Performing Economic Evaluations Alongside of Cancer Clinical Trials

July 1, 1998

Health care providers and financing organizations have become more aware of the resource constraints on the provision of medical services, thus increasing the importance of economic evaluations within the health care industry.[1,2] This has carried over to the evaluation of new, therapeutic strategies for cancer, which have traditionally been evaluated exclusively for safety and clinical efficacy.

Health care providers and financing organizations have become more aware of the resource constraints on the provision of medical services, thus increasing the importance of economic evaluations within the health care industry.[1,2] This has carried over to the evaluation of new, therapeutic strategies for cancer, which have traditionally been evaluated exclusively for safety and clinical efficacy.

The National Cancer Institute (NCI) and the American Society for Clinical Oncology (ASCO) have recognized the growing need for economic information. To address this issue, they held a joint economic workshop in 1996.

The outcome of the workshop was the development of a workbook outlining procedures and guidelines for performing economic evaluations alongside cancer clinical trials.

As a result of these efforts, guidelines and procedures are currently being developed and applied to appropriately evaluate both the economic and clinical impact of new cancer therapies in NCI-sponsored clinical trials.

Within a clinical trial, economic evaluations address the same issues as clinical investigations, but approach them from a different perspective. While the clinical protocol may be concerned with the results of a magnetic resonance imaging (MRI) scan, for example, the economic protocol will be concerned with the number of MRI scans conducted during the study period.

Evaluation of resource utilization data often requires specific assessments that account for the longitudinal nature of economic data.

Costs that can be included in an economic evaluation of cancer care consist of direct medical, direct nonmedical, indirect, and intangible costs.[3]

Direct medical costs represent costs incurred in providing care, such as payment for a chemotherapy agent or for seeing an oncologist.

 Direct nonmedical costs represent costs incurred because of illness or the need to seek medical care, and are generally paid out-of-pocket by patients and their families (eg, transportation or hotel stays required for medical treatment, or purchase of cosmetic apparel after disfiguring cancer treatment).

 Indirect costs represent costs not associated with transactions for goods or services, such as morbidity (eg, time lost from work) or mortality (eg, premature death leading to removal from the work force).

Intangible costs represent the costs of pain, grief, and suffering.

There are four types of analyses performed in clinical economics: cost-identification, cost-effectiveness, cost-utility, and cost-benefit.

A cost-identification analysis simply assesses the costs involved in medical care without regard to clinical outcome. Typically, this type of analysis is conducted for procedures or therapies with equivalent clinical efficacy.

 A cost-effectiveness analysis compares both the costs and outcomes of an intervention, with effectiveness measured in any meaningful clinical unit, such as years of life saved or number of toxic side effects prevented.

 When a medical intervention can result in several outcomes, the outcomes can be assessed in terms of patient utility. Utility is a measure of the patient’s preferences for a particular health state or for the outcome of an intervention. A cost-utility analysis compares the costs and benefits of care with the benefits measured as patient utility.

A cost-benefit analysis compares the cost of a medical intervention with its benefit by measuring costs and benefits in the same units (usually dollars), allowing for calculation of the ratio of dollars spent to dollars saved, or the net cost or net savings.

The necessity of associating monetary values with medical outcomes, such as the cost of a year of life lost or gained, makes a cost-benefit analysis a difficult measure to use for health care evaluation.

Economic analyses may incorporate the concept of perspective in assessment of costs. Costs can be considered from the perspective of the patient (eg, lost work time, travel costs, and copayments); the employer (eg, lost employee productivity, increased insurance premiums); the insurance company (eg, payments for physician visits, hospital stays, clinical procedures, and pharmaceuticals); or society.

The societal perspective takes into account the overall costs of a treatment to society as a whole. By focusing on all costs stemming from the treatment choice, the societal perspective has the advantage of being able to determine true cost savings that accrue to one member of society as a result of employing a specific treatment.

For example, employing a treatment that shortens hospital stays may significantly reduce the costs faced by an insurer, while shifting the financial burden to patients and their caregivers. Use of the societal perspective allows health care decision-makers to explicitly consider these tradeoffs.

Prospective Data Collection

Kevin A. Schulman, MD, MBA, and William L. Boyko, Jr., PharmD

In prospective analyses of the costs of care in clinical trials, the development of economic data as endpoints in the trial begins at the same time the clinical trial is being designed.[4] The initial step is to construct a study design to establish economic endpoints and data collection methods. This involves collaboration with the clinical study investigators and protocol approval from trial sponsors and individual Institutional Review Boards.

Economic case report forms need to be merged with the clinical case report forms. Special consideration must be given to the amount of data requested within the protocol for clinical and economic data, as the workload of the data collection staff is always a concern in these studies.

Final steps involve the generation of a database appropriate for the study, analysis of the data, and dissemination of the study results.

