Assessing the Patient Care Implications of “Concierge” and Other Direct Patient Contracting Practices: A Policy Position Paper From the American College of Physicians

As physicians seek innovative practice models, one that is gaining ground is for practices to contract with patients to pay directly for some or all services—often called cash-only, retainer, boutique, concierge, or direct primary care or specialty care practices. Such descriptions do not reﬂect the variability found in practices. For the purposes of this paper, the American College of Physicians (ACP) deﬁnes a direct patient contracting practice (DPCP) as any practice that directly contracts with patients to pay out-of-pocket for some or all of the services provided by the practice, in lieu of or in addition to traditional insurance arrangements, and/or charges an administrative fee to patients, sometimes called a retainer or concierge fee, often in return for a promise of more personalized and accessible care. This deﬁnition encompasses the practice types previously described. The move to DPCPs is based on the premise that access and quality of care will be improved without third-party payers imposing themselves between the patient and the physician. Yet concerns have been raised that DPCPs may cause access issues for patients who cannot afford to pay directly for care. This ACP position paper, initiated and written by its Medical Practice and Quality Committee and approved by the Board of Regents on 25 July 2015, assesses the impact of DPCPs on access, cost, and quality; discusses principles from the ACP Ethics Manual, Sixth Edition , that should apply to all practice types; and makes recommendations to mitigate any adverse effect on underserved patients. shortages, and the cost and beneﬁts associated with personalized services included in the retainer fee?

As physicians seek innovative practice models, one that is gaining ground is for practices to contract with patients to pay directly for some or all services-often called cash-only, retainer, boutique, concierge, or direct primary care or specialty care practices.
Such descriptions do not reflect the variability found in practices. For the purposes of this paper, the American College of Physicians (ACP) defines a direct patient contracting practice (DPCP) as any practice that directly contracts with patients to pay out-of-pocket for some or all of the services provided by the practice, in lieu of or in addition to traditional insurance arrangements, and/or charges an administrative fee to patients, sometimes called a retainer or concierge fee, often in return for a promise of more personalized and accessible care. This definition encompasses the practice types previously described.
The move to DPCPs is based on the premise that access and quality of care will be improved without third-party payers im-posing themselves between the patient and the physician. Yet concerns have been raised that DPCPs may cause access issues for patients who cannot afford to pay directly for care.
This ACP position paper, initiated and written by its Medical Practice and Quality Committee and approved by the Board of Regents on 25 July 2015, assesses the impact of DPCPs on access, cost, and quality; discusses principles from the ACP Ethics Manual, Sixth Edition, that should apply to all practice types; and makes recommendations to mitigate any adverse effect on underserved patients. A s physicians seek innovative practice models, one that is gaining ground is for practices to contract directly with patients to pay directly for some or all services-often called cash-only, retainer, boutique, concierge, or direct primary care or specialty care practices.
Such descriptions do not reflect the variability found in practices. For the purposes of this paper, the American College of Physicians (the College, or ACP) defines a direct patient contracting practice (DPCP) as any practice that directly contracts with patients to pay out-of-pocket for some or all of the services provided by the practice, in lieu of or in addition to traditional insurance arrangements, and/or charges an administrative fee to patients, sometimes called a retainer or concierge fee, often in return for a promise of more personalized and accessible care. This definition encompasses the practice types previously described.
This ACP position paper assesses the effect of DPCPs on access, cost, quality and other considerations; discusses ethical principles that should apply to all practice types; and proposes policies to mitigate any adverse effect on underserved patients.

METHODS
The ACP Medical Practice and Quality Committee, which is charged with addressing issues affecting the practice environment, developed these recommendations. The committee did an extensive literature review on DPCPs. Draft recommendations were reviewed by the College's Board of Regents; Board of Governors; and Councils of Early Career Physicians, Resident/Fellow Members, Student Members, and Subspecialty Societies. The ACP Center for Ethics and Professionalism and its Ethics, Professionalism and Human Rights Committee reviewed drafts of the paper and provided substantial input on the ethical considerations that should apply. The position paper and recommendations were reviewed by the ACP Board of Regents and approved on 25 July 2015.
services-which patients pay to have access to the practice. In return, patients are promised increased personalized attention.
Payment in cash at the time of service: Some do not accept insurance and require that patients pay directly for all services at the time of care. These practices are typically called direct primary care practices, direct specialty care practices, or cash-only practices. Patients that seek care from a cash-only practice may choose to obtain a health savings account or a wraparound highdeductible insurance plan.
Smaller patient panel: Some accept substantially fewer patients than traditional practices, which requires practices to "downsize" their patient panels as they transition to a DPCP.
Not every DPCP will have all 3 of these features. For example, there are practices that charge retainer fees that do not downsize their patient panels. Some accept insurance and charge a retainer fee whereas others do not accept insurance, with or without a retainer fee, and many other variations, which makes it challenging to generalize their effect on patient care. A 2013 survey (1) found that approximately 6% of physicians were in concierge or cash-only practices (up from 4% in 2012); another (2) reported that 9.6% of "practice owners" are planning to convert to concierge practices in the next 1 to 3 years. Yet ACP's own 2014 membership survey found that only 1.3% selected "retainer-based practice, concierge" as the method that best describes their basic source of compensation (3).

