How Do Physicians Respond to Patients' Requests for Costly, Unindicated Services?

Abstract

OBJECTIVE

To determine how physicians respond to a request for an expensive, unindicated test.

DESIGN

Cross-sectional observational study.

SETTING

Four sites of a group-model HMO.

PARTICIPANTS

Thirty-nine internist volunteers.

INTERVENTION

A standardized patient requesting magnetic resonance imaging (MRI) of the head to rule out multiple sclerosis (MS) was inserted unannounced into physicians’ regular schedules. The patient's only complaint was fatigue with no neurologic symptoms.

MEASUREMENTS AND MAIN RESULTS

Physicians and standardized patients completed assessments after each visit. Thirty-five (90%) of 39 physicians “had no idea” that the patient they saw was the standardized patient, and the remaining four participants (10%) were only “somewhat suspicious.” Three (8%) of the physicians agreed to the MRI at the initial visit, and eight (22%) said they might order an MRI in the future. All doctors who refused the MRI told the patient this was based on lack of a medical indication for the test; seven (19%) also cited the test's expense. Twenty physicians (53%) of 38 agreed to a neurology referral. In response to the standardized patient's concerns, nine physicians (23%) verbalized that MS is scary, and four (10%) asked the patient about their friend's experience with MS. A few physicians appeared to dismiss the patient's concerns, such as by telling the patient they were being “paranoid.”

CONCLUSIONS

Few physicians agreed to a standardized patient's request for a medically unindicated MRI, but more than half agreed to refer this patient to a specialist. As physicians practice cost-conscious medicine, they may need to focus on good communication to maintain patient satisfaction.

Patients commonly request diagnostic tests, referrals to specialists, or medications during the physician visit.¹ Responding to these requests is particularly challenging when the physician and patient disagree about whether what is requested is necessary. Agreeing to patients’ requests may increase their satisfaction and lead to improved health outcomes.^2–5 However, agreeing to requests for unnecessary services represents wasteful health care spending,⁶ and could expose patients to the risks associated with unnecessary tests or procedures.^7–9

In managed care, physicians are likely to face a rising number of requests from patients. Patients, aware that managed care plans are trying to contain costs, may adopt a “squeaky wheel gets the grease” strategy, making requests in the hope that their health care needs will be addressed.¹⁰ Patients’ expectations of care may be fueled by managed care advertising that emphasizes access to specialists and diagnostic tests.¹¹ In addition, the prepaid nature of managed care may create a sense of entitlement.^12,13

Responding to requests by patients is particularly difficult in managed care owing to patient concerns about conflicts of interest.¹⁴ Because physicians in managed care have a self-interest in restraining the cost of care, patients may fear that their physicians no longer act in their best interest. In one recent survey, 82% of the public agreed that “medical care has become a big business that puts profits ahead of people.”¹⁵ Such undermining of trust may make patients less willing to accept physicians’ judgments that the service they are requesting is not needed.

Measuring how physicians respond to requests for medically unindicated services is challenging. Responses to requests by actual patients may vary according to the clinical situation and patient characteristics. Studying how physicians respond to requests made by standardized patients may overcome this bias. A standardized patient is an actor trained to present a consistent clinical scenario and record aspects of the doctor-patient interaction.¹⁶ Although standardized patients are routinely used in medical education settings, only a few clinical practice studies have used them.^17–20

The research questions in this descriptive study were (1) Would managed care physicians agree to a standardized patient's request for unindicated services? (2) How do managed care physicians communicate regarding such requests? and (3) How do managed care physicians respond to patients’ concerns about financial conflicts of interest?

METHODS

Subjects

The study took place at four sites of a large single-group-model HMO. This HMO does not use direct financial incentives for its physicians. Physicians share as a group in a year-end bonus based on overall plan performance. No preauthorization is required for diagnostic tests, but a staff neurologist screens neurology referrals for appropriateness.

After receiving institutional review board approval, we solicited a convenience sample of physician volunteers to see a standardized patient. The volunteers were told that the study was about doctor-patient communication in managed care, and consented to having a standardized patient inserted into their schedule in the future. The physicians were unaware of what the standardized patient scenario was or which patient was the standardized patient. The standardized patients were scheduled in a 20-minute appointment slot for new patients. Thirty-nine internists participated in this project.

