Lying for Patients:  Physician Deception of Third-Party Payers FREE

IN THE past decade, managed care programs have increased their role in the daily practice of medicine.¹ Physicians have experienced pressure from managed care organizations (MCOs) to reduce utilization and curtail costs. Practice controls such as utilization restraints and preapproval criteria have limited physician autonomy.^2- 5 As MCOs impose stricter controls on access to specific medical treatments, physicians are increasingly facing an ethical dilemma—how to reconcile their professional responsibilities as patient advocates with potentially conflicting contractual obligations to third-party payers. Anecdotal evidence^6- 9 suggests that some physicians have chosen to maintain the appearance of compliance while skirting managed care guidelines in order to secure treatment. Such deception, primarily in the form of incorrect or ambiguous documentation, has been called "gaming the system."¹⁰ However, the full extent to which physicians support the deception of third-party payers, particularly for which clinical indications, remains unknown. Furthermore, physicians' attitudes and beliefs about the use of deception are not well characterized. We report the results of a study investigating physicians' attitudes toward the deception of third-party payers, and identify clinical and demographic patterns associated with the sanctioning of deception.

We developed a survey instrument to assess physicians' attitudes toward the deception of third-party payers as a means of securing patient treatment. Our intent was to present physicians with clinical vignettes that a primary care physician might face in practice, and to evaluate their willingness to sanction the use of deception and their attitudes about deception for each vignette. The survey was distributed to internists in 4 regions of the United States. We initially developed 10 vignettes of varying clinical severity and pilot tested the survey instrument among physicians in an internal medicine group at an academic health center. Based on follow-up interviews and analysis of responses, 6 vignettes that covered a range of clinical severity were selected and further refined for the final survey instrument.

Physicians were presented with clinical vignettes in which a colleague was considering deceiving a third-party payer to secure treatment for a patient. The use of questions framed for another agent committing a potentially unethical act is an accepted method for eliciting respondent perceptions.¹¹ Physicians were informed that for each vignette, (1) the patient is seeking only the specific procedure or referral discussed; (2) the patient cannot afford the procedure or referral on his or her own because of financial hardship, and no external resources are available; (3) local colleagues are unwilling to provide services unless they are reimbursed; (4) despite diligent efforts by the physician and colleagues, the patient's third-party payer is unwilling to change its reimbursement policy regarding the particular procedure or referral; and (5) the third-party payer checks documentation but does not confirm clinical findings.

To evaluate physician support for the use of deception, we employed a conceptual framework based on casuistry.¹² Casuistic analyses begin with a paradigm case and then modify aspects of the case to determine whether changes in the details of the case modify judgments about the morality of the action. In our analysis, we varied the clinical severity and risks associated with the patient's condition between vignettes. Vignettes ranged from life-threatening (arterial revascularization and coronary bypass graft) to emergent (psychiatric referral), palliative (intravenous pain medication and nutrition), diagnostic (mammography), and cosmetic (rhinoplasty). In each vignette, physicians were asked whether their colleague should deceive the third-party payer. Physicians were then asked how they, their colleagues, and society would rate the physician's deceptive behavior on a 5-point Likert scale ranging from 1 (clearly justified) to 5 (clearly unjustified) (see "Study Vignettes" section).

Respondents also provided demographic data, practice characteristics, and career and professional perceptions. Demographic data included physician age, sex, and race or ethnic affiliation. Practice characteristics included years since completion of residency, years practicing in the current metropolitan market, percentage of time devoted to direct clinical care, payer mix (percentage of patients enrolled in Medicare, enrolled in Medicaid, capitated, self-pay, or uninsured), type of clinical income (salaried or nonsalaried), and type of practice setting (academic health center, staff health maintenance organization [HMO], solo practice, or group practice with number of associates). Career and professional perception items evaluated agreement with statements concerning satisfaction with a medical career, opportunity to practice quality medical care, professional autonomy, business aspects of practice, and relationships with colleagues.¹³ Physicians were also asked to select 1 of 5 statements characterizing their primary professional responsibility.

