The Physicians Practice

Dr. Katherine Sauer Metropolitan State College of Denver Health Economics

Chapter Outline: I. A Model of the Physicians Practice II. Supplier-Induced Demand III. Small Area Variations IV. Other Issues

I. A Model of the Physicians Practice McGuire and Pauly (1991) describe physicians as utility maximizers which means that physicians value items besides profit. In this model, the physician gets utility from net income and leisure, and disutility from inducement, (the physicians own efforts to induce patients to buy more care than appears medically necessary.)

U = f ( , L, I) is net income L is leisure I is inducement

A. The income leisure tradeoff:

Assume working returns constant revenue. - working for one more hour means your income increases by w and you get one less hour of leisure.

24 Hours of Leisure Hours of Labor

Income

As the wage increases, the income line gets steeper. income w1 income w2 income w3 The physicians optimal balance between labor and income changes.

24 Hours of Leisure Hours of Labor

wage

w3 w2

Initially, the physicians income effect dominates the substitution effect. - higher wage leads him to work more There comes a point where the substitution effect dominates. - higher wage leads him to work less
L1 L3 L2 Hours of Labor

w1

An increase or decrease in income causes the physician to reevaluate the choice of how much to work

B. The income inducement tradeoff: - disutility from inducement

U1

Inducement

Income mQo +mI

A certain level of income will be achieved even with no inducement. m is the rate of profit per patient care unit

mQo

Inducement

Q0 is the number of patient care units without inducement.

Income

U mQo +mI

The optimal level of inducement is found at the tangency between the income line and indifference curve. This physician induces QI* nonmedically necessary care.

mQo

QI*

Inducement

Income

U1 m1Qo +m1I U2 m2Qo +m2I

If the profit rate per patient falls, the income line is lower and flatter. Now the optimal level of inducement is higher.

QI1* QI2* Inducement

C. Do Physicians Respond to Financial Incentives? Nassiri and Rochaix, (2006) when physicians are paid per service provided, they provide more services than when they are given a fixed total payment (capitation) Studies also suggest that physicians respond to income pressures on their practice by striving to increase their incomes (Iversen, 2004; Gruber and Owings, 1996; Quast, Sappington, and Shenkman, 2008; Rizzo and Zeckhauser, 2003, 2007).

II. Supplier-Induced Demand (SID) On becoming ill, consumers hire health care professionals. In medicine, we identify the physician as the agent, and the patient as the principal. The policy concern is that out of self-interest physicians may violate their roles as agents.

Physicians provide care, but also sell it. - competition lowers the per-patient profit rate - increased inducement An increase in physicians would cause an increase in the supply of services, but also an increase in demand for services (induced demand).

Health economists have modeled supplier-induced demand for at least two reasons: - to understand the motivations of physicians, how their incentives affect their practice - models are needed to understand the data we observe

A. Supply and Demand Model Even without SID, the market predicts that an increase in supply of physicians will increase the quantity of services. - supply shifts right higher quantity, lower price If demand increases such that the price of service rises, then SID is identified (Reinhardt Fee test). But, maybe physicians respond to competition by increasing their quality. (dont need SID to explain higher quantity)

B. Target Income Hypothesis (TIH) This argues that physicians have desired incomes that they strive to achieve or to restore whenever actual income falls below the targets. This target income model is a special case of the benchmark model, though a relatively extreme one. This model has received criticism because - income is the main focus of the physician - but not income beyond the target

If the profit per patient falls, so does the physicians income.

The TIH says this physician will try to increase their income.

m2Qo +m2I

I1

I2

Inducement

The physician can increase income by increasing inducement. The dashed line is parallel U1 Income to the original m1Qo +m1I income line, and then U2 1 tangent to the new level of 2' utility. m2Qo +m2I
2

I1

I2

I2

Inducement

This shows the change in inducement from removing income from the physician. The income effect is the U1 Income distance I1 to m1Qo +m1I I2.
U2 1 2' 2

m2Qo +m2I

I1

I2

I2

Inducement

If instead of having a target income a physician is a profit maximizer, then a reduction in the profit rate results in an income effect equal to zero. - profit-maximizer gains utility only from more net income - values inducement only if it brings in more income

C. McGuire-Pauly synthesis: the size of the income effect is critical to understanding SID behavior A lower profit rate (m) will have two effects on inducement: - substitution effect: when inducement is less profitable, providers will substitute away from it - income effect: decreased income will make inducement more attractive

D. The Parallel Between Inducement and Marketing Stano (1987) argued that an influx of new competition (which reduces the physicians profit rate) may lead physicians to induce more, or less, depending on the cost structure of the firms production and its advertising. If the physicians SID is analogous to advertising/ inducement, then inducement would usually decline in an increasingly competitive market.