THE AFFORDABLE CARE ACT AND DIRECT PRIMARY CARE
The Patient Protection and Affordable Care Act (ACA) authorizes direct primary care to be included in the insurance exchanges, as long as they are paired with a wraparound insurance policy covering everything outside of primary care (that is, direct primary care combined with a low-cost high-deductible plan). It is the only noninsurance offering to be authorized in the insurance exchanges (4

EFFECT ON ACCESS
Because DPCPs often limit their patient panels to several hundred patients compared with the typical 2500-plus panel size (6), there is concern that such downsizing, especially when associated with retainer fees, could create a barrier to lower-income persons, patients with chronic diseases, and other underserved populations.
One study (7) found that retainer physicians have smaller proportions of patients with diabetes than their nonretainer counterparts, and they care for fewer African-American and Hispanic patients. The study does not definitively address whether the case mix of retainer practices is causally driven by their retainer status or whether these practices tend to emerge in highincome areas where there are fewer African-American and Medicaid patients.
The literature (8,9) has some examples of DPCPs that charge comparatively low retainer fees, and/or do not accept health insurance, and have made their practices accessible to lower-income, uninsured, and other vulnerable populations, possibly at less cost to the patient than if the practice accepted traditional insurance.

EFFECT ON COST
The DPCP advocates assert that price transparency, competition, lower administrative costs, and patients accepting more personal responsibility for the cost of their care will keep out-of-pocket costs and overall spending down to a competitive and affordable level (10). Critics observe that DPCPs can leave the patient at risk for higher out-of-pocket costs for health care. For most families and individuals, health insurance enhances access to health services and offers financial protection against high expenses that are relatively unlikely to be incurred and those that are more modest but are still not affordable to some (11).

EFFECT ON WORKFORCE
Practices that downsize patient panels could potentially place an even greater strain on physicians that remain in the larger system. Yet DPCPs may lead more physicians to stay in practice, rather than close their practices altogether, which could have a beneficial effect on alleviating workforce shortages. The effect of DPCPs on projections of workforce shortages requires more study.

VALUE OF PERSONALIZED SERVICES
Retainer fees may cover personalized services or special amenities, such as extended patient visits, preventive services, immediate access, private waiting areas, and coordination with specialists. Practices vary in what services are included in the retainer fee. No research is available to indicate the cost benefit of such amenities.

ETHICS AND PROFESSIONALISM
The ethics policies of ACP, as embodied in the College's Ethics Manual, Sixth Edition (12), state that physicians have both individual and collective responsibilities to care for all. Such ethical considerations must guide physicians in considering the types of practices they choose to participate in and what they must do to ensure their practices provide accessible care to patients in a nondiscriminatory manner. Practices that, by design, exclude certain categories of patients should be understood to create a greater potential of being discriminatory against underserved populations and require special attention by physicians to the ethical considerations involved. Practices that downsize their patient panels also must be aware of ethical and legal considerations relating to patient abandonment.
The College also believes that physicians in all types of practices should strive to take care of patients enrolled in Medicaid. Medicaid is a crucial part of the safety net for poor persons and, under the ACA, is the principal source of coverage for patients with incomes up to 138% of the federal poverty level in states that have agreed to accept federal dollars to expand Medicaid.

ACP POSITION STATEMENTS AND RECOMMENDATIONS
The following statements represent the official policy positions and recommendations of the ACP. The rationale for each is provided in the full position paper (see Appendix, available at www.annals.org). g. the effect of physicians not participating in insurance and therefore not participating in national quality programs, interoperability with other electronic health record systems, and the associated effect on quality and outcomes.

SUMMARY
The growth of DPCPs seems to be principally motivated by physicians' frustration with paperwork, low reimbursement, and restrictions on time spent with patients. It is essential that policymakers address these and other factors. Yet it must also be recognized that DPCPs potentially exacerbate racial, ethnic, and socioeconomic disparities in health care and impose too high a cost burden on some lower-income patients.
The College supports physician and patient choice of practices that are accessible, viable, and ethical. It asserts that physicians in all types of practices must ensure that they are meeting their obligations to serve patients of all types, especially the poor and other vulnerable patients. ACP recommends that DPCPs consider ways to mitigate any adverse impact on the poor and other underserved populations.
Finally, the College calls for independent research to study the factors contributing to the growth of DPCPs and their impact on workforce, cost, and access to care especially for vulnerable populations.
From the American College of Physicians, Washington, DC.

Disclaimer:
The authors of this article are responsible for its contents, including any clinical or treatment recommendations. draft of the manuscript and Kolton Gustafson for his administrative support to the authors.
Financial Support: Financial support for the development of this guideline comes exclusively from the ACP operating budget.

Disclosures:
Mr. Doherty reports the following outside of the manuscript: board membership, consultancy, employment at ACP, expert testimony, grants (or grants pending), payment for lectures (Department of Medicine, Jefferson Hospital, Philadelphia; honoraria [pending]), payment for manuscript preparation, patents, royalties, payment for development of education presentations, stock or stock options, travel (Utilization Review Accreditation Commission [URAC]; no compensation received for board membership but travel expenses to attend URAC board meetings were reimbursed), other travel/accommodations/meeting expenses, and other. Mr. Doherty reports the following during the composition of the manuscript: grant(s), consulting fees or honorarium, support for travel to meetings for the paper, fees for participation in review activities, payment for writing or reviewing the manuscript, provision of writing assistance or administrative support, and other. Dr. Fuisz reports that she is a Managing Partner of a small private practice group; this is a direct payment practice that does not participate in Medicare or other insurance products. Dr. Hunter reports that her input to the manuscript reflects her own views and not those of the U.S. Office of Personnel Management where she is employed as the Chief Medical Officer. Authors not named here have disclosed no conflicts of interest. Disclosures can also be viewed at www.acponline.org/authors/icmje/ConflictOfInterestForms .do?msNum=M15-0366.