To determine what role practicing in a predominantly managed care market might have on physician attitudes, we identified 4 metropolitan markets in the United States with a high level of managed care penetration. We defined a high–managed care penetration market as a market in which 40% or more of the total metropolitan area population was enrolled in an HMO. Results of prior research¹⁴ have shown that HMO penetrations greater than 30% result in a significant HMO effect on provider and purchaser communities—affecting product diversity, premiums, provider reimbursement, and most important, health system organization. We also identified a low–managed care penetration market in each of the same regions, in which less than 25% of the population was enrolled in an HMO. Metropolitan area managed care penetration figures and populations were obtained from the 1995 Interstudy Competitive Edge Regional Market Analysis 6.1 and reflect market data as of July 1, 1995.¹⁴

From the list of potential candidate metropolitan areas, we identified high– and low–managed care penetration markets in each region that were most similar in population. We then selected the largest such pairing for inclusion in the study. The high-penetration markets selected (and their respective managed care penetrations and study regions) were Boston, Mass (42.1%, Northeast); Miami, Fla (43.6%, South); Minneapolis–St Paul, Minn (42.0%, Midwest); and San Francisco, Calif (42.1%, West). The low-penetration markets selected (and their respective managed care penetrations and study regions) were Pittsburgh, Pa (24.2%, Northeast); Atlanta, Ga (16.1%, South); St Louis, Mo (22.7%, Midwest); and Las Vegas, Nev (20.9%, West). Sample physicians were divided among these 8 metropolitan markets, based on the size of the local population.

We further modified our sample to assess sex effects by quota sampling within each metropolitan region to provide maximal power for sex comparisons. The vignettes were presented to half of the sample in order of increasing clinical severity, and in reverse order to the other half to test for order effects. We limited our sample to physicians exclusively board certified in internal medicine. Physicians were randomly drawn from metropolitan market listings in the 1996 American Board of Medical Specialties Directory of Board Certified Medical Specialists.¹⁵ Physicians who had board certifications in fields other than internal medicine, had completed subspecialty fellowship training (with the exception of general medicine fellowships), or were no longer active in clinical practice were excluded from analysis and replaced with a randomly selected physician of the same sex from the same metropolitan market.

Sample physicians were initially contacted by mailed questionnaire in December 1996. Physicians with incorrect addresses were followed up by local directory assistance where possible. After the start of the study, nonresponders were followed up by 2 additional survey mailings at 1 and 3 months, 2 follow-up telephone calls at 2 and 4 months, and a facsimile transmission of the survey instrument at 6 weeks. All surveys were identified by a unique identification code to ensure confidentiality of responses. Study protocol, methods, and survey instrument were approved by Georgetown University Medical Center's institutional review board, Washington, DC.

The primary outcome of interest was physicians' sanctioning of a colleague's deception of the third-party payer in each of the 6 vignettes. The effect of clinical severity on the sanctioning of deception was evaluated by the Cochran Q test and McNemar tests of deception rates in each vignette. We also tested for an association between support for deception and other variables—demographics, practice characteristics, and perceptions of career and professional responsibilities. Physician's reported personal, collegial, and societal justifications of deception in each vignette (reported on a 1- to 5-point Likert scale) were evaluated by a 3 (personal, collegial, and societal perspectives) × 6 (vignette) repeated-measures analysis of variance. A significant interaction effect was followed up using tests of simple effects by the Tukey Honestly Significant Difference test for paired differences. Perceived primary professional responsibility and career satisfaction items are also reported. Categorical variables were compared with the χ² test and expressed in proportions. Continuous variables were compared using an independent sample t test analysis. Statistical analysis was performed with 2 software packages (SAS 6.11; SAS Institute Inc, Cary, NC, and Statistica; StatSoft Inc, Tulsa, Okla).

