The Oncologist, Vol. 11, No. 9, 1025-1033, October 2006; doi:10.1634/theoncologist.11-9-1025 © 2006 AlphaMed Press
Referral to Medical Oncology: A Crucial Step in the Treatment of Older Patients with Stage III Colon Cancera Sealy Center on Aging, b Department of Preventive Medicine and Community Health, and c Department of Internal Medicine, University of Texas Medical Branch at Galveston, Galveston, Texas, USA; d Department of Breast Medical Oncology, University of Texas M. D. Anderson Cancer Center, Galveston, Texas, USA Key Words. Medical oncologists • Consultation • Colon cancer • SEERMedicare • Chemotherapy Correspondence: James S. Goodwin, M.D., The University of Texas Medical Branch at Galveston, 301 University Boulevard, Galveston, Texas 77555-0460, USA. Telephone: 409-747-1987; Fax: 409-747-3585; e-mail: jsgoodwi{at}utmb.edu Received April 20, 2006; accepted for publication August 15, 2006.
Access and take the CME test online and receive 1 AMA PRA Category 1 CreditTM at CME.TheOncologist.com
Purpose. Adjuvant chemotherapy for stage III colon cancer produces a substantial survival benefit, but many older patients do not receive chemotherapy. This study examines factors associated with medical oncology consultation and evaluates the impact of such consultation on chemotherapy use. Patients and Methods. We used the Surveillance Epidemiology and End ResultsMedicare linked database and identified 7,569 patients, aged 6699, with stage III colon cancer diagnosed from 19921999. Modified Poisson regression was used to assess the relative risk for seeing a medical oncologist and for receiving chemotherapy as a function of individual characteristics. Results. 78.08% of patients saw a medical oncologist within 6 months of diagnosis. Patients who were female, white, married, had low comorbidity scores, were diagnosed in more recent years, or had four or more positive lymph nodes were more likely to see a medical oncologist. Patients seeing a medical oncologist were 10 times more likely to receive chemotherapy (odds ratio, 9.98; 95% confidence interval, 8.2112.14), after controlling for demographic and tumor characteristics. Chemotherapy use increased over time, but was substantially lower among older, black, and unmarried patients. Conclusions. Referral to medical oncology is one of the most important factors associated with receipt of chemotherapy among older patients with stage III colon cancer. Comorbidity decreases the likelihood of receiving chemotherapy, but its effect is the same for those who see a medical oncologist and all patients combined. Ensuring that high-risk patients are referred to medical oncology is a crucial step in quality care for patients with colon cancer.
The incidence of colon cancer increases markedly with advancing age [1]. A number of studies have shown that older individuals with cancer are at risk for receiving less than optimal care [24]. In particular, several investigators have found that substantial percentages of older individuals with stage III (node-positive) colon cancer do not receive chemotherapy [5, 6], and this is particularly marked among black elderly [57]. This is of great concern, because chemotherapy is associated with substantially better survival in stage III colon cancer [813], and the beneficial effects of chemotherapy are still apparent in older adults [8, 14, 15], at least up to age 80. For example, in randomized controlled trials, the 5-year survival rate for stage III colon cancer patients without chemotherapy was approximately 50%, whereas this rate was approximately 70% for those receiving chemotherapy [912, 15, 16]. In the U.S., chemotherapy is usually administered by medical oncologists. Colon cancer, as with many solid tumors, traditionally has been seen as a surgical disease; that is, the primary treatment is surgical. With the advent of effective chemotherapy, it became important that all patients also see a medical oncologist. However, shifts in practice patterns tend to occur slowly [17, 18]. We hypothesized that a major reason why individuals with stage III colon cancer might not receive chemotherapy is that they were not referred to a medical oncologist. While there have been substantial numbers of studies of factors associated with less than definitive treatment of cancer [25, 7], there have been few studies on factors associated with the appropriate evaluation of the newly diagnosed cancer patient. In this study we used the Surveillance Epidemiology and End Results (SEER)Medicare linked database to examine factors associated with medical oncology consultation and the impact of such consultation on chemotherapy use.
Data Sources Data for this study were obtained from three large databases, which were linked through collaborative projects and agreements involving the National Cancer Institute (NCI), the Centers for Medicare and Medicaid Services (CMS), and the American Medical Association (AMA) [1921]. These databases contain: (a) population-based tumor registry data obtained from the SEER program, (b) Medicare enrollment information and claims submitted to CMS for services provided to Medicare beneficiaries, and (c) information collected by the AMA on all physicians in the U.S. (AMA Master file).
