Commentary (Hillner): Evaluating the Total Costs of Cancer

Financial costs to individuals and society represent one measure of the burden of disease. Accurate and complete estimates of the financial burden associated with cancer are a necessary component for reaching a variety of goals. These goals might include the costs and benefits of cancer control interventions, assessing the performance of health systems, and the allocation of health-care and research resources across disease categories.[1]

In this issue of ONCOLOGY, Drs. Calhoun and Bennett report on a series of pilot studies from their cancer center that, they assert, provide the framework for a large-scale future study of cancer care costs and services utilization. While I applaud the authors for their efforts in addressing this issue, the work presented herein is too superficial to provide more than a broad conceptual sketch of the framework.

Defining Costs

As the authors note, the financial "costs" associated with any illness can be broken down into various categories. Cost-of-illness estimates typically include three main elements: direct costs, morbidity costs, and mortality costs. Direct health-care costs are made up of costs typically paid by a third-party insurer through the health "system"-the expenses of tests, drugs, physician services, other personnel, supplies, and hospitalization. Most cost-of-illness studies use national level estimates or Medicare claims to calculate these direct costs.

National morbidity costs associated with cancer are estimated from disease-specific work disability. One study estimated that in 1990 about 10% of total cancer costs (~$10 billion) were morbidity-related.[1] In this same study, the largest cancer care costs were those incurred in lost incomes associated with premature death (~61%, $59 billion)

Although direct cancer costs to an insurer and morbidity costs to family and society are large, these do not directly weigh down a living cancer patient and his or her extended family. Calhoun and Bennett identify several of these additional costs: (1) the out-of-pocket direct cost for cancer care not paid for by insurers or government (copayments and medications); (2) the direct non-health-care cost, including the out-of-pocket costs to partake in the intervention (eg, transportation, child or family care, parking) or hiring services to assist or perform household responsibilities (eg, because the patient is too fatigued to prepare meals or cleaning); and (3) informal caregiver time (ie, the monetary value of the time of family members or volunteers who provide home care).

Measuring Costs

While it is easy to categorize the different forms of out-of-pocket costs, there are numerous methodologic uncertainties in their actual measurement. The current report describes a cohort of breast, prostate, and colorectal cancer patients who completed a comprehensive survey instrument and a subset who kept a diary of outof- pocket costs. The report provides no clues to the representative nature of this cohort: How many patients were approached and refused to participate? How can completeness be validated? Are diaries convenient and complete? What about an audit of receipts? What are the patient privacy issues?

A fourth category of costs that has rarely been systematically assessed at an individual level is the employment cost to the patient and the extended family in terms of time lost to treatment-ie, productivity costs associated with lost wages to the patient and possibly the family, while caring for and transporting the patient. For example, a woman may take a leave of absence or work less than a full schedule during the course of therapy. If adverse effects develop (ie, congestive heart failure or peripheral neuropathy), long-term employment costs could be substantial.

Intriguing Pilot Projects

The pilot project of 31 hairy cell leukemia survivors is tantalizing but of unknown representativeness. It is not surprising that the few survivors who attempted to get life insurance were refused. What is more concerning is the handful whose health insurance was discontinued.

The most important pilot project that the authors discuss is their assessment of the differences in types of costs incurred with chemotherapy toxicity. Direct costs to the "system" predominate in cases of neutropenia and thrombocytopenia. In contrast, costs incurred with lost employment by the patient and informal caregivers are more common in cases of neurotoxicity. While these findings do indeed show an impact on patients, there is no evidence that knowledge of these effects would necessarily influence choice of care. If an alternative treatment were available, these data would be vital to a cost-effectiveness assessment.

Conclusions

A full economic assessment of the cancer patient's specific out-of-pocket and employment costs is laudable. Yet, prior to launching such an initiative, extensive additional work is needed to address its feasibility, representativeness, acceptability, and data collection design.

Financial Disclosure: The author has no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.