Fax: (212) 659-5599
Prevalence and characteristics of patients with metastatic cancer who receive no anticancer therapy
Article first published online: 15 JUN 2012
Copyright © 2012 American Cancer Society
Volume 118, Issue 23, pages 5947–5954, 1 December 2012
How to Cite
Small, A. C., Tsao, C.-K., Moshier, E. L., Gartrell, B. A., Wisnivesky, J. P., Godbold, J. H., Smith, C. B., Sonpavde, G., Oh, W. K. and Galsky, M. D. (2012), Prevalence and characteristics of patients with metastatic cancer who receive no anticancer therapy. Cancer, 118: 5947–5954. doi: 10.1002/cncr.27658
- Issue published online: 19 NOV 2012
- Article first published online: 15 JUN 2012
- Manuscript Accepted: 20 APR 2012
- Manuscript Revised: 18 APR 2012
- Manuscript Received: 13 MAR 2012
- metastatic cancer;
- health care disparities;
- National Cancer Data Base;
- no treatment
A subset of patients who present with metastatic solid tumors never receive anticancer therapy. Reasons may include poor functional status, comorbidities, and patient preference. To the authors' knowledge, the prevalence and characteristics of this population have not previously been described.
The National Cancer Data Base was queried for patients diagnosed with metastatic (stage IV according to the American Joint Committee on Cancer) solid tumors (including those of the breast, cervix, colon, and kidney; small cell and nonsmall cell lung cancer [NSCLC]; and tumors of the prostate, rectum, and uterus) who received neither radiotherapy nor systemic therapy. Log-binomial regression analysis was used to estimate prevalence ratios (PRs) for the percentage of untreated to treated patients with stage IV cancer.
Between 2000 and 2008, 773,233 patients with stage IV cancer were identified, 159,284 of whom (20.6%; 95% confidence interval, 20.5%-20.7%) received no anticancer therapy. Patients with NSCLC accounted for 55% of the untreated population. Patients with cancers of the kidney and lung had the highest rates of no treatment at 25.5% and 24.0%, respectively, whereas patients with prostate cancer had the lowest rate of no treatment at 11.1%. Across all cancer types, older age (PR range, 1.37-1.49; all P < .001), black race (PR range, 1.05-1.32; all P < .001), lack of medical insurance (PR range, 1.47-2.46; all P < .001), and lower income (except for cancer of the uterus; PR range, 0.91-0.98 for every $10,000-increase in income [all P < .001]) were associated with a lack of treatment.
Approximately 20% of patients who present with stage IV solid tumors do not receive anticancer therapy. Although there are likely multiple reasons for this lack of treatment, including appropriate indications, these findings have potential implications with regard to health care policy and access to care. Cancer 2012. © 2012 American Cancer Society.
Clinicians caring for patients with cancer are well aware that a subset of patients who present with metastatic solid tumors never receive anticancer therapy. Although advanced solid tumors are generally not curable with current therapeutic modalities, the use of systemic chemotherapy and radiotherapy has been associated with improvements in cancer-related symptoms, and prolongation of life, across multiple tumor types.1, 2 Despite these benefits, treatment may be forgone for several reasons including comorbidities, poor functional status, lack of access to care, and personal preference.3, 4
To the best of our knowledge, there has been scant literature describing the prevalence and characteristics of patients presenting with advanced cancer who receive no anticancer therapy. In 1 such study of 6308 patients diagnosed with stage IV nonsmall cell lung cancer (NSCLC) between 1991 and 1993, only 21.5% of patients received systemic therapy.5 However, since the early 1990s, additional data have been generated to support the use of chemotherapy for advanced cancer and more tolerable treatment regimens have been developed.6 A recent analysis of patients with urothelial cancer reported that 23.8% of patients with metastatic disease received no systemic therapy.7 Studies of large populations of patients with cancer, across various tumor types, have to our knowledge not been performed previously to estimate the magnitude of potential undertreatment. Therefore, we sought to analyze the treatment patterns of patients presenting with metastatic cancer across 9 common cancer types using a contemporary cohort from a national cancer registry database.
