Kao AC, Green DC, Zaslavsky AM, Koplan JP, Cleary PD. The Relationship Between Method of Physician Payment and Patient Trust. JAMA. 1998;280(19):1708-1714. doi:10.1001/jama.280.19.1708
From the Institute for Ethics, American Medical Association, Chicago, Ill (Dr Kao); the Prudential Center for Health Care Research, Atlanta, Ga (Drs Green, Koplan, and Cleary); and the Department of Health Care Policy, Harvard Medical School, Boston, Mass (Drs Kao, Zaslavsky, and Cleary)
The Patient-Physician Relationship section editor: Richard M. Glass,
MD, Deputy Editor, JAMA.
Context.— Trust is the cornerstone of the patient-physician relationship. Payment
methods that place physicians at financial risk have raised concerns about
patients' trust in physicians to act in patients' best interests.
Objective.— To evaluate the extent to which methods of physician payment are related
to patient trust.
Design.— Cross-sectional telephone interview survey done between January and
Setting.— Health plans of a large national insurer in Atlanta, Ga, the Baltimore,
Md–Washington, DC, area, and Orlando, Fla.
Participants.— A total of 2086 adult managed care and indemnity patients.
Main Outcome Measure.— A 10-item scale (α=.94) assessing patients' trust in physicians.
Results.— More fee-for-service (FFS) indemnity patients (94%) completely or mostly
trust their physicians to "put their health and well-being above keeping down
the health plan's costs" than salary (77%), capitated (83%), or FFS managed
care patients (85%) (P<.001 for pairwise comparisons).
In multivariate analyses that adjusted for potentially confounding factors,
FFS indemnity patients also had higher scores on the 10-item trust scale than
salary (P<.001), capitated (P<.001), or FFS managed care patients (P<.01).
The effects of payment method on patient trust were reduced when a measure
based on patients' reports about physician behavior (eg, Does your physician
take enough time to answer your questions?) was included in the regression
analyses, but the differences remained statistically significant, except for
the comparison between FFS managed care and FFS indemnity patients (P=.08). Patients' perceptions of how their physicians were
paid were not independently associated with trust, but the 37.7% who said
they did not know how their physicians were paid had higher levels of trust
than other patients (P<.01). A total of 30.2%
of patients were incorrect about their physicians' method of payment.
Conclusions.— Most patients trusted their physicians, but FFS indemnity patients have
higher levels of trust than salary, capitated, or FFS managed care patients.
Patients' reports of physician behavior accounted for part of the variation
in patients' trust in physicians who are paid differently. The impact of payment
methods on patient trust may be mediated partly by physician behavior.
THE PATIENT-PHYSICIAN relationship is built on a foundation of trust.1- 4 Many
health plans have implemented payment methods that place physicians at financial
risk for providing medical services. This shift in the way that physicians
are paid has raised issues about the impact of payment methods on patient
Many practice and physician characteristics might influence patient trust,11,12 but many people are most concerned
about the potential effects of payment methods on physician behavior and decision
and how such changes could influence patients' trust in their physicians.
Trust also could be directly shaped by patients' perceptions of, and attitudes
about, how their physicians are paid.17,18
Despite the importance of these issues, no empirical studies have examined
the association between methods of physician payment and patient trust. We
surveyed adult managed care and indemnity patients to address 2 key questions:
(1) Does trust vary among patients with different perceptions of how their
physicians are paid? (2) Does trust vary among patients whose physicians are
paid differently? There is frequent discussion in the popular press about
financial incentives to provide less care. Thus, we hypothesized that patients
who thought their physicians were capitated would be least likely to trust
them. We thought that the effect of perceived method would be stronger than
the effect of actual method of physician payment.
Eligible patients were adults (≥18 years) who had a primary care
(family practice, internal medicine, or obstetrics/gynecology) physician visit
in 1995 and were enrolled in managed care or indemnity plans of a large, national
health insurer in 3 health insurance markets (Atlanta, Ga, the Baltimore,
Md–Washington, DC, area, and Orlando, Fla). In each of these markets,
less than 30% of the physicians are solo practitioners, consistent with practice
arrangements nationally.19 In Atlanta and Orlando,
salaried physicians were part of medical groups with exclusive contracting,
while those in the Baltimore-Washington area were in group practices that
contracted with multiple payers.
Patients were selected using a 2-stage, stratified sample design. We
classified all primary care physicians based on their payment method for eligible
patients. In this study, eligible patients include those whose physicians
were paid on a salary (category 1), capitated (category 2), fee-for-service
(FFS) managed care (category 3), or FFS indemnity (category 4) basis when
providing care to individual patients. For the remainder of this article,
we refer to patients of salaried physicians as salary patients, etc. Except
for those who were salaried, physicians in this study could have patients
in multiple payment method categories.
