Working Group 1: Patterns of Care

Speaker:
Vincent Mor, Ph.D., Brown University

Co-Chairs:
Paul F. Engstrom, M.D., Fox Chase Cancer Center
Jerome W. Yates, M.D., M.P.H., Roswell Park Cancer Institute

Introduction

Cancer is an age-related disease with peak incidence and mortality rates in the segment of the population 65 years and older. Thus, it is surprising that so little descriptive information is available about cancer treatment in older persons. Not only is limited information available on the kinds of cancer care older patients receive in NCI-designated cancer centers, but few data are collected in a systematic manner prospectively or retrospectively in community cancer treatment settings, in population-based studies, and in case-control studies.

There is a dearth of information on cancer care in older patients. No evidence-based or clinically-based consensus exists on how older patients who are newly diagnosed with a malignancy should be treated. Moreover, no recommended or standardized scientific approaches are available to specify the appropriate treatment for patients with age-related health problems concurrent with their cancer (i.e., comorbidity). The critical knowledge gap in this area needs urgent attention.

Working Group 1 addressed the need for information on how to identify older persons who are likely to benefit from treatment that is known to be effective in younger populations. Such information must be communicated to the research and practice communities.

In his introductory presentation, Dr. Vincent Mor gave an overview of the cancer treatment of older patients from the small number of studies conducted in the 1990s on breast, prostate, colorectal, and lung cancers. Articles that did not examine age as a predictor of care received were not included. Dr. Mor's review provided the following insights on age differences:

• Across tumors, diagnostic intensity (e.g., the probability of receiving a referral to a "chest physician") declines with advancing age.
• Older cancer patients tend to receive less aggressive treatment (e.g., they are less likely to receive radiation therapy with or without chemotherapy after surgery for colon cancer).
• When appropriate treatment has been identified for a particular tumor, older patients are less likely to receive that standard of care (e.g., breast cancer surgery plus radiation).
• Variations identified in treatment for older patients may stem from differences in physician training, geographic location, academic versus community hospital care and treatment, and rural and urban settings.
• Physician bias is evident in these studies, which suggests that treatment differences and exclusion are not based only on risk assessment and/or patient/family preferences.

Dr. Mor indicated that, in the studies examined in this review, the extent to which the preexisting health status of the newly diagnosed cancer patient (i.e., comorbidity) contributes to diagnostic and treatment variations and the higher observed rates of cancer mortality among aged cancer patients is not known. Unfortunately, obtaining detailed information on the physiological condition and the presence of comorbidities in older patients is extremely difficult. Participants in Working Group 1 agreed that any prospective study developed and supported by the NCI and NIA on cancer in older persons must include some reasonable medical measures of cancer patient comorbidities.

Research Questions

1. Can patterns-of-care studies assess the effects of comorbidity as it relates to treatment and care decisions?
2. Is multidisciplinary care more likely to address age factors in older patients?
3. Does treatment for older patients differ in cancer centers and the community setting?
4. Can comorbidity profiles be devised to characterize elderly patients who are more likely to benefit from those therapies known to be effective in younger populations?
5. At what age and stage is effective treatment less beneficial for aged patients (i.e., benefits of treatment are remote)?

Cancer Center Role

In the NIA/NCI planning meeting for the cancer centers workshop in which working group themes were selected, it was acknowledged that information on older patients treated in cancer centers might not be generalizable to all older persons. Therefore, to facilitate population-based patterns-of-care studies, cancer center leaders and experts could collaborate with the NCI Surveillance, Epidemiology, and End Results (SEER) Program leaders, as almost all NCI SEER registries are located in states or regions in which NCI-designated cancer centers are located.

In addition, the NCI and other federal agencies have increased their efforts to improve the quality of cancer care by enhancing their data collection systems and linkages among systems. Although these efforts do not focus directly on older patients, their descriptive and analytical research targets certain tumors (e.g., colorectal and lung cancers) that are among the major malignancies that affect the elderly disproportionately and will yield useful data on the elderly. For example, the NCI recently announced awards for the Cancer Care Outcomes Research and Surveillance (CanCORS) Program (October 2001). Awards totaling approximately $34 million within a 5-year period will be made through cooperative agreements with the NCI Division of Cancer Control and Populations Sciences. These projects will assess community practice patterns and disparities in care for population subgroups with colorectal and lung cancers. A research infrastructure will support six teams of scientific investigators and a coordinating center to collect common data elements describing the processes and outcomes of care for all enrolled patients through medical record abstracts, surveys, and administrative data. The data-coordinating center and one of the six research groups are cancer centers. Given the large number of older persons afflicted by colorectal and lung cancers, these new studies are likely to focus on older patients, at least to some extent.