MATERIALS AND METHODS
Data for this study were obtained from the National Cancer Database (NCDB) Public Benchmark Reports (http://cromwell.facs.org/BMarks/BMPub/Ver10/bm_reports.cfm) from 2000 through 2008. The NCDB is a national, hospital-based cancer registry jointly sponsored by the American College of Surgeons and the American Cancer Society. Since 1996, the Commission on Cancer (CoC) has required CoC-accredited programs to report cancers diagnosed or treated at their facilities to the NCDB. As of 2001, data collection has been restricted only to CoC-accredited programs, including > 1400 academic centers, cancer centers, community hospitals, and Veterans Affairs (VA) hospitals. The NCDB contains standardized data elements regarding patient demographics, tumor characteristics, and the first course of treatment. In addition, the NCDB contains information concerning patient insurance status, geographic location of residence, type of treatment facility, and income level. The NCDB Public Benchmark Reports provide access to aggregate-level data categorized by multiple variables.8 Previous studies have shown concordance between the NCDB data set and other national population and cancer registries for ethnicity distribution, age distribution, insurance status, and stage at diagnosis.9 Because no patient, provider, or hospital identifiers are included in the research data set and no protected health information is present, Institutional Review Board approval was not required for this study.
Study participants were diagnosed with American Joint Committee on Cancer stage IV lung (NSCLC or small cell lung cancer [SCLC]), breast, kidney, prostate, colon, rectum, cervical, or uterine cancer and did not receive any first-course therapy. These cancers represent 5 of the most common solid tumors diagnosed in men and women in the United States.10 The label “no first-course therapy” was selected to designate patients who refused any type of intervention, who died before the initiation of treatment, or for whom the treating physician recommended no treatment be given. Importantly, the “no first-course therapy” designation is actively selected by the cancer registrar and excludes patients with other or unknown first-course therapies.
Patient Demographics and Socioeconomic Indicators
Factors potentially associated with a lack of treatment were classified into 4 categories: clinical characteristics, patient-level demographics, area socioeconomic indicators, and facility characteristics. Data regarding stage of disease, year of diagnosis, and first-course therapy were extracted. Age at diagnosis was classified as a categorical variable using the following age groups: aged < 20 years to 49 years, aged 50 years to 59 years, aged 60 years to 69 years, aged 70 years to 79 years, aged 80 years to 89 years, and aged > 90 years. Race was defined by Facility Oncology Registry Data Standards (FORDS) criteria and categorized as white, black, Asian/Pacific Islander (API), and other/unknown. In terms of ethnicity, patients were classified as Hispanic or non-Hispanic. Primary payer/insurance type at the time of diagnosis was also defined by FORDS criteria and was grouped into the following categories: private insurance plans (health maintenance organization, preferred provider organization, managed care, or private insurance), Medicaid, Medicare (Medicare alone or Medicare with supplement), uninsured (not insured, insurance not otherwise specified, charity write-off, and self-pay), and military (VA or military/TRICARE).
Area-based indicators of patient socioeconomic status, specifically education and income, were derived by the NCDB from US census data at the ZIP code level, and divided into 5 equally proportional categories based on the observed distribution. The percentages of the population in the patient's ZIP code of residence without a high school degree were coded into the following groups: > 29%, 20% to 28.9%, 14% to 19.9%, and < 14%. The median household income for the patients' ZIP code of residence was stratified into the following groups: < $28,000, $28,000 to $32,999, $33,000 to $38,999, $39,000 to $48,999, and > $49,000.
The primary objective of the current study was to define the percentage of patients with stage IV cancer who received no anticancer therapy. An additional objective was to examine patient characteristics associated with a lack of treatment. Log-binomial regression analysis was used to estimate prevalence ratios (PRs) and corresponding 95% confidence limits (95% CL) relating the percentages of patients receiving no first-course therapy versus any therapy for stage IV cancer among categories of various socioeconomic and demographic variables. Statistical significance was defined as P < .05. All statistical analyses were performed using SAS statistical software (Version 9.2; SAS Institute Inc, Cary, NC).