All physicians with at least 8 eligible patients were included in the
study, with 2 exceptions. We sampled Atlanta physicians in category 4 because
of the large proportion of physicians in that payment category in Atlanta.
Similarly, there was a large proportion of salaried physicians in the Baltimore-Washington
area, so we sampled those physicians.
We drew a disproportionate probability sample of patients from the practice
of each sampled physician. We were particularly interested in comparisons
between managed care patients of physicians who were paid on a capitated or
FFS basis because of the prevalence of these arrangements19
and the difference in financial risk to the physician. Therefore, we sampled
approximately twice as many capitated and FFS managed care patients as other
Available measures of patient trust20,21
have not been used to evaluate whether trust differs by physician payment
method. Our 10-item scale (coefficient α=.94) was developed using psychometric
analyses of data from about 300 patients in a pilot study (Table 1). It comprises some modified questions from the Trust-in-Physician
scale20 and questions about issues such as
access to specialists and informing patients about medical options.
Patients' trust in physicians can be influenced by other attitudes.
Thus, we asked about trust in one's individual health plan (How much do you
trust [name of health plan] to control costs without adversely affecting the
quality of your care?), trust in health maintenance organizations (HMOs) in
general (How much do you trust HMOs in general to control costs without adversely
affecting the quality of your care?), and general trust in people.5,8,22- 27
The Benevolence-of-People scale28 was used
to assess patients' general trust in people.
Respondents were asked to identify their physician's method of payment
as (1) salary (physician's pay is based on a straight salary), (2) capitation
(physician's pay is based on some fixed monthly amount, which is dependent
on the number of patients in the physician's practice), or (3) FFS (physician's
pay is based on the number of office visits). Even though FFS payment was
used in managed care and indemnity settings, respondents were not asked to
differentiate between their physician's payment method in these 2 settings
because the inherent financial incentives were comparable.
Modified questions from the Picker survey on patient-centered care29- 31 that ask about patients'
experiences with their physicians also were included (Table 2). We also asked patients, "Have you ever consider changing
your doctor because of concerns about the care provided by Dr ____?" The questionnaire
asked about patients' race, education, income, health status, length of physician
relationship, and whether they had enough choice when selecting a physician.
Patients' age, sex, and length of plan enrollment were obtained from administrative
data. Information about physicians' age, sex, board certification, international
medical graduate status, and medical practice size were abstracted from the
American Medical Association's 1997 Physician Masterfile and the insurer's
We used t tests to compare patient trust by
actual and perceived method of physician payment. We also compared patients'
trust in their health plan or HMOs by actual and perceived method of payment.
The SEs were adjusted to account for the sample design.
To further assess the relationship between patient trust and payment
methods, we estimated linear regression models with Huber corrections to adjust
statistical tests for intracluster correlation (STATA, Version 5.0, STATA
Corp, College Station, Tex). Variables that were not significant at the .10
level and were not conceptually relevant were removed from the multivariate
models using backward selection.32 The dependent
variable in these models was the patient trust score. We conceptualized the
effect of payment methods on trust as both an indirect effect mediated by
physician behavior and a direct effect. In the behavior-dependent pathway,
payment methods influence physician behavior and decision making that in turn
influence patients' trust in their physicians. The average response to the
Picker questions was used to assess how well patient reports of physician
behaviors explained the relationship between payment methods and patient trust.
In the behavior-independent pathway, patients' perceptions of how their
physicians are paid directly affect their trust. We statistically controlled
for patients' perceptions of how their physicians are paid to assess the behavior-dependent
hypothesis. A variable indicating whether a patient correctly identified his/her
physician's method of payment was included in the regression analysis.
Other hypothesized predictors of patient trust used in the multivariate
analyses included patients' trust in their health plan, trust in HMOs in general,
and general trust in people. Patient characteristics included as covariates
were age, sex, race, education, income, and self-reported health status. Physician
characteristics included were age, sex, board certification, and international
medical graduate. Other independent variables were perceived choice of physicians,
length of physician relationship, and length of plan enrollment.
One of the difficulties in studying the effects of financial incentives
is that payment methods are related to structural practice characteristics
that may also influence patient trust. For example, physicians who are paid
by salary tend to practice in staff-model HMOs or group practices. We included
number of physicians in the practice in the models to partially control for
such effects. A variable also was included to control for potential regional
effects. Independent variables with missing data were imputed using regression
models. To assess the relationship between patient trust and patients' stated
intention to change physician, we estimated a logistic regression model in
which intention to change was the dependent variable.
