Economic Analysis Costing

Were the costs and outcomes adequately measured?

Essentials:

1.   Were true costs measured – both up front and downstream (see below for explanation)? Were true costs measured or were surrogates, such as hospital charges, used?

2.   Outcomes: is there evidence of effectiveness of the alternatives being compared? Does the intervention work in a real world setting? Does it work in all patients to whom it was offered?

3.   It is essential to measure effectiveness not efficacy – the latter overestimates usefulness [i.e. does the treatment do more good than harm? This is fine for laboratory animals, but it is the ethical obligation of physicians not to harm patients].

4.   Were costs and outcomes appropriately evaluated in terms of timing: a dollar today is worth more than a dollar tomorrow? Discounting of net future costs and benefits should take place. The more commonly used discount rate in published CEAs is 3%.

Costing – the real bugbear.

In order to perform a CEA both inputs and outcomes must be measured in their entirety. Costs may be fixed, variable and intangible. Costs may be direct or indirect. Intangible costs represents pain, suffering, and emotional costs to the patient and his or her relatives. They are not quantifiable. Investigators often talk of upfront and downstream costs.

An example of downstream costs follows:

“….an economic analysis of thrombolytic therapy in the setting of acute myocardial infarction, from a societal perspective, should include the subsequent need for rehospitalizations and revascularization procedures. In addition to direct costs, costs to the patient (time off from work, child care, travel, etc.), costs borne by employers, other employees, society, and non-health aspects of the intervention on society should be considered from a societal perspective (Weinstein 1996)

These indirect or downstream costs are very difficult to quantify. They represent side effects of therapy, both negative – worsened health status, and positive – reduced future health costs. Some critics suggest that the future social welfare and health costs of the patient should be included (as these do arise if the patient survives), but it is not conventional to include these factors.

True economic costing thus involves direct, indirect and opportunity costs. Marginal costs are an essential element of CEA, and Chalfin argues that total costs are less important than incremental costs between different choices. 

True costing is difficult to determine as it includes medications, personnel, furniture, lighting, telephones etc. Often surrogate markers are used, such as, in the United States, hospital charges. Another method is to multiply the average bed price from the hospital ledger by the number of days stay. Both of these methods are significantly flawed (Gyldmark 1995, Heyland 1999, Rich 1999). These methods do not represent patient specific resource usage. Some patients become acutely ill, deteriorate and are admitted to intensive care where they respond rapidly to therapy. The cost of their treatment thus is very high initially, and then wanes. Other patients continue to worsen and, as their disease progresses, their treatment becomes more expensive – due, for example, to the increase in antibiotic costs, dialysis costs etc. Intensive care patients are both different from ordinary ward patients and different from each other. Using standardised methods of surrogate costs will thus be inaccurate.

Another method of abstracting cost data is through reimbursement collection data (such as that for Medicaid). Costs are based on DRG (diagnosis related groups) diagnoses. The problem with DRGs is that patients are not admitted to ICU due to their diagnosis, rather, their problem is severity of illness and loss of physiologic reserve. ICU patients are more expensive than average, which DRGs describe, and hospitals lose financial resources using this process. Intensive care patients are thus subsidised in the DRG system.

Gyldmark (1995) looked at twenty published cost studies relating to intensive care. None of these used a registration of costs process for each patient, and were all consequently flawed. She notes that most cost studies are “bottom up” retrospective studies where data has been extracted from computerised bills of the hospital accounts system with the problem of imperfect registration of resource usage and activity. The “bottom up” approach involves recording costs at a cost-object level and multiplies units of resource use with unit costs. This approach measures only indirect costs and direct costs must be apportioned. To perform this retrospectively requires very accurate patient files or resource use registrations.

Gyldmark points out that one of the major problems with costing ICU patients is in the area of unit costs: it is often very difficult to obtain the true cost, for example, of a lab test or a medical service. This problem has not been resolved. She recommends a standardised cost model based on decision tree analysis.

Please note: these tutorials are for personal study purposes only.  They are not currently peer reviewed, and no responsibility will be taken for mistakes or inaccuracies. Reproduction of information is forbidden. All material is copyrighted by the GasWorks Group.