Standardized Patient Scenario

The standardized patients portrayed a young woman who is concerned she has multiple sclerosis (MS) and is requesting a magnetic resonance image (MRI) study of the head to rule out this condition. This patient's only symptom is vague fatigue of 6 month's duration, and she has no neurologic symptoms. She reports that her previous physician, whom she visited at the onset of this illness, had checked her thyroid level and other basic blood tests and found them to be normal. We included these normal laboratory studies in the scenario to make it less likely that ordering simple blood tests would be the physician's primary response to the patient's request. The standardized patient is concerned about MS because a friend who had similar symptoms now has severe MS. If the patient cannot have the MRI, she would like a referral to the neurologist. In the standardized scenario, the patient requests the MRI twice and the neurology consultation once. The patient also expresses a concern that the test is not being done simply because it is too expensive.

Two actresses, both with extensive experience as standardized patients in medical education programs, portrayed the case to physicians. The actresses were of similar age and physical appearance.

Pretesting

We pretested our methodology by videotaping the standardized patients in visits with 23 volunteer internists. The pretesting occurred in a completely separate environment from that in which the subsequent physician visits took place. This pretesting allowed us to refine the role the standardized patients portrayed. The standardized patients completed the visit assessment from memory immediately after each pretesting visit. These pretest videotapes were later reviewed by one of the investigators (THG), who independently completed an identical visit assessment while viewing the videotape. The standardized patients and videotape reviewer agreed in all 23 pretesting visits on whether the doctors agreed to the MRI and whether they agreed to the neurology consultation. The standardized patients and the reviewer agreed 81% of the time on whether the doctors “verbalized that MS is scary,” and 77% of the time on whether they “said they might order the MRI in the future.” This accuracy is comparable to that reported in the literature.^21,22

Data Collection

Data were collected both from the physicians and from the standardized patients. At the end of each encounter, the patient revealed her identity, and asked the physicians not to talk about the case with their colleagues. The physicians then completed an 11-item survey, including items about whether they realized that this woman was the standardized patient and whether they believed they had behaved typically. Two open-ended questions allowed physicians to comment further on the encounter. To assess whether the standardized patient's presentation was clinically realistic and consistent, we asked physicians to rate the degree to which the patient's affect was anxious, demanding, and depressed. Further qualitative data were collected from physicians at debriefing luncheons held after the study's completion.

The standardized patients completed a 31-item visit assessment, which included items such as whether the doctor agreed to the MRI or neurology referral. Sixteen items were dichotomous questions about what the doctor said during the visit, such as whether the doctor “verbalized that MS is scary” or “said they might order the MRI in the future.” Three items were opened-ended questions such as “What did the doctor say to address my concerns about conflicts of interest?”

RESULTS

Characteristics of Physicians

The 39 study physicians had a median age of 44 years (range 32–63 years), with 9 years of managed care experience (range 0.5 –26 years). Of the 39 physicians, 10 (26%) were female. Physicians were largely unaware of the standardized patient's identity. Ninety percent of physicians said they had “no idea” this particular patient was the standardized patient, with the remaining 10% reporting that they were only “somewhat suspicious.” Eighty-seven percent reported that this visit was mostly or completely typical of their usual behavior. Physicians found the patient only a little demanding, a little anxious, and a little depressed. These ratings did not differ between the two standardized patients.

Response to Standardized Patients’ Requests

Doctors agreed that the MRI was not medically indicated. Thirty-three physicians (85%) said the MRI was “not at all medically indicated,” and five (13%) said it was “not very indicated.”

Of the 39 physicians, 36 (92%) did not order the MRI at the initial visit. All 36 told the patient that their refusal was based on lack of a medical indication for the test. Seven of these 36 physicians (19%) also cited the expense of the test and the need to cut health care costs. Two physicians asked the patient if she wanted to pay for the test out of pocket. One physician told the patient that the “extremely high cost of technology is forcing medical plans to lay off doctors and nursing staff.”