A 55-year-old woman presents to a colleague as a new patient seeking referral for coronary revascularization (coronary artery bypass graft [CABG]). She is homebound because of angina. A recent angiogram indicated severe 3-vessel coronary occlusion that appeared amenable to CABG. However, she was recently forced to switch insurance companies. She remains stable but is receiving a maximal medical regimen. She reports occasional severe angina symptoms relieved by sublingual nitroglycerin administration, but there has been no progression in her chest pain symptoms. Based on a medical literature review, your colleague understands that CABG is medically indicated for such a debilitating and life-threatening condition, yet knows that the patient's third-party payer will not pay for CABG for this preexisting condition unless she has progressive chest pain. Thus, the patient can forgo CABG or have it at her own expense. However, your colleague also knows that the third-party payer will pay for coronary revascularization if he or she documents "increased frequency of chest pain" as an indication for coronary revascularization, even though such documentation would not be accurate.

A 55-year-old woman presents to a colleague as a new patient seeking referral for arterial bypass grafting for chronic arterial insufficiency to her foot. She has never smoked but has chronic arterial insufficiency secondary to chronic atherosclerotic disease. She is in pain "at rest," and has chronic nonhealing ulcers. Physical examination reveals pregangrenous skin changes that the patient reports existed before her current insurance coverage began. Based on a medical literature review, your colleague understands that arterial revascularization is medically indicated to avoid later amputation of the toes or foot. Your colleague knows that the patient's third-party payer will not pay for arterial revascularization if the pregangrenous changes were a preexisting condition, which leaves her to forgo the arterial revascularization or have the procedure at her own expense. However, your colleague also knows that the third-party payer will pay for arterial revascularization if he or she documents "new skin changes: pregangrenous" as an indication for arterial revascularization, even though such documentation is inaccurate.

A 55-year-old woman presents to a colleague as a new patient seeking referral for intravenous pain medication and nutrition as part of her comfort care at home. She has been diagnosed with incurable ovarian cancer, and is not expected to live long. She has chronic abdominal pain that is relieved only by taking oral pain medication. She also has severe, refractory nausea for 1 to 2 hours after swallowing solids or liquids. All antinausea drugs have failed to provide any relief. She reports no vomiting and no problems swallowing. Based on a medical literature review, your colleague understands that intravenous pain medication and nutrition are medically indicated as part of comfort care for a patient with such a terminal condition, yet knows that the patient's third-party payer will not pay for comfort care as long as she is still able to swallow solids and liquids without vomiting, leaving her to forgo intravenous treatment (and continue taking oral medication and nutrition), or receive it at her own expense. However, your colleague also knows that the third-party payer will pay for intravenous pain medication and nutrition as part of comfort care if he or she documents "recurrent vomiting" as the indication for intravenous pain medication and nutrition, even though such documentation would not be accurate.

A 55-year-old woman presents to a colleague as a new patient seeking referral for psychiatric treatment for depression. She notes that the physician who was caring for her died unexpectedly, and that she has been unable to get further information or referrals through his office. She claims a recurrent history of profound depression and cites several past suicide attempts, revealing well-healed scars down both wrists. She presents her prescription medication bottles and claims compliance with a maximal antidepressant therapy regimen for more than 3 months. She reports a 3-week history of daily feelings of profound sadness, guilt, and worthlessness. She has had no interest in regular activities, great difficulty concentrating, and feels completely estranged from her family. In addition, she reports fatigue, insomnia, little appetite, and a 15% loss in body weight during the past 3 weeks. She is tearful, agitated, and concerned that she is feeling worse with each passing day. She strongly denies suicidal or homicidal ideation. She is cooperative and even willing to be hospitalized—but only if her insurance will cover the costs. Based on a medical literature review, your colleague understands that urgent psychiatric referral is medically indicated, yet knows that the patient's third-party payer will not pay for psychiatric referral unless she states an intent or exhibits demonstrable evidence of imminent suicidal or homicidal behavior. However, your colleague also knows that the third-party payer will pay for a psychiatric referral if he or she documents "patient states suicidal ideation" as an indication for the psychiatric referral, although such documentation would not be accurate.