APPENDIX: ASSESSING THE PATIENT CARE IMPLICATIONS OF "CONCIERGE" AND OTHER DIRECT PATIENT CONTRACTING PRACTICES: A POLICY POSITION PAPER FROM THE AMERICAN COLLEGE OF PHYSICIANS
As physicians and other health professionals seek innovative and efficient ways to deliver high-quality health care to patients, many different practice models are emerging. One that is gaining ground is for practices to contract with patients to pay directly for some or all services-often called cash-only, retainer, concierge, private contracting, or direct primary care or direct specialty care practices.
Such descriptions do not reflect the variability found in such practices or reveal much about how they operate and their implications for patient care.
The American College of Physicians (the College, or ACP) is the largest physician medical specialty society and the second-largest physician membership organization in the United States. Further, ACP members include 143 000 internal medical physicians (internists), related subspecialists, and medical students. Internal medicine physicians are specialists who apply scientific knowledge and clinical expertise to the diagnosis; treatment; and compassionate care of adults across the spectrum, from health to complex illness.
For the purposes of this paper, the College defines a direct patient contracting practice (DPCP), as any practice that directly contracts with patients to pay outof-pocket for some or all of the services provided by the practice, in lieu of or in addition to traditional insurance arrangements, and/or charges an administrative fee to patients, sometimes called a retainer or concierge fee, often in return for a promise of more personalized and accessible care. This definition of DPCPs therefore encompasses retainer, concierge, boutique, cash-only, direct primary care, and direct specialty care practices.
The move to DPCPs is based on the premise that access and quality of care will be improved if patients have a greater responsibility to pay directly for services provided by physicians and other health professionals in the practice, without third party payers imposing themselves between the patient and the physician. The literature shows that there are potential benefits to DPCP models-including providing patients with better access and more time with physicians and fewer administrative burdens on the practice-yet concerns have been raised that they may be inequitable and cause access issues for patients, especially among those who cannot afford to pay directly for care. Overall, there is little in the way of high-quality, independent research on the effect of DPCP models on quality and access.
This paper offers ACP's assessment of the evidence on the policy and patient care implications of DPCPs to inform discussion among policymakers, researchers, the public, and physicians about the potential implications of DPCPs.

Method
The ACP Medical Practice and Quality Committee, which is charged with addressing issues affecting the practice environment, developed these recommendations. The committee did an extensive literature review on DPCPs. Draft recommendations were reviewed by the ACP Board of Regents; Board of Governors; and Councils of Early Career Physicians, Resident/Fellow Members, Student Members, and Subspecialty Societies. The College's Center for Ethics and Professionalism and its Ethics, Professionalism and Human Rights Committee reviewed drafts of the paper and provided substantial input on the ethical considerations that should apply. The position paper and recommendations were reviewed by the ACP Board of Regents and approved on 25 July 2015.

Types of DPCPs
Although there is considerable variation in DPCPs, they include 1 or more of the following features: Administrative service fees (retainer or concierge fees): Some DPCPs charge a monthly or annual administrative fee-in addition to or in lieu of some of their usual fees for billable services-which patients are expected to pay to have access to the practice. In return, patients typically are promised increased personalized attention and greater individualized resources (such as increased and more immediate access to the physician, in person or virtually).There is no established or accepted range of retainer fees, nor are there any good studies or data available on the usual retainer fees charged by such practices or on their effect on access for poorer patients. Retainer fees seem to be more common in primary care practices and less common in subspecialty and surgical cash-based practices. The DPCPs that charge a monthly or annual administrative fee are typically called concierge, retainer, VIP, or boutique practices. In this paper, we refer to them as "retainer fees" and "retainer practices" except when quoting directly from literature that uses other terms.
Payment in cash at time of service: Some DPCPs do not accept traditional insurance payment or participate in insurance contracts. In these practices, patients pay directly for all services at the time of care and those with insurance can choose to file an insurance claim on their own and be reimbursed directly by the insurer for covered services. There is no guarantee that the amount of reimbursement by the insurer will cover their out-of-pocket payments to the practice, and they may pay more out of pocket than if they received care from a physician who participates in an insurers' network and accepts their negotiated payment rate. Such practices may be combined with a health savings account and wraparound high-deductible insurance plan. Practices that require payment in cash at time of service are typically referred to in the literature as direct primary care practices, direct specialty care practices, or cash-only practices.
Smaller patient panel: Most DPCPs seem to accept substantially fewer patients than traditional practices, which results in "downsizing" of patient panels when traditional practices transition to a DPCP.
Yet not every DPCP has all 3 of these elements. Some DPCPs charge retainer fees yet accept patients' insurance coverage, including Medicare and Medicaid (although retainer fees cannot cover services already paid for by Medicare and Medicaid), and may participate in insurers' provider networks and their established payment schedules. Some cash-only DPCPs do not charge retainer fees. Some DPCPs charge retainer fees, and/or do not participate in health insurance, without downsizing their patient panels. In addition, there are some cash-based DPCPs that bill patients and collect payment after services are rendered rather than collecting the fee at the time of service.
One variation of DPCPs is sometimes called a "hybrid" practice. This practice accepts insurance and offers their patients a choice of a retainer-style practice or a more traditional nonretainer practice option. They may apply some of the additional practice income from the optional retainer fee, for those who can afford it, to keep their practice open to all types of patients. Reports indicate that often hybrid practices open up the option of a concierge or retainer program to a small portion of patients-typically less than 10%-who want more time and enhanced service from their physician (13).
Three basic questions can be asked of DPCPs to help distinguish among them: 1. Does the practice require patients to pay out-ofpocket for some or all services rendered? 2. Does the practice charge a retainer fee to some or all patients, and if so, how much is the fee and what services are included in it?
3. Does the practice have a smaller and more limited patient panel than is typical of traditional practices in the same specialty field and community?