Of the 4448 patients we attempted to telephone from January 1997 through
June 1997, 2733 patients were contacted and screened for eligibility. Of those
screened, 602 patients (22%) were ineligible. Among those who were ineligible,
303 were no longer enrolled in the health plan, 94 could not be interviewed
for reasons such as language problems and hearing difficulties, 114 could
not identify a physician as their "regular doctor," and 91 had no opinion
about their physician. Among the eligible patients, 2086 telephone interviews
were completed, for an estimated response rate of 60.1%, assuming the ineligibility
rate in the nonscreened group was comparable with that in the screened group.
Respondents were older on average than nonrespondents (46 vs 43 years; P<.01) and more likely to be women (69% vs 64%; P<.01). Respondents also were enrolled in the health
insurance plan for a shorter period than nonrespondents (7.0 vs 7.7 years; P<.05).
Compared with indemnity patients (Table 3), managed care patients (salary, capitated, and FFS managed
care patients) were younger, less likely to be white, less likely to have
an annual income higher than $45,000, and enrolled in the health insurance
plan for a shorter period. The FFS patients in managed care and indemnity
arrangements were more likely to be women and more likely to have graduated
from college. Patients with salaried physicians also tended to report slightly
poorer health status than other patients.
Approximately one third (32.1%) of patients correctly identified their
physician's payment method, while about one third of patients were incorrect
(30.2%), and the other patients (37.7%) said they did not know how their physician
was paid. Most patients (84%) completely or mostly trusted their physicians
(Table 4). However, more FFS
indemnity patients (94%) trusted their physicians to put their health and
well-being above keeping down the health plan's costs than salary (77%), capitated
(83%), or FFS (85%) managed care patients. Patient responses to other trust
questions were comparable with the overall patient trust question (results
not presented) except that at least 98% of each group of patients trusted
their physician to keep medical information private.
There were no significant differences in responses to the overall patient
trust question (Table 4) or the
other trust questions among groups of patients with different perceived methods
of physician payment. However, patients who said they did not know how their
physician was paid had more positive responses to the overall trust question
than other patients (89% vs 81%; P<.001).
The average patient trust score for salary patients was 4.32 (SD=.74).
The average for capitated patients was 4.44 (SD=.68), for FFS managed care
patients was 4.53 (SD=.60), and for FFS indemnity patients was 4.73 (SD=.38).
A difference of 0.25 points in the patient trust score is associated with
a doubling of the odds that patients said they had considered changing physicians.
Patients' trust in their health plan did not differ significantly by
payment method (Table 4), except
that FFS indemnity patients were more likely to trust their health plan than
managed care patients. The FFS indemnity patients were less likely to trust
HMOs in general than salary or capitated patients. Regardless of payment method,
patients' trust in their health plan or HMOs was lower than trust in their
Patients who thought their physicians were paid on an FFS basis were
more likely to trust their health plan to control costs without adversely
affecting the quality of care than patients who thought their physicians'
fees were capitated. There were no other significant differences in patients'
trust in their health plan or HMOs based on perceived method of physician
In initial multivariate analyses, patient age, education, income, length
of enrollment, physician age, physician board certification, and site were
not significant predictors of trust and were excluded from subsequent analyses.
In the final multivariate model, FFS indemnity patients had significantly
higher levels of trust than salary, capitated, or FFS managed care patients
(Table 5). Patients who thought
their physicians were paid on an FFS basis did not have higher levels of trust
than those who thought their physicians were paid on a salaried or capitated
basis. However, those who said they did not know how their physicians were
paid were more likely to trust them than were other patients. Patients of
salaried physicians who correctly identified their physician's method of payment
had higher levels of trust than other patients.
Patients' trust in their health plan or HMOs was positively associated
with patients' trust in their physician. Healthier patients tended to trust
their physicians more than patients who reported poorer health status. Patients
had higher levels of trust in physicians who were graduates of US medical
schools. Patients of physicians in 2-physician practices (but not those of
solo practitioners) had higher levels of trust in physicians than those of
physicians in larger medical groups. Patients who said they had enough physician
choice when selecting their physician were more likely to trust them than
other patients. Longer patient-physician relationships also were associated
with higher levels of patient trust.
When the Picker score was included in the regression model, the effects
of payment on patient trust were reduced, but still significant, except for
the difference between FFS managed care and FFS indemnity patients. This suggests
that part of the difference in trust among patients with different physician
payment methods was due to differences in health care experiences.
