Responses to the statement that personal financial incentives for physicians to encourage restraint in resource use are ethically acceptable.
Responses to the question, "How have changes during the past 10 years affected the ethic of undivided loyalty to patients?"
Responses to the question, "How have changes over the past 5 years affected your patients' trust in you?"
Sulmasy DP, Bloche MG, Mitchell JM, Hadley J. Physicians' Ethical Beliefs About Cost-Control Arrangements. Arch Intern Med. 2000;160(5):649-657. doi:10.1001/archinte.160.5.649
Copyright 2000 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2000
Although much has been written about the ethics of new methods of health care financing, little is known about the extent to which physicians experience these cost-control arrangements as ethical problems.
A cross-sectional telephone survey of 1549 physicians, 8 to 17 years after residency, randomly selected from 75 US metropolitan service areas (response rate, 74.0%).
Only 17.0% believed that financial incentives to limit services are ethically acceptable. Although 52.9% thought that physicians should try to abide by guidelines discouraging the use of interventions with possible but unproven benefit, only 14.5% thought such guidelines should be enforced by payers. Only 5.7% thought that it was morally acceptable for payers to discourage physicians from telling patients about their personal financial incentives, and only 9.1% found compliance with such restrictions morally acceptable. Changes in the health care system in the past 5 years were believed to have had a negative impact on their own patients' trust in them by 50.6%, and 80.8% believed that changes in the health care system in the past decade have diminished physicians' commitment to an ethic of undivided loyalty to patients. In multiple regression analysis, physicians who reported that the overall personal financial incentives in their practices encouraged them to reduce services were significantly more likely to have ethical objections to such incentives, to believe their own patients' trust in them had diminished, and to believe that the ethic of undivided loyalty to patients had diminished.
Many of the methods now commonly used to influence medical decision making are considered ethically objectionable by most midcareer physicians. Whether their ethical disquiet about these arrangements is justified cannot be answered from these data.
THE MANAGED care revolution has transformed the methods by which physicians are paid, and it is subjecting their clinical decisions to unprecedented levels of supervision. Many primary care and specialist physicians now have financial incentives to reduce the number of tests, treatments, and referrals they order for their patients,1- 3 and growing numbers of physicians are being encouraged to adhere to health plans' clinical practice guidelines.4,5 The world of fee-for-service reimbursement and informal, peer supervision is being supplanted by capitated payment,6,7 fee withholding, and bonuses tied to reduced clinical spending8; cost-conscious "gatekeeping" by primary care physicians9,10; and efforts by health care plans to discourage physicians from telling patients about expensive treatment options or their personal financial incentives.11,12
Proponents of these well-chronicled changes contend that at least some of these measures are necessary to control spiraling costs and to allocate more rationally the greater than $1 trillion that Americans now spend annually on medical care. They offer a vision of health care as a centrally managed set of services, quality controlled, evidence based, and designed with consistent cost-benefit trade-off policies in mind.13- 15
Some contemplate an automotive superstore model of differently priced health plans with varying cost-benefit trade-off policies and standards of care, much as Ford or Chrysler engineers strike different cost-safety balances when designing economy, midsized, and luxury car models.14,16 Others urge a single standard of medically necessary care, adhered to by health care plans that compete on price, amenities, and other grounds.13 From an ethical perspective, however, all share the following key characteristic: they justify the revolution in physician payment and in the management of clinical decision making as an affirmation of medical practitioners' moral responsibilities as stewards of collective resources.