Research Priorities

The priorities of Working Group 1 address the infrastructure for conducting retrospective and prospective patterns-of-care studies to elucidate what is occurring in the cancer treatment arena for older patients. The recommendations do not identify major research areas or malignancies that deserve in-depth pursuit in patterns-of-care studies or suggest study designs. But the priorities do call for groundwork to support potential investigations that could be conducted in partnership with cancer centers.

1. Develop a key data matrix of content area items from relevant databases in the public domain that are pertinent to conducting patterns of quality cancer care research on the treatment of older patients. Some suggested databases include NCI SEER, Medicare Managed Care, Insurance Encounter data, American College of Surgeons National Cancer Data Base, Veterans Administration studies, and Minimal Data Set reports in skilled nursing facilities.
   • This effort should begin with existing studies to ascertain what might be applied to a cancer center study effort.
   • The complexity and expense of conducting patterns-of-care studies were discussed. A major obstacle is that available data on treatment effectiveness do not include specific information on older cancer patients, and large prospective studies on the aged include few, if any, details on cancer.
   • Supplemental data are required to make the existing data collection efforts useful and applicable to older patients. Related issues include sociodemographic and ethnic disparities, continuity of care, and the best models for delivery of care.
   • Networking with the community is essential for patterns-of-care studies. Few collaborations of this nature have occurred, but compelling reasons exist for cancer center oncologists and community cancer care physicians to work together.
2. Develop a dictionary of data elements (i.e., a compendium of items and their intended use) as a resource for investigators. Although crafting geriatric assessments anew is not desirable, instruments are available that may be applicable to older cancer patients. Whether some instruments already exist that could be applied at the aging/cancer research interface should be determined.
   • To facilitate future compilations of information on older patients, a computer-accessible data set should be developed that indicates which items are in which set of survey study information.
   • Development of a progressive database requiring inclusion of certain items should be considered.
   • Potential should be built for aggregation of patterns-of-care data across selected cancer centers.
3. Incorporate the clinical expertise from NCI projects, particularly the NCI SEER projects, that is available in cancer centers to improve the quality of care of the medically underserved, aging population. Cancer surveillance efforts should be enhanced to improve the clinical information base on cancer in older persons in SEER. NCI-designated cancer centers and NCI SEER registries that facilitate special studies on cancer in the elderly should be identified in strategic geographic locations.
   • Whether patterns of care for older patients treated in cancer centers are different from patterns of care for other older patients with the same types of cancer who are treated in the community should be ascertained.
   • Cancer center consortia should be developed to address treatment efficacy and tolerance for selected tumors that primarily affect older patients.

Research Barriers

• Heterogeneity of cancer centers is a significant barrier—treatment environments vary greatly among cancer centers.
• More uncommon malignancies are seen in cancer centers than in other settings; common cancers tend to be managed in the community.
• Patients treated in cancer centers tend to be younger.
• Cooperation with local physicians in the community is required, so cancer centers must facilitate networking. Although this is not an unachievable goal, it requires a strategic collaborative planning effort on the part of both center and community sectors.

Mechanisms

• Encourage geographically compatible Community Clinical Oncology Programs (CCOPs) and NCI-designated cancer centers to conduct consortium studies.
• Organize demonstration projects within cancer centers and community settings to enhance comorbidity assessment and measurement of older patients' health problems concurrent with their malignancies.
• Consider conducting patterns-of-care studies in certain cancer centers (e.g., possibly those in New Hampshire, New Mexico, Iowa, Wisconsin, North Carolina, and Colorado) that serve reasonably well-defined catchment areas, if certain study criteria can be met.
• Consider supplemental funding for a task force to implement some patterns-of-care activities, such as funding a short-term, centralized cancer center patient data system.
• Develop incentives to facilitate study teams of geriatricians and oncologists to create model protocols for testing efficient geriatric assessments of cancer patients. These instruments may later be used in large-scale patterns-of-care studies.