The associations between trust and perceived payment method were not
affected, except that the difference in trust between those who said they
did not know how their physician was reimbursed and other patients was no
longer statistically significant. This is probably because patients who do
not know how their physician is paid report fewer problems in response to
the Picker questions (results not shown). It may be that patients with fewer
problems are less likely to be concerned about payment issues.
The associations between trust in physician and trust in the health
plan and in HMOs in general were reduced, again suggesting that part of these
associations was because those with less trust in HMOs and their health plan
had experienced more problems than other patients. The association between
patient trust and respondents' belief in the benevolence of people was substantively
similar after the Picker score was added, confirming that general attitudes
about others was not a confounding factor for the associations studied.
The association between patient trust and health status was reduced
when the Picker score was added to the model. This was expected because sicker
patients generally report more problems with their care.29,30
The associations between sex and race and trust became somewhat stronger and
statistically significant in the model including the Picker score, but the
changes in the coefficients were small. Race and sex were not significantly
associated with the Picker score.
In the regression model including the Picker score, the associations
between trust and practice size, physician choice, length of relationship,
and having an international medical graduate as a physician were all reduced,
suggesting that part of these associations were because these variables also
are associated with patients experiencing more problems with care, although
some of these associations are not significant.
Almost 40 years ago, Roemer33 wrote of
the need for studies of the full consequences of different systems of paying
the physician. Few issues in health care have provoked such heated controversy
recently as the method of paying physicians for their services. Capitation
and other at-risk payment methods are designed to encourage cost-conscious
medical practice. Some argue that payment methods like capitation could adversely
affect individual patient care by reducing use of diagnostic tests, specialist
referrals, and other medical services.34,35
Others contend that capitation may contribute to better allocation of finite
health care resources by encouraging use of preventive services36
and discouraging use of expensive medical services that have little diagnostic
or therapeutic value.
Regardless of payment method, the overwhelming majority of patients
trust their physicians, but we found that capitated patients had lower levels
of trust in their physicians than FFS indemnity patients did. The unadjusted
difference of about 0.3 points on the trust scale is substantively important.
A difference of 0.25 points corresponds, on average, to a doubling of the
odds that patients said they had considered changing their physician. There
was a comparable and significant difference in trust after controlling for
patients' perception of how their physicians were paid. Thus, the lower trust
of capitated patients is not only due to lower trust among patients who think
their physicians are paid on a capitated basis.
The lack of a direct link between patient trust and patients' perceived
method of physician payment supports the hypothesis that the differences in
trust by payment method may be mediated by patients' experiences with physician
behavior. We found that differences in patient trust by payment method were
reduced when the Picker score was included in the regression analyses. These
findings suggest that the adverse impact of capitation on patient trust may
be partially due to differences in physician behavior. This is consistent
with research showing that physicians are less satisfied with their ability
to treat patients in capitated plans according to their best judgment.37
Although the potential impact of capitation on the patient-physician
relationship has attracted the most public attention, we also found that salary
patients had less trust in their physicians than FFS indemnity and capitated
patients. Salaried physicians have no explicit financial incentive or disincentive
for the use of medical services.38 One possible
explanation for these payment differences in patient trust may be the contractual
arrangements in multitiered systems.11,39
Many physicians are members of a medical group, and the health plan or insurer
contracts directly with the group. Because the medical group is often capitated
by the health plan or insurer, physicians are usually paid by capitation or
this study, some salaried physicians were part of capitated medical groups.
These physicians, especially equity partners, may be aware of the financial
incentives faced by their group and behave in accordance with those incentives.
Furthermore, in each of the cities studied, salaried physicians also had a
bonus or withholding arrangement that was based on board certification, patient
satisfaction, financial performance, or specialty pools. Thus, they might
have had more explicit financial incentives to use services efficiently than
physicians paid on a capitated basis.
Other factors that could influence physician behavior and patient trust
include practice characteristics (eg, solo vs group practice),42
rule-based mechanisms (eg, utilization review, guidelines, and protocols),43 and education and norms.11,44- 46
For example, salaried physicians are less likely than other physicians to
be in solo or small-group practices.19 Smaller
practices may provide better continuity of care47,48
and a more personal environment than larger practices. However, solo practitioners
may have less access to informal or "curbside" consultation with colleagues.49- 51 To control for such
factors we included physician group size in the multivariate models. The effects
of payment method were significant even after controlling for these effects.
Physicians now must deal with administrative strategies such as utilization
management, clinical practice guidelines, and preauthorization procedures
that were unknown to many of them during their training and their early career.