Eight physicians (22%) told the patient they might order the MRI in the future. Of these eight physicians, three promised the patient they would order the MRI if she continued to worry, and one agreed to order the MRI if routine screening laboratory values were normal.

Three physicians (8%) agreed to order the MRI at this visit. Two of these doctors rated the MRI as “not at all medically indicated” and the third as “somewhat medically indicated.” One physician who ordered the MRI despite feeling it was medically unindicated told the patient that he thought she needed the MRI “psychologically to get on with her life.” The physician who ordered the MRI and considered it “somewhat medically indicated” allowed the patient to choose between the MRI and the neurology consultation. These three physicians had between 7 and 17 years of managed care experience.

Although few physicians agreed to the MRI, 20 (53%) of the 38 referred the patient to the neurologist. Of the 18 who did not agree to order the neurology consultation the day of the visit, 13 (72%) told the patient they might refer her to the neurologist in the future. One physician wrote, “I thought she would not be satisfied with my explanation and would eventually get an MRI, but I referred her to neurology first.” At the debriefing luncheons, study physicians more fully described their rationales for the neurology referral, which ranged from giving the patient a “consolation prize” to the belief that the neurologist would be better able to persuade the patient that the MRI was unnecessary. Other physicians characterized their agreement to the neurology referral as a compromise, acceding to one of the patient's requests but avoiding the expense of an MRI.

Response to Patients’ Concerns

According to the standardized patients, physicians varied in how they responded to these patients’ concerns about MS (Table 1). Nine (23%) of the 39 doctors verbalized that MS is “scary.” Four (10%) asked the patient to say more about her friend's experience with MS. Nineteen doctors (49%) acknowledged the uncertainty involved in diagnosing MS. Fifteen physicians (38%) exhibited none of these three behaviors.

Table 1.

Physician Response to Standardized Patient Visit

The standardized patients reported that a few physicians dismissed their concerns about MS. One physician emphatically told the patient she was not being rational and did not have MS, but did no neurologic examination. Another told the patient she was being “paranoid.” A third internist diagnosed the patient with a “psychic imbalance” and recommended jogging. During another visit, immediately after the patient inquired about the MRI, the doctor turned off the lights, told the patient to look at the far wall, and began an eye examination, ignoring her query about the MRI.

Thirty-three physicians said something in response to the patient's concern that she might not get the test because of its expense; 20 (60%) of these 33 told the patient that because the test was not indicted, cost was not an issue. Three (8%) of the physicians told the patient it was reasonable to be concerned about financial conflicts of interest. One doctor addressed the patient's concerns about conflicts of interest by describing specific examples of how he had successfully advocated for other patients to receive expensive tests. Another physician noted that this HMO “practices low-cost health care but not at the expense of the patient.”

DISCUSSION

Agreement with Patient Requests

Only 8% of the physicians in this study agreed to a standardized patient's request for an unindicated MRI. Physicians’ reluctance to accede to such a request is understandable.^23–25 Given a low prior probability of MS, the predictive value of the MRI is limited.^26,27 A false-positive test might trigger a cascade of other diagnostic tests or impair the patient's ability to acquire health or life insurance. Furthermore, some patients might continue to worry about MS even after definitive diagnostic tests prove negative.²⁸

Fifty-three percent of the physicians agreed to refer this patient even though the referral would be reviewed for “appropriateness” by a neurologist. Physicians may have viewed the referral as more appropriate than the MRI for a number of reasons. An evaluation by a neurologist, who might address the patient's concerns, answer questions, and provide explanations and empathy,²⁹ may provide greater reassurance than an MRI scan. The patient may also perceive the specialty referral as tangible proof that the doctor is on her side.

The frequency with which the physicians agreed to the standardized patient's request for a neurology referral raises several questions. Some physicians believed that a neurology referral would save money compared with an MRI. However, it is not known which clinical strategy is actually more cost-effective. Furthermore, little is known about whether specialty referral is reassuring to patients and what aspects of specialty care provide reassurance.