Prevalence of DPCPs
Geographic and market prevalence of innovative practice design are hard to determine because there is not a single source or entity that tracks the number of DPCPs. According to the American Academy of Private Physicians, there were 2400 physicians practicing some form of retainer-based medicine nationwide in 2010 and 3500 in 2011. In 2013, the Concierge Medicine Research Collective (a research and data depository created by Concierge Medicine Today) (14) estimated this number to be more than 5000 -an increase of nearly 100% in just 3 years. (Concierge Medicine Today is a national trade publication for the direct primary care and concierge medicine marketplace; it is not a peer-reviewed publication.) It is likely that there are additional physicians and practices that are not included in these estimates. According to a 2013 Medscape survey (1), approximately 6% of physicians were in concierge or cash-only practices (up from 4% in 2012). This represents a relatively large increase, and experts in the field note that this number will continue to rise as physicians are faced with uncertainty in pay cuts and look for new ways to improve their practice satisfaction (15,16). Another recent survey from the Physicians Foundation (2) found that 9.6% of "practice owners" were planning to convert to concierge practices in the next 1 to 3 years. (The Physicians Foundation is a charitable organization that funds research into changes in practice systems, including a biannual survey of physicians that is sent to more than 650 000 physicians nationwide or approximately 80% of physicians in direct patient care).
The movement to innovative practice design is occurring across all medical specialties, with 6.8% of all physicians indicating that they will stop taking insurance in favor of concierge-style medicine or "direct primary care" (2). Although many medical specialties are in these types of models, most retainer-based physicians are in primary care. Concierge Medicine Today (17) estimates that nearly 70% of current U.S. concierge physicians operating practices today are internal medicine specialists (the second most popular medical specialty in concierge medicine is family practice).
Yet ACP's 2014 member survey suggests that far fewer ACP members report receiving substantial compensation from DPCPs. When asked "which method best describes your basic compensation for patient care, whether self-employment income or paid by another entity?" only 1.3% selected "retainer-based practice, concierge" (3). (The available selections were performance-based salary, fixed salary without incentive or bonus, productivity model, time-based salary, retainer-based practice or concierge, and some other method).
In 2010, retainer-based practices were located in all but 11 states and tended to be concentrated in urban areas. The highest numbers of retainer-based practices were found in the metropolitan areas of Los Angeles, Miami, and Washington, DC. When looking at the number of practices compared with population, Naples, Florida had the highest number of retainerbased physicians as a proportion of the population (18). In addition, according to Concierge Medicine Today (17), Florida, California, Pennsylvania, and Virginia also have a significant number of persons (most aged >50 years) seeking out concierge physicians.
In addition, direct primary care was included in the Patient Protection and Affordable Care Act (ACA). This clause requires that direct primary care be included in the insurance exchanges, with the caveat that these practices be paired with a wraparound insurance policy covering everything outside of primary care (that is, direct primary care combined with a low-cost highdeductible plan). It is the only noninsurance offering to be authorized in the insurance exchanges (4). Yet direct primary care so far is not widely available through the ACA's exchanges; before 2015 "there [were] no DPC [direct primary care] practices operating in the federally facilitated exchanges, but the first DPC offering paired with a QHP [became] available in the Washington state exchange in January 2015" (5).
Given that DPCPs are recognized by the ACA, and that there seems to be growing interest in them among physicians, what do we know about their effect on patient access and patients' experience with the care provided, total health care spending, out-of-pocket costs, physician supply shortages, and the cost and benefits associated with personalized services included in the retainer fee?

Effect on Patient Access and Patients' Experience With Care Provided
To allow for increased access to physicians and other amenities in the practices, physicians in DPCPs often limit their patient panels to several hundred patients at most, which is a significant drop from the typical 2500-plus panel size (6). A 2005 study from the U.S. Government Accountability Office found that retainer physicians reported having, on average, 491 patients. In comparison, physicians saw an average of 2716 patients during the year before they started their retainer-based practices. A more recent analysis conducted for the Medicare Payment Advisory Commission of retainer practices found the following: Among the retainer-only physicians that could estimate their panel size, patient panels ranged from 100 to 425 patients. The average size was about 250 patients, just over a tenth of the average panel size these respondents reported having before starting their retainerbased practice (2265 patients). The physicians using hybrid models had kept a much larger patient panel, but had many fewer retainer patients. Over half of our respondents wished their practice size was larger (18).
A study by Alexander and colleagues (7) found that retainer physicians have smaller proportions of patients with diabetes than their nonretainer counterparts, and they care for fewer African-American and Hispanic patients. Given that minorities are already underserved and at risk for worse health outcomes, these findings suggest that retainer practices could exacerbate racial and ethnic disparities in health care. In addition, this study found that retainer practices reported caring for fewer patients on Medicaid than nonretainer physicians. The authors note that the study does not definitively address whether the case mix of retainer practices is causally driven by their retainer status or whether these practices tend to emerge in high-income areas where there are fewer African-American and Medicaid patients (7).
Another concern about downsizing patient panels is the potential for patient abandonment. When transitioning to a DPCP model that involves having a smaller patient panel, physicians will need to be aware of local and state patient abandonment laws that impose requirements, such as adequate notice, when a physician terminates a relationship with a patient. Although patients may be offered the choice to continue or stop seeing physicians once they transition to a DPCP model, once the practice's panel size is set at the new (lower) maximum number of patients it is able and willing to see, the remaining patients may have no choice but to find another physician. These patients may feel abandoned. Some may not choose to follow their physician or may not be able to follow their physician if the practice adopts a retainer fee, so they will need to find a new physician.
By not accepting health insurance or by charging retainer fees, proponents of DPCPs suggest that they can offer better access and other aspects of care that benefit their patients, including lower-income patients: a. Patients may benefit from more time with their physicians and more convenient and timely access to services. b. They may allow practices to stay in business that otherwise might close because traditional insurance does not cover their costs of delivering care.
c. They also offer the advantage of simplifying the patient-physician relationship by removing the barriers created by third-party payers and potentially lowering administrative costs.
d. By charging a retainer fee, or by being able to bill their full charges to those patients who can afford it, they can subsidize no-cost or discounted care for patients who cannot afford the retainer (13). They may also be able to have arrangements to provide discounted or free care to low-income patients (19).
In the literature, there are examples of DPCPs that charge retainer fees, and/or do not accept health insurance, which reportedly have made their practices accessible to lower-income, uninsured, and other vulnerable populations at possibly lower costs than if the practices accepted traditional insurance. These practices typically charge very modest monthly retainer fees to make them affordable to more persons, which are much lower than many other retainer or concierge practices.
For example, Time profiled Qliance Health, a Seattle company that manages direct primary care practices: Here's how it works: For a flat fee every month, patients have unlimited access to their doctor-in person and by phone or email-for routine things like checkups, cuts and burns, infections, flu shots and skin exams, as well as chronic-condition maintenance like blood tests for diabetes or high cholesterol. Under the law, every American is required to have medical insurance-but direct-primary-care patients can seek less expensive policies, because they require coverage only for hospitalizations, surgeries and other specialized care. [ . . . ] Qliance now serves some 35,000 patients; the cost of about half of them is paid by the government through traditional and expanded Medicaid programs. Treating a wide variety of patients-young and old, healthy and chronically sick, well-off and poor-Qliance claims to be saving approximately 20% on the average cost of care compared with traditional fee-forservice providers" (8).
The ACP Internist profiled another company in Rhode Island, HealthAccess RI, which is "building a network of doctors working in what it describes as subscription-based primary care. The fee is $35 per month plus an $80 enrollment fee. After that, each doctor's visit costs $10. Per its website, the company is targeting several groups, including those without insur-ance, immigrants without documents, or persons with high-deductible insurance plans" (9).