Administrative strategies like utilization review can improve quality of patient
care by limiting the use of inappropriate medical services, fostering evidence-based
disease management,52 and providing clinical
feedback and education to physicians and patients.53
Utilization review can undermine the clinical autonomy of physicians by deterring
them from recommending certain services.54,55
Because these administrative strategies may be a new experience for many physicians,
they may feel uncomfortable with them. They may also think they compromise
the quality of care.46 This discomfort may
be perceived by the patient and diminish trust. Even when the physician is
comfortable with these strategies, the patient may find the manifestations
of these changes disturbing. Regardless of the clinical appropriateness of
procedures, medications, and hospital days, previous patterns of care created
the expectation that more was better. Greater application of utilization review
and preauthorization in managed care settings40,41,54,56
may partly explain the observed differences in trust between FFS patients
in managed care and indemnity settings.
When there are no financial incentives affecting how care is provided,
as is the case with salaried physicians, health care organizations may rely
more on administrative strategies.11,40
This may partially explain the lower trust of patients with salaried physicians.
The administrative strategies used by managed care organizations and medical
groups deserve more attention in future studies.
One reason patients who said they did not know how their physicians
were paid were more likely to trust their physicians than other patients may
be that those with greater trust in their physicians are less concerned about,
or interested in, how their physicians are paid.57
This may be partly because they have fewer problems with their care. Among
patients of salaried physicians, those who correctly identified their physicians'
payment method had higher levels of trust than those who did not. This may
be because patients who have the most favorable attitudes toward salaried
group practice are most likely to be aware that their physicians are paid
Patients who trusted their health plan were more likely to trust their
physician. While patients' confidence in their health plan may influence patients'
trust in physicians, this relationship is likely to be bidirectional. Thus,
our multivariate model may underestimate the effect of payment method on patient
trust. Patients' trust in their physician was higher than trust in their health
plan. This is consistent with declining trust in all social institutions,23 and a natural inclination to trust an individual
more than an organization. Patients with a longer relationship with their
physician and patients who said they had enough choice of physicians had greater
physician trust than other patients. The former association may be due to
both selection effects (patients who do not trust their physician are less
likely to maintain a relationship) and the well-established finding that repeated
positive interactions engender trust. We are not sure why patients with more
choice trust their physicians more. It may be that such patients have an opportunity
to select physicians who have the characteristics that are important to them
or the exercise of choice gives them a sense of commitment to their physician.
The association between health status and trust may be due to the fact that
patients with more health problems are more likely to have concerns about
the availability of specialty or discretionary services.58
The finding that white patients had higher levels of patient trust than
other patients is consistent with negative views about disparities in the
health care system held by many blacks,59- 62
also manifested in other areas including business, law, and government.23 Patients of non-US medical graduates had lower levels
of trust. Possible explanations for this include patients' attitudes about
physician training, linguistic differences, and cultural or racial differences
between patients and physicians.63,64
We surveyed enrollees of a single large health insurance plan in 4 metropolitan
areas, and our results may not generalize to other health insurance plans
or other cities. However, the sites represented 3 diverse medical care markets
and a wide range of practice settings. Some of the observed differences between
patient groups may partly reflect self-selection. We statistically adjusted
for measured differences in the patient groups, but there may be unmeasured
differences as well. However, the small effect of perceptions of how physicians
are paid argues against significant selection effects and biased reporting
that would account for differences in patient trust based on payment method.
We were not able to quantify and model the effects of secondary incentives
such as withholdings and bonuses. However, salaried physicians in large medical
groups at all 3 sites had bonuses or withholdings as part of their payment.
The findings in this study are consistent with the hypothesis that secondary
incentives encourage less use of medical services and could affect the patient-physician
relationship. Physician behavior and clinical decision making were not directly
assessed. Patient reports of physician conduct accounted for a substantial
amount of the variance in patient trust, but we do not know how much of the
unexplained differences by payment are due to actual differences in physician
Regardless of whether the use of financial incentives promotes more
or less appropriate use of medical services, the discordance between patients'
experience and expectations and recent changes in the management of health
care may lessen trust in their physician. Those who feel that recent changes
will compromise the quality of care might argue that eliminating certain elements
of managed care would correct the situation and increase patient trust. Those
who think these changes are necessary and beneficial would counter that patients
will adapt to, and ultimately accept, the changes with more information and
The American health system is in a period of tremendous flux, with changing
rules, relationships, and expectations for all parties. Despite these changes,
trust will remain a crucial element in the patient-physician relationship,
with a profound impact on the therapeutic process. We need to understand more
about physicians' behaviors and practice patterns that promote trust as well
as how patients' expectations, attitudes, and knowledge affect their trust.
The consequences of trust for medical professionalism and managed care credibility
are too important to allow uninformed perceptions to guide policy.