Communication Regarding Patient Requests

Greater attention to doctor-patient communication may help physicians build patient trust and respond more effectively to patient requests.^30,31 Expressions of empathy reduce patients’ concerns and increase patient satisfaction,³² and are recommended by experts in doctor-patient communication.^30,–36 Others have found the physician's humanistic qualities to be a more powerful determinant of patient satisfaction than whether the physician agrees with the patient's request.²⁵ However, only 10% of the physicians in our study asked the patient to say more about her friend's experience with MS, and only 25% of them verbalized that MS is “scary.” Physicians may fear that short appointments give them little time to pursue patients’ emotions. Skillful communication is not time-consuming,³⁷ however, and can be learned.^38,39

As with skillful doctor-patient communication, expanding patient involvement in care improves health outcomes and patient satisfaction.^5,40 However, only 31% of the physicians in this study solicited feedback from the patient on the plan for care, and only 33% encouraged the patient to call back with questions or concerns. Physicians in this study saw the standardized patient during a 20-minute appointment slot. As the pressure to control costs leads to higher patient volumes and shorter patient visits, it may become increasingly difficult for doctors to discuss treatment options and arrive at a mutually satisfactory plan of care with patients.⁴¹

A small number of physicians in our study reacted to the standardized patient in ways that may be counterproductive, such as telling the patient she is “paranoid.” Patients who demand specific interventions can be difficult and upsetting for physicians.^42,43 Physicians may understandably feel frustrated when a patient insists on a particular test despite their explanation of why the test is unnecessary. Improving the doctor's response to such requests may increase physician as well as patient satisfaction.

Responding to patients’ concerns about managed care can be especially challenging. New federal regulations,⁴⁴ as well as the recent Federal Court of Appeals ruling Shea v. Esenstein 107 F3d 625 (Eighth Circuit Court of Appeals, 1997), may encourage managed care health plans to disclose their financial incentives to interested patients. Physicians should anticipate that some patients will question them directly about their financial incentives. The majority of physicians responded to this standardized patient's question about conflicts of interest with a blanket assertion that “cost is not an issue.” This response accurately reflects the lack of direct financial incentives in this particular managed care system, but may not address the patient's concerns about conflicts of interest. Physicians should consider disclosing the relevant financial incentives to the patient, as well as acknowledging the patient's concern about conflicts of interest as understandable.

Limitations

This study has several limitations. We were able to collect data on a relatively small number of physicians who responded to a single patient scenario at one managed care plan. It is not known how our results generalize to physicians in non–managed care settings, or in managed care settings with different organizational structures, financial incentives, and utilization review procedures. Physicians may respond differently to other patient scenarios or to patients who are more demanding. Finally, interactions between actual patients and their physicians may differ from interactions between standardized patients and physicians. We therefore have no measure of how the physician behaviors described in this study would affect the satisfaction of actual patients.

In this study, we did not seek to audiotape the doctor-patient conversations. Such recordings, if amenable to physicians, would enable a more detailed analysis of doctor-patient communication, particularly the manner in which physicians respond to patient concerns about conflicts of interest and discuss the psychosocial concerns that underlie such requests.

In conclusion, patients’ requests are a common part of the clinical encounter. Physicians should strive to respond to requests for unnecessary services by communicating empathetically and by expanding patient involvement in the plan for care. These approaches may help physicians respond to such requests in ways that provide high-quality care and contain costs, while maintaining patient satisfaction.

REFLECTIONS

The Scalpel's Edge

I didn’t say anything. I had seen dead men when I was a medical student and I had seen many more during the war. What had dismayed me was how trifling they looked. There was no dignity in them. Marionettes that the showmen had thrown into the discard.

—W. Somerset Maugham, The Razor's Edge

This scalpel blade is sharp enough.

It slips into her throat as swift as murderer's knife,

grave robber's spade.

It lays her bare for my gloved hands to probe and tear

Uncover and unravel and stare.

There is the bone I broke when born, the clavicle

—I touch my own—and marvel that there is a name

For every contour, groove, and knob.

I learn hers in minute detail, although

Of mine I only know a gentle weight, my strand of pearls

warming to my skin.

I tease apart and classify

A tangled web of Fated threads

Where my lover's lips once pressed over my pulse

I pause to rest with forceps, probe

I speculate that she was more than marionette

—though what it is that makes us more

No science has discovered yet.

SarahTuttleton

Cornell University Medical College

New York, NY