A 55-year-old woman presents to a colleague as a new patient seeking referral for a mammogram. Her younger sister was recently diagnosed with breast cancer. The patient has never had a mammogram but is concerned that she too may have breast cancer. She has not noted any new breast lumps, skin changes, or nipple discharge. Your colleague's physical examination reveals no evidence of any breast lumps or lesions. Based on a medical literature review, your colleague understands that an annual screening mammogram is medically indicated for women older than 50 years, especially if they have a family history of breast cancer. Your colleague knows that the patient's third-party payer will not pay for a screening mammogram, leaving the patient to forgo screening or have it at her own expense. However, your colleague also knows that the third-party payer will pay for a mammogram if he or she documents "suspicious breast lump" as an indication for the mammogram, even though such documentation would not be accurate.

A 55-year-old woman presents to a colleague as a new patient seeking referral for rhinoplasty (the surgical alteration of nasal bones). She has no significant medical history and currently has no medical complaints. However, she is sad about feeling less attractive with each passing year. In addition, she becomes tearful when recounting how she has been teased throughout her life about having a prominent nose. She is realistic in not expecting changes in her nose to radically change her life, but she would like to improve this one feature of her body. She denies any injury to her nose and reports no problems breathing. Based on a medical literature review, your colleague understands that rhinoplasty is not medically indicated but would certainly make the patient feel better about herself. Your colleague knows that the patient's third-party payer will not pay for rhinoplasty as an elective procedure, leaving the patient to forgo the procedure or have it at her own expense. However, your colleague also knows that the third-party payer will pay for rhinoplasty if he or she documents "deviated septum, problems breathing" as an indication for the surgery, even though such documentation would not be accurate.

Of 602 randomly selected physicians, we obtained accurate contact information for 471. A further 18 physicians were excluded because of retirement from practice or having moved out of the sample area, and they could not be replaced by random sampling. At the conclusion of data collection in June 1997, we received 169 responses (37.3% of contacted physicians). Respondents and nonrespondents did not differ by sex, age, or years in practice.

Respondents were predominantly women (52.7%) and white (81.7%), with a mean age of 41.6 years. Most respondents were salaried (78.1%) and located in high–managed care penetration markets (56.8%). A total of 40.8% practiced at academic medical centers. On average, respondents spent 69.4% of their time in clinical work, had been in practice 10.1 years after residency, and had spent 9.3 years in their current metropolitan market (Table 1).

Our principal analysis was a determination of physician support for deception of a third-party payer to secure patient treatment (Table 2). Most physicians sanctioned deception to obtain authorization for coronary bypass surgery (57.7%) and arterial revascularization (56.2%). Many physicians also approved deception to obtain intravenous pain medication and nutrition (47.5%), mammography (34.8%), and psychiatric referral (32.1%). Few physicians supported deception for cosmetic rhinoplasty (2.5%). More than a quarter of physicians (26.6%) did not support deception in any vignette, whereas 13.6% sanctioned deception in all vignettes except cosmetic rhinoplasty.

Sanctioning of deception was associated with clinical severity of vignettes (Cochran Q = 200.25; P<.001). McNemar tests indicated that sanctioning of deception decreased as clinical severity decreased (Table 2). In other words, the deception rates for each vignette are arranged in descending order as follows: (coronary artery bypass ≈ arterial revascularization) > intravenous pain medication and nutrition > (psychiatric referral ≈ mammography) > cosmetic rhinoplasty.

Analysis of deception by practice location identified significant variation. Physicians practicing in high– managed care penetration markets were more likely to sanction deception in arterial revascularization (64.1% vs 45.7%; P = .02) and mammography (44.1% vs 22.5%; P = .004) vignettes. These physicians also provided higher rates of deception approval for all other vignettes, including cosmetic rhinoplasty, although differences were not statistically significant (psychiatric referral, P = .10; intravenous pain medication and nutrition, P = .18; CABG, P = .28; and cosmetic rhinoplasty, P = .008).

