Physician fees and procedure intensity: the case of cesarean delivery
Introduction
A major topic of public policy concern in the U.S. in recent years has been the rapid growth in the expenditures of the major public insurance programs, Medicaid and Medicare. Medicaid, which provides health insurance to low income women and children, disabled, and elderly, has grown by almost 400% over the past decade, with total state and federal expenditures of US$148 billion in 1996. Medicare, which provides universal coverage to the elderly, has grown by 250%, with expenditures of US$203 billion in 1996. These two programs now amount to 18% of the federal budget, and have accounted for 30% of the growth in the budget over the past decade.1 As a result, there is enormous pressure to rein in program costs. But doing so through channels such as restricted eligibility or reduced beneficiary benefits has run into considerable political difficulty. Instead, much of the attention to cost control in these programs has focused on provider reimbursement, for example through reducing physician fees.
In theory, cutting fees to physicians is a simple mechanism for controlling costs. In practice, behavioral responses by physicians render this a much more difficult proposition than simple static analysis would imply. On the one hand, there is concern that lowering fees might lead to deterioration in quality of care for the publicly insured. On the other hand, there is also concern about physician `induced demand', whereby providers may make up a large share of fee reductions through increased volume, limiting any cost savings from reductions in fees.
In fact, there are enormous health care literatures on the consequences of fee changes in these two programs. But these literatures have had quite different focuses. The literature on Medicare has focused on the effect of fee changes on the quality of care, and it has largely concluded that lowering fees will raise treatment intensity; as a result, estimates of the savings to the government from fee reductions include an offset for induced demand responses. The literature on Medicaid, on the other hand, has focused on patient access, assessing whether reductions in fees may lead providers to avoid potential Medicaid patients. But there has been remarkably little work on the effect of reduced Medicaid fees on the quality of care of Medicaid patients, conditional on gaining access to the system.
It is therefore tempting to apply the induced demand findings of the Medicare literature to estimate the effects of fee changes in the Medicaid program. In fact, however, there is reason to believe that the evidence from the Medicare program may not apply to the case of Medicaid. The reason emerges from the basic economics of physician behavior. In a standard model such as that of McGuire and Pauly (1991), fee changes will have both income and substitution effects on physicians. As McGuire and Pauly point out, the induced demand hypothesis is simply the implication of a model where income effects dominate substitution effects. This may be a relevant concern in the Medicare program, since treatment of Medicare patients is often performed by physicians who specialize in the Medicare population. But we present data showing that Medicaid patients are a smaller share of a physician's patient pool, so that income effects will be less important. In this context, it is therefore possible that substitution effects will dominate, so that lowering fees will lower treatment intensity.
We investigate this issue within the context of a particular example, cesarean section delivery of childbirth. Cesarean delivery is a particularly useful example because the underlying costs of the procedure in terms of physician time and intensity are considered to be similar to the alternative, vaginal birth, yet reimbursement has traditionally been higher.2 Moreover, there is tremendous variation across state Medicaid programs in the reimbursement of cesarean delivery. In this paper we consider the effect of variations in this differential on the likelihood of cesarean delivery for the Medicaid population.
In particular, we use discharge abstract data on Medicaid-paid births from nine states over the period 1988–1992 to model the effect of the cesarean reimbursement differential on substitution towards cesarean delivery in the Medicaid population. We match to these data information on Medicaid fee differentials, and we estimate the effect of these differentials on the mode of delivery for childbirth. We find that, on net, there is a strong positive effect of fee differentials on use of cesarean delivery, in contrast to the induced demand findings of the Medicare literature. We estimate that a US$100 increase in the absolute differential between cesarean and vaginal delivery reimbursement increases the cesarean delivery rate in the Medicaid population by 0.7 percentage points. We also find that the lower fee differentials between cesarean and normal childbirth under the Medicaid program than under private insurance can explain between one-half and three-quarters of the difference between Medicaid and private cesarean delivery rates. This finding suggests that cutting reimbursement for cesarean delivery under Medicaid lowers the intensity of treatment of childbirth.
Our paper proceeds as follows. Section 2sets up a simple model of physician behavior which provides the framework for integrating the previous views. Section 3reviews the previous literature on physician behavior. Section 4presents our data and estimation strategy, and our results are presented in Section 5. Section 6concludes.
Section snippets
Theory
We begin our analysis with a brief theoretical formulation of the physician intensity decision. Our analysis closely follows the models of McGuire and Pauly (1991) and Gruber and Owings (1996). Our goal here is to highlight the ambiguous prediction for the effect of fee changes on treatment intensity, and to focus on the relative role of income and substitution effects.
The physician obtains utility from income, but disutility from inducement. His separable utility function has the form:
Physician fees and treatment intensity
There is a sizeable health economics literature on the effects of fee changes on procedure intensity.3 Much of this work has
Data and empirical strategy
Our primary source of data for this analysis is the Healthcare Access and Utilization Project (HCUP) data, collected by the Agency for Health Care Policy and Research (AHCPR). This survey collects information from a random sample of discharges across eleven states, over the 1988–1992 period.
Basic results
Our basic findings are shown in Table 2, which presents logit estimation of Eq. (vi). These coeffecients can be interpreted by multiplying them by p.(1−p), which is 0.146+our sample mean. We use the absolute fee in columns (1) and (3), expressed in hundreds of dollars, and the normalized (by private fees) fee in columns (2) and (4). In contrast to the Medicare literature, we find a highly significant positive coefficient on the fee differential, either specified as absolute dollars, or
Conclusions
Provider fee policy remains the tool of choice for policy-makers in trying to rein in program costs. As a result, the effect of fees on treatment intensity has received enormous attention in the context of the Medicare literature, but there has been scant attention paid to this question in the context of Medicaid. This is unfortunate, because as we have argued the results for Medicare cannot be naively applied to the case of Medicaid. If the induced demand type results that have been found for
Acknowledgements
Gruber acknowledges research support from the National Sciences Foundation and the National Institute on Aging; Kim acknowledges research support form the National Institute on Aging. We are grateful to Joe Newhouse and two anonymous referees for helpful comments.
References (32)
- et al.
Physician-induced demand for surgery
Journal of Health Economics
(1986) - et al.
Physician response to fee changes with multiple payers
Journal of Health Economics
(1991) Induced demand—can we ever know its extent?
Journal of Health Economics
(1986)- Adams,...
- American Hospital Association. Hospital Statistics, various years. AHA, Chicago,...
- Baker and Royalty,...
Volume responses to exogenous changes in medicare's payment policies
Health Services Research
(1992)- Currie, Gruber, Fischer,...
- Decker,...
- Dranove, Wehner,...