Physicians, in this vision, can best serve consumers by allying themselves with the allocative and quality management missions of the health care plans in which they participate.17 Many commentators, believing that bedside rationing has always been part of the physician's job, hold that the only realistic way to implement a spirit of stewardship is to encourage physicians to make cost-conscious allocative decisions at the bedside.18- 20
Bedside responsibility for the management of collective resources fits awkwardly, and at times clashes sharply, with the Hippocratic ideal of undivided clinical loyalty to patients.21,22 This older conception of professional obligation presents undivided loyalty to patients as essential to physicians' clinical credibility and efficacy.23- 25 Some commentators have argued that the economic exigencies of the time require adjusting this moral standard.26- 29 Others have expressed concern about conflicts of interest and the possible erosion of patient trust as a result of recent changes in health plan payment and management practices.30- 35
Medical practitioners today find themselves caught in an intense ethical and economic crossfire between both conceptions. Surprisingly, although there have been multiple studies of physicians' satisfaction and behavior under the payment and management systems being introduced by managed health care plans,9,36- 41 there are almost no data concerning physicians' beliefs about the ethical acceptability of these changes, ie, whether physicians believe these changes to be morally right and morally justifiable. We therefore conducted a national survey of more than 1500 practicing physicians and asked about their ethical beliefs regarding financial incentives and other management practices of health care plans. We also inquired about their perceptions of how recent changes in the American health care system have affected patient trust in physicians and the fundamental ethical principles of the medical profession.
The Young Physicians Survey (YPS) of 1991 was designed to elicit opinions regarding the medical practice environment and to assess the career satisfaction of physicians who had recently entered practice. This survey was designed to represent all allopathic physicians (and all osteopathic physicians completing allopathic residencies) who were younger than 45 years and had from 2 to 9 years of practice experience in 1991. This survey was based on a complex sampling design, including weights that adjust for sampling design and correct estimates to reflect the American Medical Association Physician Masterfile distribution of young physicians by age, sex, and country of medical education (United States or other). The survey design, including pretesting and weight construction, is described in detail elsewhere.42 In 1997, we resurveyed a subsample of the physicians who had participated in the YPS in 1991. Eligible physicians in 1997 had completed residency training 8 to 17 years before the survey, practiced in one of the 75 largest metropolitan service areas (MSAs) in the United States, were younger than 52 years, and spent at least 20 hours per week in patient care.
The survey consisted of 102 closed-ended questions. A number of questions asked during the 1991 YPS were asked again to compare responses and to assess temporal changes since the 1991 survey. Additional questions about payment structures, financial incentives, moral beliefs, and attitudes were included in our survey. Care was taken to minimize bias in the language of these new questions. For items hypothesized to generate negative responses, the statements were framed positively to minimize the extent to which measured negative responses would reflect framing.43 Response categories were balanced, neutral response options were given, and negative responses were placed last (eg, statements were of the form, "X is ethical," and responses were of the form, agree, neither agree nor disagree, and disagree). These new questions were first pretested by the research team with a focus group of general internists, general pediatricians, and general surgeons, and subsequently revised. Further revisions of the questions were made after the interview schedule was field tested by professional survey designers from CODA, Inc (Silver Spring, Md) for length, clarity, comprehension, and validity using the technique of cognitive pretesting. The exact wordings of the questions about ethics are given in Table 1.
Since contractual arrangements for individual practitioners are often multiple and complex, any individual, regardless of health maintenance organization (HMO) penetration in the area, may have many or a few patients in managed care contracts. We therefore asked physicians, "How would you describe your overall personal financial incentives in your [main] practice? On balance do these incentives . . . favor reducing services to patients? favor expanding services to patients? favor neither?" We have shown elsewhere that this item has construct validity, as evidenced by the fact that it correlates well with related constructs, such as physicians' percentage of patients in managed care plans, their percentage of patients in capitated arrangements, and percentage of physicians reporting capitation as the financial arrangement of the dominant plan in which they participate. This item is also correlated with the HMO penetration in the MSA.44
Trained professional interviewers from Mathematica Policy Research, Inc, Princeton, NJ, using their computer-assisted telephone interviewing system, conducted telephone interviews. Physicians' offices were contacted in advance and appointments were made to conduct the interviews at a time convenient for the physicians. Up to 30 telephone calls were made to attempt to conduct each interview. The average number of calls per completed interview was 19. Physicians who participated were not compensated. The interview schedule took approximately 30 minutes to administer.
The study was approved by the Institutional Review Board of the Georgetown University School of Medicine, Washington, DC.