Estimating Costs

Data for the economic arm of the study should be collected during initial inpatient hospitalizations as well as during the follow-up period. In addition to chart review as a data collection means, regular telephone interviews can be conducted to obtain data associated with specific categories of postdischarge resource consumption.

For the hospitalization phase of an economic analysis, costs can be estimated from patient charges on hospital bills. A hospitalwide cost-to-charge ratio can be obtained from the Medicare cost report data, and this information can be used to estimate costs based on patient charges. Data regarding physician visits can be collected and assigned CPT-4 codes, which are then assigned costs using the Medicare fee schedule.[5,6]

For the posthospitalization phase of an economic analysis, costs can be estimated using a variety of methods. Rehospitalization costs can be estimated with the same methods used for initial hospitalization.

Costs for chemotherapy should consider both the cost of the drug as well as physician time. Drug cost can be estimated using pharmaceutical wholesale prices, while physician time can be valued using the Medicare physician fee schedule for chemotherapy administration.[5,7]

Costs of radiation therapy can be estimated based on standard regimens using the Medicare fee schedule and should include both initial cost components and costs associated with weekly radiation therapy team visits.

Costs of transfusions (packed red blood cells, white cells, platelet standard units, platelet apheresed units, fresh frozen plasma, and whole blood) can be estimated based on proprietary cost data from individual hospitals.

Provider costs can be separated into three categories: physicians, nurses, and home health. Physician visit costs can be estimated using the Medicare physician fee schedule. Costs associated with both nurses and home care visits can be estimated by multiplying the length of the visit by the hourly cost for each nurse or service, respectively.[8,9]

Outpatient surgeries and procedures can be estimated by assigning each procedure a CPT-4 code and assigning a relevant cost based upon the Medicare fee schedule.

Retrospective Data Collection

Joe Laver, MD, and Tammy Pajeau, MS

 Although there is an increasing need for economic data on new therapeutic agents, cooperative cancer group clinical trialists have been reluctant to incorporate economic analyses prospectively into their studies because of concerns over funding of the study, overburdening of the data management operations and statistical center, and considerations of conflict of interest.

The Cancer and Leukemia Group B (CALGB) canceled a clinical trial designed to evaluate the efficacy and cost-effectiveness of use of hematopoietic colony-stimulating factor vs oral antibiotic supportive care following non-small-cell lung cancer chemotherapy.[10] The cancellation followed a 2-year planning period. In the end, it was felt that, although economic analyses are important, there was not sufficient funding to support data management efforts for both clinical and economic evaluations.

The Eastern Cooperative Oncology Group (ECOG) and the Pediatric Oncology Group (POG) have devised an alternative strategy for evaluating economic outcomes alongside phase III clinical trials based on a retrospective data collection method. Recently, the retrospective method has been used to evaluate costs of care for cooperative clinical trials conducted by the Southwest Oncology Group (SWOG), POG, and ECOG.[11-13]

Clinical information is obtained directly from the statistical coordinating center of the clinical trial. These data sets include information on patient demographics, clinical status, clinical outcomes, toxicities, survival, and use of important resources such as days with antibiotics, number of days of colony-stimulating factors, use of platelet and red blood cell transfusions, and development of grade 3-4 infections.

One method of economic analysis involves imputing standardized costs for each of the resources utilized. An alternative method involves the use of patient billing informaton. Using this method, economic data are obtained for a subset of patients participating in the clinical trial. The data include detailed financial accounts for costs and charges for patients at some of the major institutions in the study; the data are obtained from hospital finance departments and reviewed by economic analysts.

Specific cost-to-charge ratios are included in these reports for each major resource within many departments, including blood bank, room, laboratories, radiology, pharmacy, and microbiology. Specific resource counts and charges are available for each of the major resources in each department, with costs of each resource imputed based on internal cost-to-charge ratios. In some institutions, professional fees are also included in the data set.

The clinical and economic data sets are then combined in an effort to evaluate clinical and economic outcomes. Because economic data are obtained for only a subset of patients on the trial, modeling efforts such as decision analyses are used. Thus, the final result is an estimated cost of care and resource utilization profile for patients on each of the treatment arms.

As more medical centers develop integrated systems of delivery, information on posthospitalization resource utilization costs of care can be made available.

Comparison of Retrospective and Prospective Data Collection Efforts

Charles L. Bennett, MD, PhD, and Peter D. Weinberg, BS

Of the two methods of collecting economic data in conjunction with clinical trials, the prospective method involves greater planning, as shown in Table 1, and additional personnel (Table 2).

In general, core competencies associated with clinical trial co-investigators, statistical support, economic analysts, and policy analysts are required for both methods. However, the prospective data collection effort requires additional input and personnel to pilot test the case report forms, obtain patient consent and release forms for access to information from all medical providers, train and monitor data collectors, and oversee the completeness and quality of the data collection efforts.