Effect on Physician Workforce Shortages
Practices that downsize patient panels and promise a very limited number of patients per physician could potentially place an even greater strain on physicians that remain in the larger system. Their decreased panel size must be compensated for by other physicians who may already be overburdened, given current shortages in the primary care workforce. In contrast, retainer arrangements and other DPCP models may lead more physicians to stay in practice, rather than closing their practices altogether, which could help alleviate the shortage of primary care physicians. At present, there are little data or analyses on the effect of DPCPs with smaller patient panels on projections of workforce shortages.

Effect on Health Care Costs and Patient Out-of-Pocket Expenses
Advocates of DPCPs believe that price transparency, competition, lower administrative costs, and patients having "skin in the game" (accepting more personal responsibility for the cost of their care) will keep out-of-pocket costs and overall spending down to a competitive and affordable level (10). Even in cash-only DPCPs that do not accept insurance, it is argued that patients may have less overall out-of-pocket costs than what they typically would pay under traditional insurance models after deductibles, copayments, and restrictions on covered benefits are considered, and they get better access and quality (20). Yet there is very little independent evidence on their effect on overall health care spending and patient out-of-pocket costs.
One study by Klemes and colleagues (21), often cited by proponents of DPCPs, looked at the effect of the personalized health care model used by MD-Value in Prevention (MDVIP). (MDVIP is a collective direct primary care group with practices in 43 states and the District of Columbia). This study found these practices had an estimated cost savings to the health care system in the 5 states it studied-New York, Florida, Virginia, Arizona, and Nevada-of $119.4 million for 2010. This represented a $2551 savings per patient mostly because of reductions in hospital utilization, including a 79% reduction in hospital admissions for Medicare patients and a 72% reduction for those with commercial insurance between ages 35 to 64 years in MDVIPaffiliated practices. In addition, the study found that MDVIP-affiliated practices had 49% fewer hospitalizations for avoidable admissions (including heart failure, diabetes, and pneumonia) than traditional primary care practices (21). However, ACP's review of the study's method found that it had no adjustment for baseline health or socioeconomic factors that affect medical utilization. Therefore, there is reason to be highly skeptical of its findings. The potential adverse effect of DPCPs generates considerable concern because some charge substantially higher retainer fees than reported by Health-Access RI or Qliance Health, particularly on access for underserved populations. Retainer and cash-based practices, critics contend, can leave patients at risk for higher out-of-pocket costs for health care: a. In practices that do not accept insurance, patients are at a greater financial risk for unexpected health care costs. For most families and individuals, health insurance enhances access to health services and offers financial protection against high expenses that are relatively unlikely to be incurred and those that are more modest but are still not affordable to some (11).
b. In addition, the burden placed on patients to submit the forms and wait for the reimbursement from their insurance company will probably be something to which most patients are not accustomed to and may find overwhelming. This is in contrast with the goal of patient empowerment, which is common among many of these practice models.
c. Most insurers will only pay the maximum amount allowed to an in-network provider-and in many cases, deductibles and coinsurance are higher for out-ofnetwork care. As a result, patients may end up paying more out of pocket for primary care and specialty services than if the services had been provided by an innetwork practice.
d. Under the ACA, most preventive and screening services are covered by insurers at no out-of-pocket expense to patients. It is unclear how direct payment practices bill for such preventive services. If they require payment out of pocket at the time the service is rendered, as is the case for most other services, patients may forgo receiving the recommended preventive services or end up paying more out of pocket for them, even if patients subsequently submit a claim for reimbursement to their insurer.