Because some response categories generated few responses, the "agree strongly" and "agree somewhat" responses to the question about the ethics of financial incentives were collapsed to generate an ordered, categorical, dependent variable consisting of the responses of "agree," "neither agree nor disagree," "disagree somewhat," and "disagree strongly." An ordered logit model was estimated for this variable. Independent variables consisted of data regarding the physicians' sociodemographic characteristics, practice characteristics, payment structure, financial incentives, reasons influencing their choice of type of practice, and characteristics of their medical school education (the latter 2 drawn from the physicians' responses on the 1991 YPS).
Similarly, because there were few responses in some of the categories, the "very positively," "somewhat positively," and "not affected one way or the other" responses to the question about trust were collapsed, as were the "somewhat negatively" and "very negatively" responses, to create a dichotomous dependent variable. The responses to the question about the ethic of undivided loyalty were also dichotomized into "same or increased" vs "decreased" responses. Logistic regression models were estimated for both dependent variables using the same independent variables as above.
Interviews were completed for 1549 physicians. The response rate was 74.0%. Of the respondents, 75.9% were men and 69.5% were non-Hispanic white. Although 35.5% were generalists (general practitioners, family physicians, general internists, or general pediatricians), 58.2% reported that they provided at least some primary care. Only 5.7% were employed in staff-model HMOs. In all, 66.0% worked in physician-owned practices, and the physician was the sole owner of 29.5% of these practices. Of the respondents, 81.1% contracted with at least 1 preferred provider organization, 54.2% contracted with at least 1 independent practice association, and 74.5% contracted with at least 1 HMO. A "formal gatekeeping arrangement" (ie, an arrangement that controlled patient entry into the health care system) was reported by 57.7%. Respondents did not differ significantly from nonrespondents in ethnicity, sex, or their answers to the 1991 YPS questions.
As shown in Figure 1, more than three quarters of the physicians believe that personal financial incentives to encourage restraint in testing, treatment, and referrals are not ethically acceptable. Similarly, more than three quarters believe that professional commitment to the traditional medical ethic of undivided loyalty to patients has diminished during the past decade (Figure 2). More than half believe that their own patients' trust in them has diminished in the past 5 years (Figure 3).
For interventions of unproven benefit, 52.9% of physicians were neutral or agreed that physicians ought to comply with guidelines to discourage their use, and 55.0% were neutral or agreed that physicians ought to comply with guidelines discouraging interventions of small, proven benefit. However, 79.8% and 76.6% of physicians, respectively, were opposed to payers enforcing either of these types of guidelines.
Most physicians (87.3%) disagreed strongly that efforts by health care payers to discourage physicians from telling patients about coverage restrictions were ethically acceptable, and 78.4% disagreed strongly that efforts to discourage disclosure of physician financial incentives were ethically acceptable. However, the physicians were somewhat less ready to deem physician compliance with such efforts ethically unacceptable. Only 69.3% disagreed strongly that physician compliance with efforts to discourage disclosure of coverage limitations is ethically acceptable, and only 67.5% disagreed strongly that physician compliance with efforts to discourage disclosure of financial incentives is ethically acceptable. Nonetheless, only 7.2% agreed that physician compliance with efforts to limit disclosure of coverage limitations would be ethically acceptable, and only 5.7% agreed that it would be ethically acceptable for physicians to limit disclosure of their financial incentives even if encouraged to do so by payers.
We estimated multivariate regression models to identify factors associated with physicians' beliefs about the ethics of financial incentives to control costs, the profession's commitment to the ethic of undivided loyalty to patients, and their own patients' trust in them. As shown in Table 2, physicians reporting an overall personal financial incentive to provide fewer services to patients in their practices, physicians who had changed their practices since 1991 but stayed in the same MSA, white Hispanic physicians, and female physicians were significantly and independently more likely to find financial incentives to limit services morally unacceptable. Physicians in primary care specialties and those who had at least 1 parent who was a physician were less likely to find financial incentives to limit services morally unacceptable. We tested for interaction between being a primary care physician and reported financial incentives, but found none. There was a trend for more primary care physicians than specialists to report financial incentives to reduce services (17% vs 15%; P=.07). However, primary care physicians with incentives to reduce or to increase services held statistically equivalent views regarding the ethics of financial incentives once the incentives themselves were controlled for in the model.
