substitution possibilities among medical inputs for treating a particular condition. More specifically, the BLS practice of separately pricing and weighting distinct medical item strata neglected the medical care sector’s substitution across various strata (e.g., physician services, prescription pharmaceuticals, laboratory tests), thereby overstating price increases.
A well-known example of this problem involves inpatient and outpatient hospitalizations. One way in which managed care has reduced overall hospitalization costs is to shift many surgeries from inpatient to outpatient environments. A consequence of this substitution is that both the severity of illnesses and the complexities of surgery for the average patient have increased over time for both inpatient and outpatient procedures, resulting in an increase in per diem costs in both settings. Because BLS priced these inpatient and outpatient procedures separately and then used constant weights over time to aggregate them, their weighted sum increased over time, even though providers’ and insurers’ total hospitalization costs (inpatient plus outpatient) for these surgical cases declined. Since the 1998 revisions, the BLS has treated inpatient and outpatient hospitalizations as a combined bundle (stratum), although the frequency with which the substrata weights will be updated has not been made clear to the panel.
Another example of substitution across strata involves the treatment of a mental disorder such as depression (see Frank et al., 1998; Berndt et al., 2001). The clinical literature has demonstrated that a number of alternative treatments for depression involving various combinations of psychotherapy and antidepressant drugs have, on average, equal expected outcomes. In the MCPI, the services of psychotherapists are classified in a separate stratum from antidepressant pharmaceuticals. Over the last decade psychotherapy-intensive treatments for depression have been reduced, and they have been replaced by either a combination of psychotherapy and pharmacological treatments or by pharmacological treatments alone. Although the movement away from psychotherapy-intensive procedures has, in many cases, reduced total costs per treatment episode, this cost reduction is not captured by current BLS procedures because the inputs come from distinct strata, each of which is priced separately and reweighted infrequently. Moreover, this problem would not be resolved by treating inpatient and outpatient hospitalizations as a combined stratum.
BLS has been moving toward the pricing of an episode of an illness, rather than pricing medical inputs. For example, for those hospitals and physicians billing for services provided to Medicare beneficiaries, BLS is now collecting selected price quotes based on Medicare’s Diagnostic Related Group (DRG) classification scheme, a system that to some extent encompasses episodes of treatment.
A major set of issues now facing BLS is how to broaden the new approach to encompass the entire treatment of medical conditions, not just hospital stays. One possibility is the following. First, the BLS could choose a subset of diagnoses or illnesses (perhaps between 15 and 40, depending in part on the costs involved in