Patients
Identification of Patients Who Saw a Medical Oncologist In the physician claims for our eligible patients, there was 65% agreement between the two sources on whether or not the physician was a medical or hematology oncologist. Based on a review by Baldwin et al. [21] of the quality of physician characteristics in the claims, the general recommendation has been to search both sources for evidence of a particular physician specialty. This is the approach we used in our study. Hence, if a patient had a physician claim within the 6-month period from the month of diagnosis and the physician specialty (primary or secondary) was medical oncology or hematology based on either the AMA or CMS data, we defined the patient as "having seen a medical oncologist."
Measures
Each patients preoperative comorbidity level was determined from claims data using the Charlson Comorbidity Index as adapted by Klabunde et al. [22, 23]. The claims data were also used to define receipt of chemotherapy based on Medicare codes for chemotherapy administration as described in a previous publication [24]. Patients were identified as "receiving chemotherapy" if they had one or more of these codes on any claim within 6 months after their colon cancer diagnosis. These codes included the International Classification of Diseases, Ninth Edition, clinical modification (ICD-9-CM) procedure code of 99.25 for a hospital inpatient or outpatient facility claim of chemotherapy (injection or infusion of cancer chemotherapeutic substance); the common procedure terminology codes of 96400 to 96549, J9000 to J9999, and Q0083 to Q0085 for a physician or out-patient claim of chemotherapy administration; and the revenue center codes of 0331 (chemotherapy injected), 0332 (chemotherapy oral), and 0335 (chemotherapy i.v.) for an outpatient claim of chemotherapy. The ICD-9-CM V codes of V58.1, V66.2, or V67.2 for follow-up examination or care after chemotherapy were also used [24].
Statistical Analysis Logistic regression for assessing the adjusted odds ratio is widely used when the event of a dichotomous outcome is rare[25,26].However, logistic regression may overestimate risk association when the probability of outcome is high (in this case, seeing medical oncologists). Therefore, this study used the modified Poisson regression model, which allows for adjusting the odds ratio so that relative risk can more easily be approximated [27]. A series of modified Poisson regression models was then constructed to estimate the likelihood of being seen by a medical oncologist after successively including groups of variables representing: (a) sociodemographic characteristics, (b) tumor characteristics and SEER registry area, and (c) comorbidity and contextual factors. Modified Poisson regression models were also generated for the likelihood of receiving chemotherapy. All statistical analyses were performed with Statistical Analysis Software version 9.0 for Windows XP (SAS Institute, Inc, Cary, NC).
Table 1
We next examined characteristics associated with seeing a medical oncologist. In bivariate analyses, younger patients, married patients, those with no or low comorbidity, and those with four or more positive lymph nodes were more likely to see a medical oncologist (Table 1
In Table 2
Of the 7,569 patients diagnosed with stage III colon cancer, 58.51% received chemotherapy initiated within 6 months of diagnosis, which we thus categorized as adjuvant chemotherapy. There was a 2.2% annual increase in the likelihood of receiving chemotherapy. As expected, the receipt of chemotherapy was closely tied to seeing a medical oncologist: 73.32% of patients seeing a medical oncologist received chemotherapy, while 5.79% of patients who did not see a medical oncologist received chemotherapy. Conversely, 97.83% of those who received chemotherapy saw a medical oncologist, while only 50.22% of patients who did not receive chemotherapy saw a medical oncologist ( 2 = 2,433.48, p < .0001).