Cost-Benefit Effect of the Personalized Services Included in Retainer Fees
The retainer fee charged by many DPCPs may cover personalized services or special amenities, such as extended patient visits, preventive services, immediate access, private waiting areas, and coordination with specialists. Practices vary in what the retainer fee includes and can range from several hundred to thousands of dollars per year. Although many of these services seem to be benefits, there is no research to indicate the cost benefit of many of these "extra amenities:" a. For example, research shows that reduced time spent with patients likely causes dissatisfaction among both the physician and the patient and could interfere with the quality of care (22). However, research does not show a direct correlation between visit or appointment length and outcomes of chronic diseases or quality of care. Many factors likely play a role in delivering high-quality care to patients, including time with a physician. In addition, the benefits of increased time spent in a physician-patient interaction likely plateau at some point. Although it is clear that short appointments, as a result physicians feeling rushed, do not contribute to high-quality care, research does not exist on what an optimal visit time should be.
b. Retainer practices note that they are able to see their patients more often throughout the year. Once again, there is no evidence to suggest that this is always necessary or effective. With all of the "amenities" offered by these practices, it is important to do a costbenefit analysis to understand the true effect of the "extras" in a practice. At this time, no research or data are available to indicate that many of these amenities in a practice yield better clinical outcomes. It is important to be aware of the potential for overutilization of physician time and medical services.
c. Supporters of these practices models maintain that physicians are able to offer different services than a traditional practice, such as ensuring coordinated and comprehensive care, allowing access to physicians at any time, permitting longer appointments with the physician, offering chronic disease management, and providing cost-effective convenience (23). However, there is a lack of good evidence on the effect of DPCP models on outcomes and cost. Because many DPCPs do not participate in quality measurement programs, it is difficult to systematically assess their effect on outcomes and costs compared with other practices.

The ACP supports physician and patient choice of practice and delivery models that are accessible, ethical, and viable and that strengthen the patient-physician relationship.
Innovations that result in accessible, ethical, and viable practice models that also strengthen the patientphysician relationship should be encouraged. In addition, because patients and physicians have different needs, we support physicians' and patients' choice of effective, efficient health care delivery options that best meet their needs.

2.
Physicians in all types of practices must honor their professional obligation to provide nondiscriminatory care, serve all classes of patients who are in need of medical care, and seek specific opportunities to observe their professional obligation to care for the poor: a. Physicians should consider the potential impact of changes in their practices that could have the effect of making it more difficult for poorer patients to access their practices and consider steps to mitigate any such impact.
b. Physicians should consider ways to reduce barriers in their own practices that may limit access for the growing numbers of patients enrolled in Medicaid; federal government and state governments should address barriers to physicians accepting more Medicaid patients.
American College of Physicians is strongly committed to advocating for increased access to quality health care for all, regardless of race, ethnicity, socioeconomic status, or other factors. The College believes that racial and ethnic disparities in health care are unacceptable and supports policies that increase access to health care for all (24). The College reinforces its position that all patients, regardless of race, ethnic origin, gender, primary language, socioeconomic status, sexual orientation, cultural background, age, disability, or religion, deserve high-quality health care. Research shows that racial and ethnic minorities tend to receive poorer quality care than nonminorities, even when access-related factors, such as insurance status and income, are controlled.
Further, longstanding ACP ethics policies state that physicians have both individual and collective responsibilities to care for all, which the College sees as fundamental to ethics, professionalism, and the privilege of self-regulation. Also key in evaluating and structuring any practice arrangement are the fundamental ethics principles of beneficence and distributive justice that are central to the profession's ethical foundation, which can create an essential tension between physician autonomy and physician duty to care that is at the core of this issue. Caring for all and doing one's fair share honors medicine's societal contract and the substantial societal investment in medical education.
The ACP Ethics Manual, Sixth Edition includes various important guidelines related to patient care and our complex and changing health care system that are relevant to different practice models: Concern about the impact of the changing practice environment on physicians and insured patients should not distract physicians or society from attending to the unmet needs of persons who lack insurance or access to care.
Physicians have an obligation to promote their patients' welfare in an increasingly complex health care system. This entails forthrightly helping patients to understand clinical recommendations and make informed choices among all appropriate care options. It includes management of the conflicts of interest and multiple commitments that arise in any practice environment, especially in an era of cost concerns. It also includes stewardship of finite health care resources so that as many health care needs as possible can be met, whether in the physician's office, in the hospital or longterm care facility, or at home.
The patient-physician relationship and the principles that govern it should be central to the delivery of care. These principles include beneficence, honesty, confidentiality, privacy, and advocacy when patient interests may be endangered by arbitrary, unjust, or inadequately individualized programs or procedures. Health care, however, does take place in a broader context beyond the patientphysician relationship. A patient's preferences or interests may conflict with the interests or values of the physician, an institution, a payer, other members of an insurance plan who have equal claim to the same health care resources, or society.
Physicians have a responsibility to practice effective and efficient health care and to use health care resources responsibly. Parsimonious care that utilizes the most efficient means to effectively diagnose a condition and treat a patient respects the need to use resources wisely and to help ensure that resources are equitably available. In making recommendations to patients, designing practice guidelines and formularies, and making decisions on medical benefits review boards, physicians' considered judgments should reflect the best available evidence in the biomedical literature, including data on the cost-effectiveness of different clinical approaches. When patients ask, they should be informed of the rationale that underlies the physician's recommendation (12).
Such ethical considerations must guide physicians in considering the types of practices they choose to participate in and what they must do to ensure their practices provide accessible and ethical care to patients in a nondiscriminatory manner. Practices that, by design, exclude certain categories of patients should be understood to create a greater potential of being discriminatory against underserved populations and requires special attention by physicians to the ethical considerations involved.
The College also believes that physicians in all types of practices should strive to take care of patients enrolled in Medicaid. Medicaid is a crucial part of the safety net for poor persons and, under the ACA, is the principal source of coverage for patients with incomes up to 138% of the federal poverty level in states that have agreed to accept federal dollars to expand Medicaid. Currently, more than 65 million persons are enrolled in Medicaid, the single largest federal coverage program, and enrollment is projected to continue to grow in all states. The College policy strongly supports Medicaid expansion, and ACP has assisted ACP chapters in making the case for Medicaid expansion to their state legislators and governors. It would be inconsistent for the College to urge Medicaid expansion while not also urging its members to increase their acceptance of Medicaid enrollees. Further, without an increase in physician acceptance of Medicaid patients, many of the newly enrolled will be challenged in finding access to care, especially primary care.
Federal and state governments have a responsibility, however, to reduce barriers to physicians accepting more Medicaid patients. Low Medicaid reimbursement that sometimes does not cover the cost of care has been shown to be associated with physicians accepting fewer Medicaid patients. The College has urged Congress to restore a program, created by the ACA, which paid primary care physicians and related subspecialists no less than the Medicare rates for designated primary care services and vaccinations in calendar years 2013 and 2014. This program expired on 1 January 2015. The College also is assisting its state chapters to encourage their states to fund the program, as an estimated 15 states plan to do, should Congress not restore it. In addition, Medicaid administrators should reduce regulatory requirements that serve as a barrier to physicians accepting more Medicaid patients.
The College acknowledges that physicians, especially those practicing in large salaried institutions, may not have direct control over the patient populations served by such institutions. Nevertheless, the College strong urges physicians in all practice settings to advocate for policies that ensure access to care for poor; minority; lesbian, gay, bisexual, and transgender; and other underserved populations and seek opportunities on their own to serve such populations.