Physicians who sanctioned deception spent less time in clinical practice than those who did not. In the arterial revascularization vignette, physicians sanctioning deception reported spending 64.8% of their time in clinical practice compared with 75.4% who did not (P = .04). Percentage of clinical time was also lower for physicians who sanctioned deception in intravenous pain medication and nutrition (61.5% vs 78.0%; P = .001), psychiatric referral (61.1% vs 74.2%; P = .02), and mammography (57.8% vs 76.5%; P<.001) vignettes. Clinical experience, as evaluated by years in practice after residency, was not associated with deception in 5 of 6 vignettes; physicians suggesting deception for psychiatric referral had been in practice significantly longer (13.4 vs 8.5 years; P = .02). Practice setting was not associated with sanctioning of deception in 5 of 6 vignettes; physicians practicing at academic health centers trended toward greater sanctioning of deception in the mammography vignette (42.6% vs 28.6%; P = .06).

Sanctioning of deception was not associated with physician sex in any vignette. Physician race and age were also not associated with sanctioning of deception, although both trended toward significance in the mammography (P = .05) and psychiatric referral (P<.06) vignettes. Analysis of demographic differences in rates of deception, however, were limited because of low statistical power. In addition, the order of vignettes was not associated with differences in deception attitudes. Because of the low proportion of physicians sanctioning deception in the cosmetic rhinoplasty vignette, we could not evaluate demographic or practice effects for this case. Comparisons of deception sanctioning by demographic and practice characteristics are provided in Table 3 and Table 4.

Physicians rated the use of deception as most justified for a patient requiring coronary bypass surgery and arterial revascularization, and least justified for a patient requesting cosmetic rhinoplasty. A repeated-measures analysis of variance of the justifiability of deception ratings indicated a significant interaction between clinical vignette and perspective (F_10,1510 = 5.64; P<.001). Analysis of simple effects showed several patterns. First, there were significant differences in physicians' reported justifiability ratings from the 3 perspectives for 5 clinical vignettes. In only the coronary artery bypass vignette did respondents believe that they, colleagues, and society would agree on the justifiability of deception. Second, physicians consistently assumed that society would provide greater justification ratings than they would. Third, physicians assumed that they and colleagues would provide comparable justifiability ratings for the 4 more severe vignettes. Physicians were somewhat less tolerant of a deceptive act than they believed their colleagues might be for the mammography and cosmetic rhinoplasty vignettes. Finally, differences between physicians' reported colleague and society justifications were found only in the intravenous pain medication and nutrition vignette. Justification ratings for deception were comparable in all other vignettes.

Ratings of the justifiability of deception paralleled variations in the sanctioning of deception associated with managed care markets. Physicians practicing in high-penetration markets reported significantly greater acceptance of deception than those practicing in low-penetration markets in the arterial revascularization vignette (2.2 vs 2.7; P<.05). Greater physician justification of deception in high-penetration markets was also reported in the coronary bypass graft (2.2 vs 2.6), psychiatric referral (2.8 vs 3.2), and mammography (2.8 vs 3.4) vignettes (P<.05 for all).

The high rates of sanctioning of deception were consonant with physicians' reported professional obligations. Most physicians (76.4%) believed their primary professional responsibility was to practice as their "patient's advocate; working within the rules and restrictions of third-party payers, so long as those rules do not significantly compromise my patient's interests." Physicians reporting patient advocacy as a primary responsibility were more likely to sanction deception for arterial revascularization (62.1% vs 25.0%; P = .001), coronary bypass (63.2% vs 29.2%; P = .02), intravenous pain medication and nutrition (53.7% vs 23.1%; P = .005), and mammography (37.1% vs 24.8%; P = .03). Physicians sanctioning deception in the psychiatric referral vignette reported significantly greater dissatisfaction (P = .002) with their degree of professional autonomy; dissatisfaction with professional autonomy was not associated with sanctioning of deception in other vignettes. Sanctioning of deceptive behavior was unrelated to other measures of career satisfaction. Professional responsibility and career perceptions are reported in Table 5.