Between 2000 and 2008, a total of 5,395,382 patients with breast, cervical, colon, and kidney cancer; NSCLC; SCLC; and prostate, rectal, and uterine cancer were identified in the NCDB. A total of 4,173,128 patients with stage 0 to III disease or an unknown disease stage were excluded (Fig. 1). Overall, 773,233 patients with stage IV cancers remained for the analysis. The distribution of stage IV cancers by primary tumor site is shown in Table 1 and illustrated by Figure 2.
|Untreated patients with stage IV, no.||12,079||297||87,400||21,085||4305||18,816||6582||7313||1407||159,284|
|Total with stage IV, no.||47,417||1217||353,748||99,206||26,140||119,748||59,074||57,148||9535||773,233|
|95% CL||25.1%- 25.9%||22.0%- 26.9%||24.6%- 24.9%||21.0%- 21.5%||16.0%- 16.9%||15.5%- 15.9%||10.9%- 11.4%||12.5%- 13.1%||14.1%- 15.5%||20.5%- 20.7%|
|Median rate of no high school education by ZIP code|
|Median income by ZIP code|
The distribution by stage at presentation, categorized by primary tumor site, is shown in Figure 3. The median age of patients with stage IV cancer was 60 years to 69 years. The lowest median age group was for patients with cervical cancer (aged 50 years-59 years) and the highest median age group was for patients with kidney cancer (aged 70 years-79 years). Overall, 32.6% of the patients with stage IV disease had private insurance (252,176 patients), 50.3% had Medicare (388,834 patients), 6.0% had Medicaid (46,260 patients), and 4.4% were uninsured (33,715 patients). The percentage of uninsured patients in the NCDB study population was greatest for cancers of the cervix (12.1%) and breast (5.8%) and was lowest for kidney (3.8%), colon (3.9%), and prostate (3.9%) cancer. The majority (621,197 patients; 80.3%) of patients with stage IV cancer were white. Overall, 12.5% (96,763 patients) of patients were black, 3.7% (28,713 patients) were Hispanic, and 1.9% (14,720 patients) were API.
Percentage of Untreated Patients
Of the 773,233 patients with stage IV cancers, 159,284 (20.6%) received no anticancer therapy. The percentage of patients with stage IV tumors receiving no anticancer therapy, categorized by the site of the primary tumor, is detailed in Table 1. Patients with lung cancer accounted for 68% (108,485 patients) of the total of untreated patients with stage IV tumors. Patients with cancers of the colon (18,816 patients; 11.8%), kidney (12,079 patients; 7.6%,), breast (7313 patients; 4.6%), prostate (6582 patients; 4.1%), and rectum (4305 patients; 2.7%) accounted for the remaining majority of untreated patients (Fig. 2). Kidney cancer, NSCLC, and uterine cancer had the highest rates of no treatment at 25.5% (95% CL, 25.1%, 25.9%), 24.7% (95% CL, 24.6%, 24.9%), and 24.7% (95% CL, 22.0%, 26.9%), respectively, whereas prostate and breast cancer had the lowest rates at 11.1% (95% CL, 10.9%, 11.4%) and 12.8% (95% CL, 12.5%, 13.1%), respectively.
Patient Characteristics and Association With Lack of Treatment
Tables 2 and 3 present the distribution by patient age, gender, race/ethnicity, insurance status, and socioeconomic status of the untreated population with stage IV disease categorized by primary tumor site. Among the untreated patients, the majority (33.1%) were aged 70 years to 79 years. Overall, 20.8% of the untreated patients with stage IV disease had private insurance (33,108 patients), 61.6% had Medicare (98,138 patients), 5.3% had Medicaid (8403 patients), and 4.8% were uninsured (7639 patients). Among the untreated patients with stage IV disease, 79.5 % were white (126,676 patients), 13.1% were black (20,790 patients), 4.0% were Hispanic (6326 patients), and 1.8% were API (2853 patients).