A logistic regression model of factors associated with physicians' belief that commitment to the traditional ethic of undivided physician loyalty to patients has diminished in the past 10 years is shown in Table 3. Physicians reporting overall financial incentives to provide less testing, treatment, and referrals for patients in their practices and physicians working in areas with a high HMO penetration were more likely to report that commitment to the traditional ethic of undivided loyalty had diminished. Those working in staff-model HMOs were less likely to report such a change.
As shown in Table 4, physicians reporting overall financial incentives to provide fewer tests, treatments, and referrals in their practices and physicians practicing in settings other than group practices, HMOs, hospitals, or hospital clinics were more likely to report diminished trust in them on the part of their patients. Female physicians and African American physicians were less likely to report that their patients' trust in them had diminished.
American medicine today faces the following fundamental, unresolved ethical question: Should the profession recommit itself, in the face of the managed care revolution, to the Hippocratic ideal of undivided loyalty to individual patients, or should the profession reconceive its role as one of balancing 2 commitments, ie, the care of the patient and the stewardship of collective health care resources? The latter conception fits more or less neatly with the ongoing transformation of the approaches of health care plans to the payment and management of physicians. The former, traditional vision is sharply at odds with this transformation.
This study is the first national survey examining the actual ethical beliefs of practicing physicians regarding this transformation. The data reveal deep moral disquiet among this cohort of seasoned but still relatively young physicians regarding these new financial arrangements. Although many ethicists have argued that personal financial incentives for physicians to limit tests, treatments, and referrals are ethically acceptable,26- 29 only 17% of our respondents agreed. The physicians' responses suggest that a seismic shift is taking place in the moral foundations of the profession. Fully 80.8% believe that commitment to the traditional ethic of undivided loyalty to patients has decreased, and more than half believe that their own patients' trust in them has diminished during the past 5 years.
These findings might be discounted by some as nonspecific indicators of physician dissatisfaction with change, or as a reaction to diminished professional autonomy. Nonetheless, the tension between competing obligations is sharper whenever physicians are encouraged to take account of the needs of third parties when making treatment decisions. It seems reasonable for physicians to have moral concerns about this situation. Further, our study is consistent with qualitative data from Jecker and Jonsen,45 who conducted 3 focus groups of physicians in Sacramento, Calif. They found that these physicians believed that financial incentives to limit tests, treatments, and referrals were associated with diminished patient trust in them and even led some physicians to question their own motives when they thought it medically appropriate to withhold certain interventions.
In multivariate regression analyses, physicians who reported financial incentives to limit tests, treatments, and referrals were significantly and independently more likely to find such financial incentives morally troubling, to believe that commitment to the ethic of undivided loyalty to patients has eroded, and to report diminished patient trust in them. Although still relatively new, the use of such financial incentives by managed care organizations appears to be increasing.1,46 It remains to be shown whether these incentives have any major impact on physician behavior or on health outcomes.40,47,48
Donelan et al49 have found that physician reports of troubling financial incentives are correlated with HMO penetration. Grumbach et al50 have demonstrated that physicians with financial incentives based on productivity are less satisfied with their practices. We have shown that although all groups of physicians find these arrangements ethically troubling, physicians who report that they practice under personal financial incentives to reduce services find these arrangements more ethically troubling than their colleagues who do not practice under such circumstances. Physicians with financial incentives to reduce services may be experiencing especially acute moral tension and confusion. They may feel caught between appeals to their financial self-interest designed to limit interventions for their patients, and appeals to their traditional professional ethic of mitigating self-interest to safeguard their patients.28,31 An abrupt reversal of incentive structures may come as a shock to physicians, perhaps the most concrete signal they could receive that the traditional moral order in medicine is being challenged. Physicians imbued with a sense of professional identity that has stressed altruism and loyalty to patients51,52 may experience the appeal to financial self-interest as a means of controlling costs as a direct threat to their self-understanding as professionals.