E. What Do the Data Say? Two criticisms were raised about much of the earlier SID work. - many of those studies could not distinguish between the SID model and the conventional supply and demand model. - many estimates of the SID effect proved to be statistically flawed

McGuire (2000) showed that the implications of availability on fees in when physicians operate in monopolistically competitive markets are not so clear. Feldman and Sloan (1988) show that if physicians can adjust their quality in response to increased competition, then higher fees could result even when there is no inducement.

Rizzo and Blumenthal (1996) use surveys of physicians to compare their desired incomes to their actual current incomes. Physicians with greater gaps were found to demand greater price increases. Rizzo and Zeckhauser (2003) find the physicians whose current incomes fall below their reference incomes are observed to show greater income growth than the average of other physicians in subsequent periods.

III. Small Area Variations (SAV) Are physicians themselves always well informed? Medical and surgical use rates can vary even in a small geographical area.

A. How to Measure The Coefficient of Variation (CV) relates the standard deviation of observed medical use rates to the mean of the same measure. - then divide by the mean (adjusts for size of the rate being studied) 0.00 to 0.10 is low variation 0.10 to 0.20 is moderate variation 0.20 and up is high variation

Extremal Ratio ratio of the largest rate observed across the small areas to the smallest rate observed

Variations by Medical Procedure Type

B. Causes of the Variation Much of the SAV work focuses on - contribution of socioeconomic characteristics of the population - role of the availability of supplies of hospital and physician services The studies together reached two conclusions: 1) Supply variables are important and demand characteristics play a somewhat lesser role though both are statistically and materially significant. 2) much variation is still left unexplained

C. The Physician Practice Style Hypothesis Practice style probably varies among physicians due to an incomplete diffusion of information on medical technologies. Wennberg (1984) argued that much of the observed variation is closely related to the degree of physician uncertainty with respect to diagnosis and treatment.

Epstein and Nicholson (2005) find that a physicians residency has relatively little influence on his practice style. Stronger influences are his peers with the hospital where he practices as well as his peers in the other hospitals in his region.

Studies show that information programs directed at physicians can alter their behaviors and thus presumably their practice styles. One study found that an informational program significantly affected the tonsillectomy rates in 13 New England areas.

Wennberg and Fowler (1977) found that morbidity and many socioeconomic variables were not sufficient in explaining the variations in a region. Concluded that variations in use rates probably are due largely to practice style differences across the small areas.

Phelps and Parente (1990) found that standard demand and supply variables typically account for between 40 and 75 percent of the variation in their study of 134 separate diagnostic categories. Escarce (1993) found that 43 percent of the variation in cataract surgery rates for the Medicare population is explained by socioeconomic variables.

D. Social Cost of Inappropriate Utilization The most important issue in the SAV literature is the proposition that substantial variation in utilization rates is an indication of inappropriate care. Phelps and Parente find that the welfare loss due to variations from true practice in the nation total $33 billion.

Inefficiency from Misinformation about Benefits

MC

MB as perceived by Doctor 1

True MB

MB as perceived by Doctor 2

R1

R*

R2

Rate of Utilization of Intervention X

IV. Other Issues A. Physician Pricing and Price Discrimination Physicians in private practice have some degree of monopoly power (monopolistic competitors) Arbitrage of physician services isnt possible. Price discrimination is therefore possible. By segmenting their market and charging different prices to different patients for the same services, physicians can increase revenue.

B. Paying for Outcomes: fee-for-service vs fee-for-outcome Fee-for-outcome is rare in the developed world but fairly common among traditional healers (ex: in Africa). Fee-for-outcome is difficult because of the uncertainty in treating heath issues. Outcome based contracts are more successful when both the patient and the physician work together.

Key Points Our benchmark model depicts the physician as someone who values positively net income, and leisure, and dislikes inducing patient demand. The model shows that for the supplier induced demand (SID) hypothesis to be supported, the physicians income effect must be positive and substantial. Small area variations can be understood as a result of uneven diffusion of medical information to these same physicians.

Discussion Questions: What forces limit a providers ability and willingness to engage in SID? In the profit-maximizing model of SIC, what are the costs to the physician of inducement? Assuming that SID is prevalent and substantial, what are the implications for policy?