In a multivariate analysis of chemotherapy use among all patients with node-positive colon cancer (Table 3
Other investigators have reported that older age, African-American ethnicity, and presence of comorbidity are associated with a lower likelihood of receiving chemotherapy after a diagnosis of stage III colon cancer [5, 7, 28, 29]. Because we found that these factors were also associated with a lower likelihood of seeing a medical oncologist, we next asked if disparities in referral to a medical oncologist were the explanation for the lower rates of chemotherapy in these patient groups (Table 3
In this population-based study of patients aged 66 years and older diagnosed with node-positive colon cancer, approximately 22% of patients did not see a medical oncologist within 6 months of diagnosis. As expected, seeing a medical oncologist was closely linked to receipt of chemotherapy. The chance of receiving adjuvant chemotherapy was 10 times higher for patients who saw a medical oncologist than for those who did not. Among patients who did not see a medical oncologist, only 6% received chemotherapy. Adjuvant chemotherapy has a large survival benefit in this population of patients with stage III colon cancer. The number needed to treat to prevent one death at 5 years is approximately five, which is substantially better than for most other instances of use of adjuvant chemotherapy. Moreover, randomized trials and cohort studies have not found a lower efficacy in patients of advanced age [8, 14]. Nevertheless, only 58.51% of the stage III colon cancer patients aged 66 and older received chemotherapy during this period, a finding made previously by other investigators [5, 14, 28, 29]. Our findings suggest that lack of referral to medical oncology is one of the most important factors in the undertreatment of older patients with colon cancer. Certain groups of patients were at particular risk for not seeing a medical oncologist. These included older patients, black patients, men, unmarried individuals, and those with three or more comorbidities. In addition, there was significant geographic variation in the likelihood of seeing a medical oncologist. These factors associated with seeing a medical oncologist are consistent with findings from studies in other cancers [30, 31], and also in a study of colon cancer patients treated in a single HMO [32]. Our findings were encouraging in that there was a temporal trend of an increasing likelihood of seeing a medical oncologist throughout the 1990s.
Many of the patterns of chemotherapy use were explained by the patterns of referral to medical oncology. For instance, the increase in patients with node-positive colon cancer seeing a medical oncologist over the 1990s appeared to result in the increase in chemotherapy use over the same time period. However, some disparities persisted. While the results showed that older individuals, black patients, and unmarried patients were less likely to see a medical oncologist (Table 2
Even though adjuvant chemotherapy produces substantial improvements in survival for node-positive colon cancer, it is not clinically plausible that all such patients should receive chemotherapy. For example, very old patients or those with other severe complex disorders possess life expectancies that change the riskbenefit analysis [33]. In that regard, it is interesting that seeing a medical oncologist had little effect on the likelihood of those with multiple comorbidities receiving chemotherapy (Table 3 Our study has some limitations. First, the potential for inaccurate coding and missing values exists for any claims-based study, and clinical information available from billing data is not as detailed as that available from chart review [34]. Second, while the SEERMedicare database is an excellent source to study older patients diagnosed with colon cancer, results are limited in their generalizability to younger populations across the U.S. In addition, the population in the SEERMedicare database is somewhat more ethnically diverse, has a higher percentage of urban residents, is more highly educated, and has a higher income than the general older population [19, 35]. Also, this study does not include information on factors such as physician and health system characteristics or patient preferences that might influence referral to a medical oncologist or receipt of chemotherapy, particularly if patients are not interested, willing, or able to receive chemotherapy. In order to design effective interventions, it is important to delineate the mechanisms producing less than optimal care. This involves describing the trajectory of care in increasing detail and examining the impact at each step in that trajectory of factors associated with less than optimal care [36]. Identifying the precise step(s) whereby certain patient characteristics are associated with less than optimal care should assist in the design of targeted, cost-effective interventions.
The current study represents a first step in that process, exploring whether a major block in receipt of chemotherapy is at the level of referral to a medical oncologist. As mentioned earlier, colon cancer, like virtually all solid tumors, has traditionally been viewed as a "surgical disease," that is, the primary therapy for localized cancers is surgical resection. Thus all or almost all patients are seen by a surgeon early in the diagnostic pathway, and surgeons are regarded as the primary treating physician. With the advent of effective adjuvant chemotherapy for some localized cancers (e.g., breast and colon cancer), it has become important to involve medical oncologists in the evaluation and treatment of such patients. Such a process can be easily accomplished in the context of multidisciplinary cancer centers, but becomes more challenging in community practice [37]. Clearly, interventions to increase appropriate use of chemotherapy in stage III colon cancer should focus on increasing the percentage of such patients seen by a medical oncologist. It is reassuring that this percentage increased during the 1990s (Table 1
The authors indicate no potential conflicts of interest.
This study used the linked SEERMedicare database. The interpretation and reporting of these data are the sole responsibility of the authors. The authors acknowledge the efforts of the Applied Research Program, NCI; the Office of Research, Development and Information, CMS; Information Management Services (IMS) Inc.; and the SEER program tumor registries in the creation of the SEERMedicare database. This research was supported by the UTMB Center for Population Health and Health Disparities (P50 CA105631) and grants from the National Cancer Institute (R01CA104949) and the Agency for Healthcare Research and Quality (R24HS011618). Dr. Giordano is supported by NIH 1K07 CA 109064-02.
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||