Female physicians were more likely to feel that financial incentives to limit care were unethical. Another recent survey found that female physicians consistently had more negative attitudes toward managed care.53 Perhaps this is partially explained by recent work in moral psychology and the tendency of women to emphasize care over justice in their responses to morally troubling situations.54 That physicians who moved practices were more likely to find financial incentives to reduce care ethically troubling seems understandable. They might have moved to escape such incentives, or the move might newly have caused them to confront the ethical problems such incentives can pose. Perhaps physicians who have at least 1 parent who is a physician find these incentives less troubling because they have grown up with firsthand knowledge of the financial aspects of medical practice.
Physicians with financial incentives to reduce services were also more likely to believe that the ethic of undivided loyalty to patients has diminished within the profession during the past 10 years. Those who worked in staff-model HMOs were less likely to believe this. However, these are not incompatible findings. Most staff-model HMOs are established, older enterprises with salaried physicians. They have generally not used financial incentives to reduce services. Ironically, then, physicians employed in staff-model HMOs have probably been relatively shielded from the changes that have been transforming health care financing. Most of the growth in managed care has come in the form of new HMOs "without walls." These enterprises have been more likely to manage practice through financial incentives as well as other utilization review techniques. Thus, it makes sense that our model shows that overall HMO market penetration in a region is associated with the belief that the ethic of undivided loyalty to patients has diminished, whereas employment in a staff model HMO is inversely associated with this belief.
So-called gag rules that prohibit physicians from disclosing to patients coverage restrictions or the financial arrangements between physicians and insurers have been sharply criticized.12,55 Although explicit contractual gag rules now are widely outlawed and are probably unusual in the current generation of contracts, more subtle strategies for discouraging such disclosures have been reported.11,12,55 These might include, for example, clinical practice profiling in conjunction with suggestions that failure to meet the company's behavioral expectations could result in deselection.11 Physicians in our survey were almost unanimous in their opposition to these practices on the part of payers.
Respondents' only marginally less intense ethical opposition to physician compliance with such efforts suggests that physicians expect each other to resist any efforts by payers to limit what they may disclose to patients. These data suggest the potential for sharp conflict if managed care organizations attempt to influence what physicians tell their patients.
Among our respondents, there was considerably more ethical acceptance of clinical practice guidelines than of direct financial incentives, although there was still considerable opposition to guidelines. A recent survey has suggested that some of the opposition to guidelines may be attributed to physicians' perceptions that they are primarily instruments for cost-control, not quality improvement.56 Ethicists have also questioned the moral status of practice guidelines.57,58
Our survey respondents were nearly evenly divided in their opinions about physician adherence to clinical guidelines discouraging interventions with "possible but unproven efficacy" and those "that have a small, proven advantage over standard interventions but cost much more." This suggests that at least some physicians may be willing to depart from the practice of pursuing all possible clinical advantage for their patients, no matter what the cost. Nonetheless, the fact that so many were even opposed to guidelines about interventions of unproven benefit is perplexing. This may reflect concerns about autonomy. On the other hand, physicians, like many patients, simply may view experimental treatments as uniformly beneficial even before this has been clearly proven. Very few were willing to endorse the proposition that either type of guideline "should, in general, be enforced by health care payers." The fact that more than three fourths of those surveyed objected to enforcement by payers, even if the intervention is not of proven benefit, suggests that concerns about professional autonomy may play a role in the formulation of their ethical beliefs, at least with respect to this issue.