Policymakers should recognize and address pressures on physicians and patients that are undermining traditional medical practices, contributing to physician burnout, and fueling physician interest in DPCPs.
The growth of DPCP models seems to be associated with the enormous pressures being placed on traditional practices and on the patient-physician relationship, including an unequal playing field in negotiating contracts with insurers, insurer consolidation and market dominance, excessive administrative requirements, payments that undervalue physicians' contributions to patient care and threaten practice sustainability, pressure to reduce time with patients, disparities in primary care reimbursement compared with other specialists, and poorly designed electronic health records and associated regulatory (meaningful use) requirements.
A 2013 Medscape physician survey (1) found that the least-satisfied specialties were plastic surgery (41% reported satisfaction in their practice), internal medicine (44%), and diabetes or endocrinology (45%). In addition, this survey found that physicians value their relationship with patients and the intellectual challenge and satisfaction of being able to solve problems and make patients healthier. Another report conducted by the RAND Corporation and the American Medical Association (25) found that physicians who perceived themselves or their practices as providing high-quality care reported better professional satisfaction. Physicians, especially those in primary care, were frustrated when demands for greater quantity of care limited the time they could spend with each patient, which detracted from the quality of care in some cases. The report also noted that physicians' income stability was an important contributor to overall professional satisfaction (25).
The College strongly believes that policymakers must address and ameliorate the pressures that are undermining the patient-physician relationship, contributing to professional dissatisfaction and physician burnout, and is therefore fueling the growing interest among physicians in DPCPs and other practice arrangements.

Physicians in all types of practice arrangements must be transparent with patients and offer details of financial obligations, services available at the practice, and the typical fees charged for services.
The College believes that transparent health care information is useful for a wide range of stakeholders, including patients and consumers, employers and purchasers, health plans, health care professionals, and policymakers. The Institute of Medicine defines health care transparency as making available to the public, in a reliable and understandable manner, information on the health care system's quality; efficiency; and consumer experience with care, which includes price and quality data, to influence the behavior of patients, providers, payers, and others to achieve better outcomes (quality and cost of care) (26). Research has found that transparency can help a patient and their families make informed health care choices. Increased health care transparency can allow for increased trust in the patient-physician relationship and strengthen our health care systems. Transparency can also improve quality, safety, and efficiency throughout the health care system because of competition and/or the availability of clinical benchmarks.
Medical practices should be transparent and present information to current and new patients in a simple and easy to understand manner: Patients should be able to know whether a practice is accepting new patients, what the typical charges are for services, whether they will be required to pay for services at the time they are rendered, and what insurance a practice accepts.
Practices that charge retainer fees should disclose the amount of the fee, the services included in, and services that are not included in the retainer and the payment policies for the excluded services.
Practices that do not accept insurance have an obligation to ensure that patients are fully aware of the arrangement they are entering. In addition, the facility should ensure that the signage, marketing materials, and practice Web sites all clearly indicate the practice's policies.
Physicians who are making changes to their practices that affect patient access, such as downsizing their patient panel or making changes in insurance and financial arrangements that affect patients' access and costs, have a particular responsibility to inform patients well in advance of implementing such changes. The College is concerned about the potential consequences of downsizing patient panels on access to underserved patients, especially minorities and lowerincome patients, as previously discussed. In addition, when transitioning to a DPCP model that involves having a smaller patient panel, physicians will need to be aware of local and state patient abandonment laws that impose requirements, such as adequate notice, when a physician terminates a relationship with a patient. 6 The College strongly recommends that physicians be aware of their local community and economic environment and design their practices to ensure appropriate access to care among a range of patients. Practices should not charge retainer fees that have the intent or result of diminishing access for lower-income and other vulnerable patients.

