics) database study spanned the period of January 1, 2000, through June 30, 2009 (study period). Newly diagnosed stage 4 NSCLC cases between July 1, 2000, and June 30, 2008 (enrollment period), were identified by an algorithm developed for use with healthcare claims data and International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) coding (Appendix).6-8

Inclusion criterion was patients 20 years and older diagnosed with primary lung cancer (ICD-9-CM 162. xx) plus secondary malignancy (ICD-9-CM 197.xx or 198.xx). Codes appeared on 1 or more inpatient or 2 or more outpatient claims, with the claim for primary lung cancer falling within 6 months prior or 30 days following index date (date of first metastatic disease). Patients with small cell lung cancer standard-of-care treatment9,10 or evidence of another primary cancer were excluded.

Segments of Disease

Disease management was divided into 3 segments: diagnosis, active treatment, and end-of-life (Figure). Diagnosis was defined as 90 days prior to index until initiation of active treatment. Active treatment was defined as claims for NSCLC agents and spanned the time between end of diagnosis and the beginning of end-of-life segment. End-of-life was defined as 30 days prior to death.11 Death was defined by a published proxy using ICD-9-CM claims.12-14 Claims for all 3 segments were required for inclusion. Patients receiving active treatment prior to index date were excluded.

Time Periods

The emergence of new treatments allowed stratifica- tion into 3 separate time periods (Figure). Patients were classified into time periods by index date. Patients with an index date between July 1, 2000, and December 31, 2002, were in the pre-period (prior to newer NSCLC treatments). Those with an index date between January 1, 2003, and December 31, 2005, fell into the transition period. Transition signified emergence of biologics, TKIs, and pemetrexed as treatment options. Patients in the current period had an index date between January 1, 2006, and June 30, 2008. These patients were diagnosed when all classes of active treatment currently listed in the NCCN compendia were available options, with the exception of the ALK inhibitor crizotinib, approved after this study. Costs incurred by individual patients fell into the time period of their index date.

Study Variables

The primary variable was total direct cost-of-care per patient per month (PPPM). Secondary variables were cost per disease segment PPPM, effect of comorbidities on cost, current total and monthly cost for patients who had active treatment lasting at least 5 months (“gold standard” patients), percentage of each segment related to overall cost, and percentage of each active treatment class relative to active treatment segment cost and overall cost. Other variables included gender, age at index, comorbidity score (National Cancer Institute Combined Index),15 and region.

Statistical Analysis

Categorical data were compared using the χ2 test and reported as a percentage. We compared continuous data using the Wilcoxon rank-sum test and reported in median and interquartile range (IQR). Costs were reported in dol- lars, discounted to 2009 using published inflation rates. We estimated multiple least squares models of the log transformation of costs regressed on indicator variables for time period as well as patient baseline characteristics (age, gender, 4 region indicators, and comorbidity score) to account for possible changes in patient profiles over time.


Baseline Characteristics

A total of 969 patients met study criteria (27 patients in the pre-period, 439 patients in the transition period, and 503 patients in the current period). Due to the low number of patients in the pre-period, statistical analyses were limited to transition and current periods for comparison. Baseline characteristics (Table 1) demonstrate age at index (61 years in transition; 63 years in current), comorbidity score, region, length from index until death (9.9 months in transition; 8.8 months in current), and active treatment Cost of Lung Cancer Care length (4.8 months in transition; 4.2 months in current) all of which were all significantly different between transition and current groups at baseline.

Primary Objective

Overall, total monthly cost for the observation period was $10,281 PPPM (IQR $6613-$15,359). There was no significant difference in total monthly cost between transition and current periods (Table 2). The multiple regression model showed baseline confounders had no effect on total cost between the 2 groups. Age, region, and length were significant independent predictors of cost difference.

Cost per disease segment is listed in Table 2. Patients in the pre-period were included in overall cost description, but due to small sample size, were excluded from statistical analyses. Overall diagnosis cost was $6601 PPPM (IQR $3376-$11,881), with median length of 3.3 months. Active treatment cost was $9287 PPPM (IQR $5221-$16,463), with median length of 4.6 months. End-of-life cost was $12,215 PPPM (IQR $4958-$24,703) with a set length of 1 month. There were no significant differences between transition and current periods for any disease segment; adjustment for baseline confounders had no effect on this estimated cost difference. Significant predictors of costs were age (total and all segments of disease), gender (total and diagnosis), and region (total and all segments of disease), and comorbidity score of 2 or greater (diagnosis segment only).

Secondary Objectives

Overall Costs

Drug cost was 23% of overall cost in the transition period, and 21% in the current period. New drug classes represented 2% (biologics), 4% (TKIs), and 2% (pemetrexed) of overall cost in transition, and 5% (biologics), 2% (TKIs), and 4% (pemetrexed) of overall cost in current. Drugs given in the end-of-life segment are included in drug classes to prevent biasing of cost results. Non-drug costs were the highest component of overall cost for both transition (77%) and current (79%) periods.

Active Treatment Costs

PDC represented the greatest cost of active treatment classes at 19% in transition and 11% in current periods. Biologics represented 4% in transition and 9% in current; TKIs represented 6% in transition and 3% in current; and pemetrexed represented 3% in transition and 6% in current. Non-drug costs were the greatest active treatment costs, accounting for 62% in transition and 65% in current. These costs included hospitalizations, surgery, nursing care costs, radiation therapy, outpatient visits, and other costs that are not drug-related.

Active Treatment Usage

PDC was the most common regimen used (87% in transition, 89% in current). Use of biologics increased 3-fold from 6% in transition to 18% in current; TKI use decreased from 28% in transition to 17% in current; pemetrexed use increased from 13% in transition to 25% in current.

Effect of Comorbidities

PDF is available on the last page.
Diabetes Compendium
Oncology Compendium