The fact that most of the physicians we surveyed reported the belief that changes in their work environments during the previous 5 years have had a negative affect on their own patients' trust in them is noteworthy. Trust is at the center of the physician-patient relationship.33,59
Although our study reports physicians' subjective impressions regarding their patients' trust in them, our findings are consistent with recent direct surveys of patients.60 Further, patient reports regarding their trust in their physicians have been shown to be tightly correlated with the perceptions of physicians, lending credibility to the reports of the physicians in our survey.61
On multivariate analysis, physicians who reported an overall financial incentive to limit services were significantly more likely to report the belief that their patients' trust in them has diminished. Our data cannot establish a cause-and-effect relationship between these phenomena. However, it stands to reason that the dual agency inherent in these arrangements might feed patient skepticism about their physicians' loyalty.22,31,32,34,35 Also, evidence is emerging that patients in capitated and prepaid plans do, in fact, have less trust in their physicians.58,62
It might be argued that outcomes research could restore (or sustain) patient trust if it could be shown conclusively that outcomes do not differ systematically according to arrangements between plans and physicians. However, across-the-board differences in clinical outcomes would be exceedingly difficult to show empirically. Moreover, trust and trustworthiness hinge on more than the equivalence of clinical outcomes: the perception and the reality of divided loyalty and conflict of interest matter morally and psychologically.22 Virtues such as trustworthiness, fidelity, altruism, and honesty cannot be re-cast in terms of outcomes.63
This study sampled a cohort of physicians who have been in practice 8 to 17 years. Arguably, these are the physicians who have witnessed the greatest degree of change in health care during their professional lives and they are therefore unrepresentative of physicians as a whole. On the other hand, this study has substantial generalizability because, to our knowledge, it is the first national, random sample of physicians regarding these issues. Although this cohort may not be representative of all physicians, they are a cohort critical to the health care system, ie, those at or entering the prime of their careers. What they think is therefore extremely important.
We also recognize that we have surveyed all types of physicians, not just primary care physicians. Some might discount the views of specialists, since they are not as directly affected by current changes in health care financing. However, our specialty distribution reflects that of the entire nation. Like many specialists, 58.2% of those in our survey report practicing at least some primary care, and there are systems in which specialists, not generalists, are capitated.64
Our data depict a potentially very unstable situation that has received insufficient attention from policymakers and bears careful monitoring. Methods that health care payers now widely use to control clinical spending are sharply at odds with physicians' ethical beliefs and values. These data do not answer the question of whether the traditional ethic should be preserved or modified, or whether the ethical issues in managed care are worse than those in fee-for-service. Rather, they underscore the gravity of the issues and the depth of ethical turmoil physicians now experience.
Some commentators have suggested strategies by which managed care organizations and regulators might restructure health care financing while preserving patient trust. Such suggestions have included the development of ethical standards in the industry, promoting mission-oriented, nonprofit managed care organizations; increasing physician control; effective marketing; improving physician communication skills; patient education; and government regulation.65- 67 Our findings highlight the need to explore all possible solutions to this problem.
In addressing this dilemma, the profession now confronts a defining moral question: should it reassert its allegiance to an ethic of undivided fidelity to individual patients, or should it adopt a new ethic that takes on the role of stewardship of collective medical resources?68 Potent moral and public policy arguments can be marshaled on both sides, reflecting the importance of such social goods as the preservation of scarce resources, fair allocation of benefits and burdens, and the care, solidarity, and fidelity that we owe to persons in crisis. Neither denial of economic reality nor dismissal of the importance of trust at the bedside contribute usefully to the debate.
This survey bears repeating. If the managed care revolution continues unabated, our findings might only represent a transient phenomenon, and physicians might show themselves quite willing to change their ethical attitudes. Depending on one's ethical perspective, such a change could be interpreted as moral progress or accommodation. On the other hand, physicians' sense of moral disquiet might only intensify with time.
Careful research and monitoring of changes in the structure of the health care system, the attitudes and beliefs of physicians and patients, and the impact of these changes on patient care will help to inform debates about the future of health care financing.
Accepted for publication May 27, 1999.
This study was supported in part by a grant from the Robert Wood Johnson Foundation, Princeton, NJ, and grant No. 17-C-90395 from the Agency for Healthcare Research and Quality, Rockville, Md.
We thank Andy Epstein, MPP, and Tomer Seifan for their technical assistance.
Reprints: Daniel P. Sulmasy, OFM, MD, PhD, Sisters of Charity Chair in Ethics, John J. Conley Department of Ethics, Saint Vincents Hospital and Medical Center, 153 W 11th St, New York, NY 10011 (e-mail firstname.lastname@example.org).