In this study, many physicians sanctioned the use of deception to obtain medical care for a patient when denied authorization by a third-party payer. Sanctioning of deception varied by the clinical severity of a patient's condition, but deception was sanctioned by most physicians in our 2 most clinically severe vignettes. Physicians also reported high levels of justification for the deception of third-party payers. Because most physicians report that it is their duty to work as patient advocates within the rules of third-party payers until those rules compromise their patients' care, this underlying attitude may represent physicians' main justification for sanctioning deception and the view that deception is justifiable.

Although physician support for deception was high in the overall cohort, we noted significant variations between locations. Physicians practicing in high–managed care penetration markets were more likely to sanction deception than were those in other markets. Although only significant in 2 scenarios, physicians in high-penetration markets trended toward higher rates of deception in all clinical vignettes. A similar pattern was observed for physician ratings of the justifiability of deception, which were also higher in high-penetration HMO markets. Taken together, these data suggest that HMO market penetration affects physicians' willingness to sanction the use of deception. This finding is consistent with the hypothesis that, as managed care brings more restrictions to a market, physicians practicing in that market become more willing to support deception as a means of circumventing these restrictions.

We hypothesized that deception rates would be highest in vignettes of greatest clinical severity and immediate patient risk. The relative ranking of physician sanctioning of deception in the vignettes supports this hypothesis. Furthermore, physicians showed greater support for deception in life-threatening illness than in urgent care needs, and in cases of palliative and diagnostic care than in cosmetic care.

Previous studies have documented physician willingness to use deception to secure medically indicated patient care. A survey¹⁶ of physicians practicing in the Northeast found that nearly 70% of respondents would knowingly misdocument a screening test as a diagnostic service to ensure coverage of service. Primary care physicians have historically misdocumented diagnoses to reduce the risk of patient stigmatization or insurance and employment discrimination.¹⁷ However, results of recent studies¹⁸ show that patient concerns about maintaining insurance coverage and ensuring continued access to care through service reimbursement have become more frequent reasons for deliberate miscoding. Similar uses of inaccurate documentation to secure patient care have been reported by other health care professionals.¹⁹ In fact, an evaluation²⁰ of ethical dilemmas in outpatient practice identified patient cost considerations as the primary source of ethical problems arising from office visits.

The advent of managed care and increased reliance on utilization review has further restricted a physician's ability to practice independently, increasing patient and physician frustration.²¹ These effects are logically felt most in markets where managed care has the greatest effect on health care delivery. Our data suggest that physicians in such markets are more willing to respond to managed care's restrictions by condoning deception. Physician deception in these markets might not even be limited to life-threatening vignettes but also used in other cases in which a physician believes a procedure is in the patient's best medical interests. This sentiment can reach extremes. That physicians who condoned deception to secure coverage for cosmetic rhinoplasty—the survey's only medically nonindicated vignette—were exclusively from high–managed care penetration markets underscores the extent of this sentiment. Resistance to such practices may also explain physicians' unwillingness to cooperate with third-party payers in other areas, including disclosure of patient information²² and suspected patient health insurance fraud.²³

Both MCOs and professional societies are aware of this trend in physician behavior.^10,23- 27 Managed care firms have responded to inaccurate claims by increasing the rigor of utilization review and terminating contracts with physicians suspected of using deception. Professional societies have standing guidelines that prohibit fraudulent misrepresentation in dealings with insurers. However, these guidelines are not fully developed. The American College of Physicians' Ethics Manual²⁵ states that physicians are not obligated to lie to a third-party payer for a patient, but it does not expressly condemn deception to obtain care. The American Medical Association's Council on Ethical and Judicial Affairs²⁶ and the Woodstock Theological Center²⁷ specifically enjoin physicians from using deception or gaming the system and instead, urge physicians to work through appeals processes at MCOs. Although President Clinton's recently announced Patient Bill of Rights and similar Congressional efforts^28- 29 might facilitate such appeals, patients' expectations of their physicians, coupled with MCOs' fundamental reliance on decreased use of services to reduce costs, suggests continued conflict and only minimal gains.