Physicians in practices that
In addition, the ACA requires that health plans cover preventive services at no out-of-pocket cost to the patient. As a result of the ACA's mandated benefits, millions of Americans may now access services, such as flu shots, cancer screenings, wellness visits, and tobacco use cessation care, at no cost. According to the U.S. Department of Health and Human Services (27), 71 million Americans received expanded evidence-based preventive service coverage without cost sharing in 2011 and 2012 as a result of the ACA. Physicians in a DPCP model should follow the ACA's policy and should not charge an additional fee for preventive services included in the ACA benefit.
The recommendations of the sixth position are intended to provide guidance to physicians who are participating in or considering a retainer model on the steps that should be taken to ensure that there is no discriminatory effect on patients; they do not imply a College position of support or nonsupport for retainer fees. The ACP Ethics Manual, Sixth Edition states the following: "Physicians cannot and should not be expected to advise patients on the particulars of individual insurance contracts and arrangements. Patients should, however, expect their physicians to honor the rules of the insurer unless doing so would endanger the patient's health" (12).
The College has long advocated for universal health insurance coverage based on evidence that persons without health insurance coverage are at greater risk for preventable deaths and injuries. Health insurance can be the gateway for patients to have access to physicians and other health care. Nevertheless, the College recognizes that having health insurance, by itself, does not ensure access if there are not enough physicians in the community to take care of patients seeking care. In addition, insurance plans that charge excessively high deductibles and other cost-sharing plans that have excessive administrative requirements, inadequate benefits, and inadequate payments may create a barrier to patients getting needed care and physicians accepting patients with such coverage.
In practices that do not accept insurance, patients may be at greater financial risk for unexpected health care costs. For most families and individuals, health insurance enhances access to health services and offers financial protection against high expenses that are relatively unlikely to be incurred and those that are more modest but are still not affordable to some.
The College acknowledges that there are DPCPs, including some that do not accept insurance, which have strived to structure their practices to provide access to care for low-income and other vulnerable patients, including Medicaid patients and the uninsured. The features and effect of such arrangements merit further analysis.
These recommendations for DPCPs that do not accept insurance are intended to provide guidance to physicians who are participating in or considering such models on the steps that should be taken to ensure that there is no discriminatory effect on patients; they do not imply a College position of support or nonsupport for practices not accepting health insurance.
8. Physicians should consider the patient-centered medical home (PCMH) as a practice model that has been shown to improve physician and patient satisfaction with care, outcomes, and accessibility; lower costs; and reduce health care disparities when supported by appropriate and adequate payment by payers.
The College strongly supports PCMHs as a way for physicians to organize their practices to provide accessible, high-quality, and patient-centered care to patients in all socioeconomic groups, which would be supported by care management fees, shared savings, and other support from payers. Studies suggest that PCMHs can result in increased physician, staff, and patient satisfaction and reduced health care disparities (28 -31). Because PCMHs are typically supported by risk-adjusted care management fees, opportunities for shared savings, performance-based payments, and other forms of practice support, they can offer an alternative way for physicians to accept insurance and sustain and improve their practices. Further, the PCMH practice model allows physicians to improve physician satisfaction, reduce burnout, and improve patients' experience with the care provided to create better outcomes and potentially lower costs. In some respects, the risk-adjusted care management fees provided to such practices can be viewed as accomplishing the same goal as retainer fees (providing another source of practice revenue to support patient-centered care) but with the critical distinction that because payers are providing the care management fee, they do not present an additional cost to patients that could serve as a barrier to care for lower-income persons.

The College calls for independent research on DPCPs that addresses the following:
a. the number of physicians currently in a DPCP, where DPCPs are located geographically, projections of growth in such DPCPs, and the number of patients receiving care from DPCPs; b. factors that undermine the patient-physician relationship, contribute to professional burnout, and make practices unsustainable and their effect on physicians choosing to provide care through DPCPs (with practice arrangements that charge retainer fees, minimize participation with insurers, and result in smaller patient panels; the policy changes to ameliorate such factors should be considered); c. the impact and structure of DPCP models that may affect their ability to provide access to underserved populations; d. the impact of DPCP models on the health care workforce (especially the supply of, and access to, primary care); e. patients' out-of-pocket costs and overall health system costs; f. patients' experience with the care provided, quality of care, and outcomes; and g. the effect of physicians not participating in insurance and therefore not participating in national quality programs, interoperability with other electronic health record systems, and the associated effect on the quality of patient care and patient health outcomes.
Currently, independent research on the effect of DPCP models of care is limited. Although a few studies have suggested that these types of care are associated with better outcomes, they come from results from in-dividual DPCPs and are not independently verifiable. Some physicians believe that retainer medicine may be able to play a role in increasing the supply of primary care physicians because of an increase in physicians' satisfaction with their practices, although each retainer physician sees fewer patients (18). Yet there is little quantitative evidence on the effect these practices are having on the overall physician workforce. In 2005, the U.S. Government Accountability Office's review of available information on beneficiaries' overall access to physician services (32) suggests that concierge care does not present a systemic access problem among Medicare beneficiaries at this time. Yet much has changed since 2005; more physicians have chosen to participate in DPCPs, and the available research has not kept pace with practice trends. Therefore, ACP's review of the evidence does not allow us to draw conclusions on the effect of DPCPs on access to care and the cost or quality of care for Medicare or other patient populations. It is essential that this research gap be closed to gain a better understanding of the effect of the movement toward DPCPs.

Summary
Physicians who are considering or have entered into a DPCP seem to be principally motivated by frustration with the paperwork, low reimbursement, and restrictions on time spent with patients that are associated with traditional practice arrangements that accept insurance. It is essential that policymakers address and ameliorate these and other factors, which may contribute to physicians leaving medicine; closing their practices; or entering into DPCPs that charge retainer fees, do not participate in insurance, and/or see fewer patients.
Although the growing physician interest in DPCPs may be an understandable reaction to such external factors, it must also be recognized that such models potentially exacerbate racial, ethnic and socioeconomic disparities in health care and impose too high a cost burden on some lower-income patients.
The College asserts that physicians in all types of practices must ensure that they are meeting their professional and ethical obligations to serve patients of all types, especially the poor and other vulnerable patients. Because of the potential for DPCPs to create barriers for lower-income and other vulnerable patients, this position paper of the College poses policy, practical, and ethical issues that should be reviewed by physicians who are considering a DPCP model; further, physicians should consider the steps they should take if they are already in a DPCP to ensure that lower-income and other vulnerable patients are not disadvantaged or discriminated against.
We continue to recommend that physicians consider PCMHs as an alternative that has been shown to decrease physician burnout and improve patient care when supported by care management payments and other support by payers.
Finally, we call for independent research to study the factors contributing to the growth of DPCPs and their effect on workforce; cost; and access to care, especially for vulnerable populations.