In contrast, the retrospective data collection effort is highly dependent on the data that are obtained from electronic data sources and does not require the additional personnel of the prospective method.

Each method has benefits and drawbacks (Table 3). The primary benefit of the prospective method is its completeness. When economic endpoints are part of the original study design, data requirements from a variety of sources can be anticipated, including inpatient, outpatient, and nonmedical sources of care. This allows for an estimate of total costs of care to be derived, and can include information on social costs and out-of-pocket expenditures.

In contrast, the main benefit of the retrospective data collection effort is related to the ease and relatively low costs of obtaining medical cost information. In a retrospective analysis, unnecessary data collection can be avoided by review of the clinical results.

Comparisons of the drawbacks of each method are also important. The prospective data collection effort requires a go/no-go decision at the time of initiation of the clinical trial, whereas this would not be needed in a retrospective analysis. For example, if the investigational agent is found to be clinically ineffective, the retrospective economic analysis would not be performed.

The retrospective data collection effort is limited by including data only from those sites with access to electronic data on medical care.

The use of the randomized clinical trial setting itself can be a drawback to economic studies. For example, patient selection bias can occur because of the generally strict inclusion criteria for these studies. Patients accrued to randomized trials are often generally healthier than those outside the trial setting and generally have better outcomes.

Clinical trials are typically conducted in academic institutions with access to highly trained staff and high-tech facilities that may not be available in the general clinical setting.

Also, clinical trials stipulate diagnostic follow-up at intervals more often than in a regular treatment setting, which may be more costly or, conversely, may save costs by leading to earlier diagnosis of adverse events. Each of these factors can affect the results of a cost analysis and must be considered when analyzing the data.

Conclusions

Economic evaluations are playing an increasingly important role in the assessment of clinical treatment strategies for cancer patients. Use of comprehensive data on the cost and effectiveness of cancer therapies obtained from economic studies of clinical trials can assist physicians and patients to make appropriate treatment decisions.

While prospective or retrospective data collection efforts have been successfully completed for clinical trials of cancer patients, each method has unique benefits and drawbacks. Future studies using both methods are needed.

References:

1. Schulman KA, Yabroff KR: Measuring the cost-effectiveness of cancer care. Oncology 9:523-533, 1995.

2. Gold MR, Siegel JE, Russell LB, et al: Cost-effectiveness in Health and Medicine. New York, Oxford University Press, 1996.

3. Eisenberg JM, Schulman KA, Glick HA, et al: Pharmacoeconomics: Economic evaluation of pharmaceuticals, in Strom BL (ed): Pharmacoepidemiology, 2nd ed, pp 469-494. Chincester, John Wiley & Sons, 1994.

4. Mauskopf J, Schulman K, Bell L, et al: A strategy for collecting pharmacoeconomic data during phase II/III clinical trials. Pharmacoeconomics 9:264-277, 1996.

5. American Medical Association: Physicians’ Current Procedural Terminology, CPT-4 ’94. Chicago, American Medical Association, 1993.

6. Health Care Financing Administration: Revisions to payment policies and adjustments to the relative value units under the physician fee schedule calendar year 1995; final rule. Federal Register. December 2, 1995.

7. Medical Economics Company, Inc.: The RED BOOK. Montvale, Medical Economics Data, 1995.

8. American Nurses Association: 1994 average salaries. Hospital and Health Care Report. Rochelle Park, Wyatt Data Services, April, 1994.

9. National Association for Home Care: Basic Statistics About Home Care. 1995.

10. Bennett CL, Smith TJ, George SL, et al: Free-riding and the prisoner’s dilemma: Problems in funding economic analyses of phase III cancer clinical trials. J Clin Oncol 13:2457-2463, 1995.

11. Bennett CL, Waters TM: Economic analyses in clinical trials for cooperative groups: Operational considerations. Cancer Invest 15:448-453, 1997.

12. Bennett CL, Golub R, Waters TM, et al: Economic analyses of phase III cooperative cancer group clinical trials: Are they feasible? Cancer Invest 15:227-236, 1997.

13. Pajeau MS, Lane D, Bennett CL, et al: Economic analysis of G-CSF use with intensive treatment for pediatric lymphoma and T-cell leukemia (abstract). Blood 90(suppl 1):316-I-73a, 1997.

Dr. Schulman is director, Clinical Economics Research Unit, Georgetown University Medical Center, Washington. Dr. Boyko is a fellow of the Clinical Economics Research Unit, Georgetown University Medical Center. Dr. Laver is director, Division of Hematology/Oncology, Department of Pediatrics, Medical University of South Carolina. Ms. Pajeau is a research analyst, Chicago VA Health Care System, Lakeside Division. Dr. Bennett is senior research associate, Chicago VA Health Care System, Lakeside Division, and associate professor of medicine, Lurie Comprehensive Cancer Center of Northwestern University Medical School, Chicago. Mr. Weinberg is a research analyst at Northwestern University.