Physician Incentives and Treatment Choice

• Date: 01 October 2015
• Author: Kevin E. Pflum
• Published date: 01 October 2015
• DOI: http://doi.org/10.1111/jems.12117

Physician Incentives and Treatment Choice

KEVIN E. PFLUM

University of Alabama

Box 870224, 361 Stadium Drive, Tuscaloosa, AL 35487

kpflum@cba.ua.edu

I analyze the impact of physician competition for patients on treatment selection and an insurer’s

ability to induce its preferences through a supply-side payment mechanism. Informed patients

choose the physician whose treatment practice best fits their preferences,aligning physician incen-

tives with patient preferences. An insurer’s ability to counter these incentives is not monotonic

in how informed patients are, however. When demand is either perfectly inelastic to treatment

practices because patients are completely uninformed or a sufficient proportion of the market

is informed relative to the number of physicians in the market, then an insurer can induce its

preferences through supply-side payment rules. Otherwise, more intensive policy levers such as

utilization review must be employed. Programs that increase patient information can therefore

improve the efficiency despite generating stronger incentive to treat according to patient prefer-

ences. I also explore how noisy signals of illness type and diagnostic testing further complicate

the insurer’s problem.

1. INTRODUCTION

As consumers increasingly turn to online reviews for a variety of products, it is no

surprise that there are now dozens of web sites providing reviews of physicians as

well.1In addition to these third-party sites, some insurers are providing their own

review systems to help enrollees select physicians based on the personal experiences

of other enrollees. For example, WellPoint has partnered with Zagat—an established

provider of restaurant and hotel reviews—to provide patient reviews of physicians for

their 35 million enrollees.2Physician reviews have to some extent always been part

of the business as patients have long shared their physician experiences with friends

and family members; however, the wealth and availability of physician information

provided by these new review systems is likely to have a much larger impact on how

physicians compete for patients.

Of course, physician competition for patients has not been developed only as a

consequence of patient reviews as there has always been some patients who seek second

opinions or switch providers when they are sufficiently unhappy with their physician’s

treatment recommendations causing physicians to factor in their patients’ preferences

when proposing a treatment (Givens, 1957; McCarthy, 1985; Macpherson et al., 2001;

Berry, 2007). Nonetheless, there is concern in the medical community that providing

I am grateful to Paul Pecorino, Paan Jindapon, Paula Cordero Salas, Paul J. Healy, Suhui Li, Thomas Buch-

mueller, and participants of the Ohio State University theory workshop, the 2012 Annual Conference of the

Southern Economic Association, and the 2013 International Industrial Organization Conference for helpful

discussions and comments. I also thank two anonymous reviewers for providing insightful comments and

suggestions that greatly improved the manuscript.

1. Examples include consumerreports.org , ratemds.com , vitals.com and drscore.com.

2. “Zagat Gets Into Doctor Ratings,” (http://blogs.wsj.com/health/2007/10/22/zagat-gets-into-doctor-

ratings-business/, accessed July 23rd,2013).

C2015 Wiley Periodicals, Inc.

Journal of Economics & Management Strategy, Volume24, Number 4, Winter 2015, 712–751

Physician Incentives and Treatment Choice 713

more information about physician treatment practices may exacerbate moral hazard

by increasing the pressure to provide the treatments patients want. Exemplifying this

concern, 55% of the respondents to a recent South Carolina Medical Association sur-

vey reported ordering tests they felt were inappropriate because of pressure to avoid

receiving a poor review from patients (Falkenberg, 2013). Moreover, these concerns are

particularly salient now given that, pursuant to the Affordable Care Act of 2010, the

Centers for Medicare and Medicaid Services (CMS) is preparing to make even more

information available to prospective patients by publishing its own physician perfor-

mance data that include patient ratings and information on how well physicians’ medical

interventions succeed.

In light of these efforts to provide more information regardingphysician treatment

practices to patients, the objective of this study is twofold: Explore how competitive

pressure among physicians due to informed patients may impact physician treatment

selection and identify how and when supply-side payment rules can overcome these

incentives to induce physicians to follow an insurer’s preferred treatmentpractice. These

objectives are accomplished using a stylized model of treatment and physician selection

that captures a couple critical features of the market. First, physician treatment selection

is modeled as a choice among discrete treatments in which the cost and benefit of each

treatment are dependent on the patients’ illness severity or “type” so that it is more

efficient to use one treatment over the other for each type although the more efficient

treatment varies with the type.3For example, an intensive medical intervention, such

as surgery, may be a more appropriate treatment choice for some men diagnosed with

prostate cancer, while a less intensive treatment option, such as active surveillance or

radiation, is more appropriate for others.