Although many physicians believed deception was appropriate in the face of restrictions on patient care, a large proportion also refused to sanction deception. More than a quarter did not recommend deception in any study vignette. Written comments of some respondents indicated that they would try to secure coverage and overturn previous service denials by discussing the vignette patient's situation with the third-party payer, despite the study vignettes' instructions otherwise. Although these physicians may be successful in their appeals, their acceptance of restrictions on medically indicated treatment has significant implications for patient care. Furthermore, current case law holds that physicians who are willing to accept an MCO's limits on provision of care have no immunity from civil liability related to the withholding of care.³⁰ Respondents who feel strongly compelled to follow contractual rules might be surprised to find that compliance with MCO rules does not decrease their exposure to malpractice or legal obligation to patients.³¹ In contrast, the smaller group of physicians who sanction deception in all medically indicated vignettes (cosmetic rhinoplasty excluded) support the equally troubling conclusion that some physicians view deception as an acceptable method of dealing with third-party payer limitations. Acceptance and deception, rather than advocacy, set dangerous precedents for future practice for both the physician involved and the profession.

Despite the nature of our findings, deception is not new to medical practice. Physicians have been known to advocate deception in several situations: lying when they believe it will benefit a patient,³² choosing not to inform terminally ill patients of their true diagnosis,³³ and intentionally miscoding to protect patient confidentiality¹⁷ or to secure reimbursement for treatment.⁶ In more spurious cases, physicians have also used deception for personal benefit, as a means of disguising malpractice, for career advancement, or to defraud payers.^34- 36 Hospitals have also been implicated for deceptive practices in dealing with third-party payers, from subtle "diagnosis related group creep"³⁷ to more extreme cases of Medicare fraud.³⁸ However, until recently, deception was unrelated to access to care. Physicians were historically able to treat uninsured patients and those with limited access to care by distributing treatment costs across the remainder of their paying practice.³⁹ As cost shifting and other methods of providing unreimbursed care have disappeared under managed care, so have the avenues by which physicians can treat patients who lack coverage for specific services.

Our study has some limitations. Although physicians indicated a significant willingness to sanction deception, we are unsure of the extent to which such willingness translates into the actual use of deception in practice. Are physicians as willing to deceive in practice when faced with the possibility of civil or criminal sanctions? Conversely, does the frequent process of attempting to secure patient care and repeated frustration further embolden a physician? What are the implications of recent legislative and judicial attempts to address this issue? Although we cannot be certain, we believe our data indicate that physicians, at a minimum, accept and are willing to use deception as a means of responding to third-party payer restrictions. In addition, our relatively low response rate limited our availability to evaluate effects such as practice in a staff model HMO, patient payer mix, and physician demographics.

In conclusion, the practice of medicine now extends beyond the patient-physician relationship. Physicians and patients are contractually bound to third-party payers and their rules. On occasion, such obligations conflict with physicians' obligations to act as patient advocates. Although using deception to solve these impasses may succeed in the interim, their long-term costs in loss of integrity and increased practice oversight are high. Situations that produce deception can ultimately only be solved by direct confrontation and frank dialogue between physicians, patients, and payers. Alternatives to deception include broadening existing appeals processes on behalf of individual patients and political advocacy for health care reform. Refusal to initiate a social dialogue regarding the appropriate balance between medical and economic considerations places medicine at risk of becoming a practice of equal parts patient care and subterfuge.

Accepted for publication March 25, 1999.

Presented in part at the Society for General Internal Medicine Mid-Atlantic Regional Meeting, Philadelphia, Pa, March 21, 1997; Society for General Internal Medicine Annual Meeting, Washington, DC, May 2, 1997; and Association of American Health Plans Annual Meeting, New Orleans, La, June 14, 1997.

Reprints: Kevin A. Schulman, MD, Center for Clinical and Genetic Economics, Duke Clinical Research Institute, Duke University Medical Center, PO Box 17969, Durham, NC 27715 (e-mail: schul012@mc.duke.edu).