There were significant differences in the percentage of untreated stage IV cancers across gender, age, and racial categories (Fig. 4). Females had a higher likelihood of being untreated compared with men for kidney cancer (28.2% vs 24.0%; PR, 1.18 [P < .001]), SCLC (21.6% vs 21.0%; PR, 1.03 [P = .035]), rectal cancer (18.5% vs 15.2%; PR, 1.22 [P < .001]), and colon cancer (16.1% vs 15.3%; PR, 1.06 [P < .001]). Older patients were more likely to be untreated than younger patients across all cancer types. For every 10-year increase in age, the percentage of untreated patients increased by 35% (NSCLC: PR, 1.35; 95% CL, 1.34, 1.36 [P < .001]) to 49% (colon: PR, 1.49; 95% CL, 1.47, 1.51 [P < .001]). Blacks were significantly more likely to be untreated for stage IV cancer than whites for all tumors except uterine cancer and SCLC (PR range, 1.05-1.32; all P < .001). Likewise, among patients with kidney cancer, NSCLC, SCLC, and prostate cancer, Hispanics were significantly more likely to be untreated than whites (PR range, 1.15-1.41; all P < .001). Across all cancer types, patients with no insurance (PR range, 1.90-2.19; all P < .01), Medicaid (PR range, 1.12-1.83; all P < .01), or Medicare (PR range, 1.73-2.45; all P < .001) were significantly more likely to be untreated than patients with private insurance.
There were statistically significant differences with regard to the percentage of untreated patients with stage IV cancer by median income and educational status (Table 3). For every 5% increase in the population without a high school education living in a patient's ZIP code, the percentage of untreated patients increased significantly (PR range, 1.02-1.09; all P < .001 except uterus). Conversely, for every additional $10,000 of median income for the population living in a patient's ZIP code, the percentage of untreated patients decreased significantly (PR range, 0.91-0.98; all P < .01 except uterus).
On the basis of the current analysis, representing 773,233 patients diagnosed with 8 common solid tumors in the United States over an 8-year period, nearly 1 in 5 patients presenting with stage IV solid tumors never receive anticancer therapy. In a hypothesis-generating analysis, we also observed that older age, racial and ethnic minority status, lack of medical insurance, and lower income were all associated with an increased likelihood of not receiving treatment.
The percentage of untreated patients varied by primary tumor site from 11.1% to 25.5%. Not surprisingly, patients with breast and prostate cancer were the least likely to remain untreated, which may be related to the high rates of screening for these cancers11 and the relatively tolerable hormonal therapies available for the treatment of metastatic disease. Patients with NSCLC had the highest rate of being untreated and accounted for > 50% of untreated patients during this time period. The percentage of untreated patients increased with age across all tumor subtypes. These findings are not unexpected because age-related comorbidities and decline in functional status have clear implications with regard to the risk-to-benefit ratio of systemic therapy.12 A recent study by Wang et al13 indicated that, in veterans aged > 65 years with NSCLC, advanced age was an even stronger predictor than comorbidity for receiving treatment. In this population, a high percentage of patients with metastatic disease did not receive treatment (66% of patients aged 65 years-74 years, 77% of patients aged 75 years-84 years, and 90% of patients aged > 85 years).13 These results further highlight the potential disconnect between the “efficacy” of therapies in the setting of clinical trials and the “effectiveness” of such therapies when applied to the general population of patients with advanced cancer.