Second, patients select physicians based on their treatment practices. That is, pa-

tients have preferences over treatmentsthat depend on their type and out-of-pocket costs

while insurance creates a wedge between the private and social benefit of each treatment

by shielding patients from the full cost. Because of the discrete nature of the treatment

and the differences in their benefits, insurance does not drive all patients to demand

the most costly treatment—even if there is no difference in the patient’s out-of-pocket

costs—reflecting the argument made by Mendel et al. (2012) that patients are not likely

to demand, or alternatively accept, the recommendation for a more costly procedure

such as surgery simply because they are insured. Instead, patients prefer the treatment

that leaves them with the highest expected benefit and select a physician whose treat-

ment practice accomplishes this when informed about physician treatment practices. In

consequence, physicians can gain market share by following treatment practices that

more closely match patient preferences vis-`

a-vis other physicians. However, the amount

of market share that a physician can capture is dependent on the proportion of patients

who are informed. The insurer’s challenge is to design its payment rules in order to over-

come this competitive pressure and induce physicians to follow its preferred treatment

practice.

The principal finding of this study is that an insurer’s ability to induce its prefer-

ences through supply-side payments critically depends on how informed patients are

relative to the number of physicians competing for patients. Specifically, it is a simple

3. The discrete nature of treatment choice is in contrast to much of the literature examining optimal

insurance and physician agency that frequently consider scenarios in which a physician chooses a quantity or

“intensity” of medical care to provide a patient for a particular diagnosis or illness. Examples include Rochaix

(1989); Selden (1990); Ellis and McGuire (1986, 1990); Ma and McGuire (1997); Gal-Or (1999); and Chon´

eand

Ma (2011).

714 Journal of Economics & Management Strategy

matter for an insurer to induce its preferred treatment practice by making a physician

internalize the social gain from the treatments when patients are completely uninformed

and there is no competition via treatment choice. Accomplishing this when some pa-

tients are informed, however, requires that the physician earns relatively more when

she treats marginal patients with their less preferred treatment in order to counteract

the physicians’ incentive to attract patients by following a practice that patients prefer.

When a sufficient proportion of patients are informed (i.e., demand is sufficiently elastic

to treatment practices), the insurer will be able to induce its preferred treatment practice

in equilibrium. However, when demand is insufficientlyelastic (but not perfectly inelas-

tic) relative to the number of physicians in the market, an insurer may not be able to

simultaneously prevent physicians from increasing profit by gaining market share from

utilizing the patients’ preferred treatment more and prevent physicians fromincreasing

profit by utilizing the more profitable but less preferred treatment.

The intuition behind these finding is as follows. A physician who chooses a treat-

ment practice that is more preferred by patients than the other physicians’ practices

stands to gain a proportionally large market share as she takes a small proportion of

informed patients from a large number of competing physicians. To counteract this

incentive, the insurer must make sure that the physician gives up a sufficiently large

amount of profit on each patient she treats with the less preferred (by the insurer)

treatment to overcome her increase in profit obtained by increasing market share. Ac-

complishing this, however,generates an incentive to instead treat more patients with the

now more profitable (and less preferred by patients) treatment as the physician stands

to lose only a small number of informed patients. In consequence, when too few patients

are informed (but not all are uninformed), the insurer cannot prevent both deviations

as they are mutually exclusive. As more patients become informed, however, the loss

in patients from selecting a less desired treatment practice increases to the point where

both deviations become unprofitable. As a result, if information on physician treatment

practices is more readily available, then an insurer will be able to better exercise control

over physicians. The availability of diagnostic testing suffers from a similar problem if

patients are also responsive to physician testing practices. However, diagnostic testing

also complicates the insurer’s problem as a physician’s private gain from testing, which

is increasing in the demand elasticity (proportion of informed patients), may be too large

to overcome through the payment mechanism. In consequence, even when an insurer

can induce its treatment practices, policies that increase patient information could also

result in overtesting.

Although there is a large literature that has examined how insurers can exert

control over a physician’s treatment choice,4this paper shares features with Dranove

(1988), Chernew et al. (2000), and Liu and Ma (2013) in particular. Dranove (1988)

explores how physician-induced demand (PID) can arise in equilibrium even when

patients know the physician’s recommendation strategy. Dranove also explores how

physician competition will impact the degree of PID similarly showing that if patients

are responsive to physicians’ recommendation strategies, then thephysicians’ incentives

will be more aligned with the patients’ reducing PID. However, the focus of the study is

on how PID can arise in equilibrium and not on how an insurer can use payments to alter

4. See Gaynor (1994) and McGuire (2000) for detailed review of the literature on physician agency and

PID. Gaynor (1994) emphasizes research that examines physician–insurer agency, whereas McGuire (2000)

places an emphasis on research examining PID. L´

eger (2008) provides a review of the more recent literature

also focusing on PID and the issue of overtreatment.