Medicare spent nearly a billion dollars in 2002 on chemotherapy for newly diagnosed breast, colorectal, and lung cancer patients alone, but surprisingly little is known about the extent to which cancer chemotherapies help or harm such elderly patients. This unsettling paradox is the direct result of the well-described under-enrollment of elderly on the clinical trials of chemotherapy. In the absence of trials with representative patients, treating oncologists, patients, and policy-makers are left to extrapolate results of clinical trials that were conducted on young and comparatively healthy individuals to the general population with cancer, who older and have greater comorbid disease. Surprisingly, basic results of this common extrapolation are unknown. Practitioners lack even basic descriptive studies reporting survival of elderly Medicare beneficiaries following receipt of most standard chemotherapy regimens. We seek to remedy this critical gap in clinical knowledge by studying the survival outcomes of both unselected elderly Medicare beneficiaries and their carefully chosen clinical trial counterparts following receipt of the same standard chemotherapy regimens. We do so by integrating--in a novel way--both clinical trial and observational data sources and methods, and we consider four common cancers among the elderly: breast, colorectal, lung, and pancreas cancer.

Our five Aims acknowledge that the survival outcomes following receipt of a given chemotherapy regimen may depend both on the type of cancer patient who is treated and the setting in which the treatment is delivered and these Aims are to:

* Describe survival outcomes of site- and stage-specific cohorts of elderly Medicare beneficiaries following treatment with first-line "standard" chemotherapy regimens in the usual care setting.

* Compare the survival outcomes following diagnosis of site- and stage-specific cohorts of elderly Medicare beneficiaries who were treated with "standard" chemotherapy regimens in Aim 1 to that of untreated, but otherwise matched similar elderly Medicare beneficiaries with cancer.

* Compare the attributes of the site- and stage-specific cohorts of elderly Medicare beneficiaries who receive first-line treatment with one of several "standard" chemotherapies in the usual care setting to those of elderly Medicare beneficiaries treated with nominally identical therapies in the clinical trial setting.

* Compare the survival outcomes of both types of site- and stage-specific cohorts of elderly Medicare beneficiaries from Aim 3 after first-line "standard" chemotherapies according to treatment setting.

* Applying data, cohorts, and modeling approaches from Aims 1-4, develop clinically useful prediction models that yield regimen-specific estimates of survival outcomes for community-based elderly patients as a function of treatment and patient and provider factors available in the dataset we construct.

This work has both a clinical and an evaluative goal. Clinically, oncologic decision-making sits on a razor's edge. Highly cytotoxic drugs are given to patients who are often already gravely sick. The decision to offer such therapy -- and to accept it -- depends on careful, individualized weighing of risks and benefits, preferences, and probabilities. Yet neither oncologists nor their patients currently have access to accurate information for the majority of patients who are elderly and treated in the community. This project will substantially narrow this critical information gap. At the same time, evaluatively, we seek to measure the degree to which clinical practice in oncology is meeting the potential of many NCI-sponsored clinical trials. The gap between knowledge and practice is known to be wide in many areas, and is a major focus of the NIH Roadmap. By explicating the separate role of patient and provider factors in explaining chemotherapy outcomes among the elderly, we will begin to identify the barriers. We will also identify targets for intervention to insure the optimal cancer care for elderly patients.