There are likely multiple reasons why patients with metastatic cancer may not receive anticancer therapy, including appropriate reasons such as poor functional status, comorbidities, and patient preference. Furthermore, there are likely several factors that contribute to patients presenting with advanced, rather than early stage, cancer. Tumor biology, with rapid disease onset and progression, may play a predominant role in some patients. However, racial/ethnic minority status and/or poor access to care have previously been identified as factors associated with an increased likelihood of presenting with an advanced stage of disease at diagnosis.14-16 For example, Halpern et al demonstrated that nonwhite patients who were without insurance or with Medicaid were at a significantly increased risk of presenting with advanced stage disease for breast, colorectal, bladder, prostate, uterine, thyroid, kidney, and lung cancers.9 A propensity score analysis of patients treated at VA hospitals versus fee-for-service Medicare patients indicated that VA patients had equivalent or better survival rates for colon cancer, NSCLC, SCLC, lymphoma, and multiple myeloma, suggesting potentially superior preventive care or cancer screening.17 Other studies have observed decreased screening rates in racial minorities, underinsured patients, and the elderly.18, 19 Taken together, these findings suggest that certain patient groups may not only be more likely to present with cancers at later stages, but may also be less likely to receive any anticancer therapy whey they are ultimately diagnosed.
It is interesting to note that the development of better tolerated and/or more effective treatments may have some impact on the number of patients who are untreated. For example, in the current study, the percentage of untreated patients with metastatic kidney cancer was 26.8% (CL, 23.4, 29.3) in the pre-vascular endothelial growth factor receptor tyrosine kinase inhibitor (VEGFR TKI) era (2000-2005), and decreased to 22.8% (CL, 21.3, 24.0) in the VEGFR TKI era (2006-2008). However, even with the introduction of better therapies, there remains a large percentage of untreated patients.
A strength of the current analysis is the large data set representing a diverse population of patients and the focus on a contemporary time period. In addition, because cancer registrars actively report “no first-course treatment” to the NCDB, the assignment of patients to the “untreated” category was not simply based on an absence of data regarding treatment. The trends observed, particularly with regard to the increased percentage of untreated patients with advancing patient age, add some further confidence to the validity of the data. However, the current study has several limitations. Foremost, we used the NCDB Public Benchmark Reports and did not have access to individual patient data. Therefore, we could not perform multivariable analysis to determine which potentially related variables associated with an increased prevalence of untreated metastatic cancer retained independent significance. Therefore, the findings regarding patient characteristics associated with an increased likelihood of being untreated can only be considered hypothesis-generating. Further evaluation using individual patient data are warranted. Although the NCDB includes data for >75% of new cancer cases in the United States annually, CoC-approved hospitals often provide more cancer-related services (including cancer screening, chemotherapy, and radiotherapy) than non-CoC-accredited hospitals and therefore these findings may not be generalizable to non-CoC-accredited hospitals and may, in fact, be an underestimation of the prevalence of this group. Finally, the current study focused on patients with advanced/metastatic disease at the time of the initial cancer diagnosis and the percentage of untreated patients with metastatic disease may be substantially lower for patients initially presenting with earlier stage disease and later developing metastatic recurrence.
In summary, a large percentage of patients presenting with advanced solid tumors at the time of their initial cancer diagnosis receive no anticancer therapy. Further studies are necessary to better define the reasons for this lack of treatment, particularly because these relate to access to care, health literacy, and treatment disparities.
Funding received from the Doris Duke Charitable Foundation.
CONFLICT OF INTEREST DISCLOSURE
Dr. Wisnivesky is a member of the research board of EHE International. He has received lecture fees from Novartis Pharmaceuticals, acted as a consultant for United BioSource Corporation, and received a chronic obstructive pulmonary disease grant from GlaxoSmithKline.
- 72012 (suppl 5; abstr 285)., , , et al. Frequency of cisplatin administration in patients presenting with advanced urothelial carcinoma in the community. Clin Oncol. 30,
- 11Cancer screening in the United States, 2012: a review of current American Cancer Society guidelines and current issues in cancer screening [published online ahead of print January 19, 2012]. CA Cancer J Clin. doi: 10.3322/caac.20143., , .
- 13Impact of age and comorbidity on nonsmall-cell lung cancer treatment in older veterans [published online ahead of print March 26, 2012]. J Clin Oncol., , , , , .