Keywords

Introduction

Primary care matters. It is associated with better health, better health-care quality, and lower costs [1,2,3]. It is also a field that has been under extreme stress due to the hard work of transformation of the delivery system and payment inadequacy. As in all of medicine, there is diversity of structure, process, and quality across the many sites of primary care. There is also great diversity and change underway in payment for primary care. This chapter seeks to define key attributes of effective primary care and to then discuss financial model s to support the delivery of such care. A model can be evaluated from the perspective of the payer or the provider, but ultimately the different and sometimes conflicting views will need to be reconciled by evidence of the success or failure of any model based upon societal goals and realistic metrics. Additional chapters in this book provide details on specific programs and models of effective care—both in terms of cost and quality. In this chapter an attempt is made to set the stage for these more detailed presentations by providing more general information on finances and the lexicon of payment “reform.” A principle that underlies any successful financial model is that it is sustainable and it supports and nurtures what matters and minimally stimulates what does not. Any model is also only as good as its implementation and no single approach or model will achieve all goals.

Defining Effective Primary Care

In 1996 the Institute of Medicine (IOM) defined primary care . “Primary care is the provision of integrated, accessible health care services by clinicians who are accountable for addressing a large majority of personal health care needs, developing a sustained partnership with patients, and practicing in the context of family and community.” [4].

The IOM also defined quality care as addressing these six aims [5]:

  • Safe: Avoiding harm to patients from the care that is intended to help them

  • Effective: Providing services based on scientific knowledge to all who could benefit and refraining from providing services to those not likely to benefit (avoiding underuse and misuse, respectively)

  • Patient-centered: Providing care that is respectful of and responsive to individual patient preferences, needs, and values and ensuring that patient values guide all clinical decisions

  • Timely: Reducing waits and sometimes harmful delays for both those who receive and those who give care

  • Efficient: Avoiding waste, including waste of equipment, supplies, ideas, and energy

  • Equitable: Providing care that does not vary in quality because of personal characteristics such as gender, ethnicity, geographic location, and socioeconomic status

A defining concept in primary care is the patient-centered medical home (PCMH). In 2007 a coalition of primary care physician specialties adopted Joint Principles of the PCMH [6]. In summary the Principles state:

  • The patient-centered medical home (PCMH) is an approach to providing comprehensive primary care for children, youth, and adults. The PCMH is a health-care setting that facilitates partnerships between individual patients, and their personal physicians, and when appropriate, the patient’s family.

  • Personal physician—each patient has an ongoing relationship with a personal physician trained to provide first contact, continuous, and comprehensive care.

  • Physician-directed medical practice—the personal physician leads a team of individuals at the practice level who collectively take responsibility for the ongoing care of patients.

  • Whole person orientation—the personal physician is responsible for providing for all the patient’s health-care needs or taking responsibility for appropriately arranging care with other qualified professionals. This includes care for all stages of life, acute care, chronic care, preventive services, and end-of-life care.

  • Care is coordinated and/or integrated across all elements of the complex health-care system (e.g., subspecialty care, hospitals, home health agencies, nursing homes) and the patient’s community (e.g., family, public, and private community-based services). Care is facilitated by registries, information technology, health information exchange, and other means to assure that patients get the indicated care when and where they need and want it in a culturally and linguistically appropriate manner.

Quality and Safety Are Hallmarks of the Medical Home

  • Evidence-based medicine and clinical decision-support tools guide decision-making.

  • Physicians in the practice accept accountability for continuous quality improvement through voluntary engagement in performance measurement and improvement.

  • Patients actively participate in decision-making and feedback is sought to ensure patients’ expectations are being met.

  • Information technology is utilized appropriately to support optimal patient care, performance measurement, patient education, and enhanced communication.

  • Practices go through a voluntary recognition process by an appropriate nongovernmental entity to demonstrate that they have the capabilities to provide patient-centered services consistent with the medical home model.

  • Patients and families participate in quality improvement activities at the practice level.

    Enhanced access to care is available through systems such as open scheduling, expanded hours, and new options for communication between patients, their personal physician, and practice staff.

It goes on to comment on payment principles necessary to sustain this model:

Payment appropriately recognizes the added value provided to patients who have a patient-centered medical home. The payment structure should be based on the following framework:

  • It should reflect the value of physician and non-physician staff patient-centered care management work that falls outside of the face-to-face visit.

  • It should pay for services associated with coordination of care both within a given practice and between consultants, ancillary providers, and community resources.

  • It should support adoption and use of health information technology for quality improvement.

  • It should support provision of enhanced communication access such as secure e-mail and telephone consultation.

  • It should recognize the value of physician work associated with remote monitoring of clinical data using technology.

  • It should allow for separate fee-for-service payments for face-to face visits. (Payments for care management services that fall outside of the face-to-face visit, as described above, should not result in a reduction in the payments for face-to-face visits.)

  • It should recognize case mix differences in the patient population being treated within the practice.

  • It should allow physicians to share in savings from reduced hospitalizations associated with physician-guided care management in the office setting.

  • It should allow for additional payments for achieving measurable and continuous quality improvements.

These are good principles and articulate important aspects of any financial model that will support effective primary care. However, there are many possible methods to implement these principles.

To individuals, effective primary care is care that meets their personal needs, regardless of how an expert may rate it. In today’s world of value-based care goals, effective primary care in the eyes of payers and the public is the care that produces the highest possible quality result at the lowest cost. Additionally, it is the care that will reign in growth of national health-care expenditures to a sustainable rate.

Why Primary Care Matters and Where the Money Goes

The IOM asks primary care clinicians to practice in the context of family and community. The context of community can take many forms, from seeking to improve the public health infrastructure, to advocating for better schools and housing and to being a good steward of resources. Inevitably, there must be a limit on health-care spending, although historically America has spent an ever-increasing percentage of its gross domestic product on health care. Any limitation on spending will create a zero sum game where increases in one area must be offset by savings in another. On a larger social scale, America does this today. Compared to other economically advanced countries, we spend substantially more on health care in a presumed effort to achieve health. In turn, we spend less on social programs. Other nations spend a similar amount in the aggregate, but proportionally more on social programs, and achieve better health statistics than the United States [7].

It is likely that increased and sustained investment in primary care will be dependent on the ability of primary care to blunt spending trends. There is reasonably strong evidence that primary care supply, especially in relationship to specialist supply, is associated with improved quality and reduced cost [8]. Whether or not increased numbers of primary care providers will blunt cost spending trends is less clear [9].

It seems obvious that primary care capacity is central to achieving access to care and preventive services. Care coordination is generally a primary care function, even if some other disciplines do prominently coordinate care across the medical neighborhood, at least for select populations (e.g., dialysis patients) or on a time-limited basis (e.g., cancer treatment). If patients are to be cared for in lower cost settings than hospitals, inpatient and outpatient, there must be a primary care infrastructure for patients to get that care. Without other change, just spending more on primary care is unlikely to achieve the triple aim of better health, better health care, and lower cost. But without a strong primary care system, the aim is merely an imaginary dream. To attain the goals of the aim, primary care needs to transform. Many activities to create advanced centers of primary care are in progress. Today, although it varies by region and population of interest (e.g., children compared to the vulnerable elderly), approximately 6–8% of health-care spending is on primary care visits. These visits comprise roughly half of all patient visits, though the ratio of primary care to specialist visits is shrinking [2].

The financial models of greatest interest to geriatricians involve Medicare, though the dual-eligible population is very relevant, and models that combine Medicare and Medicaid expenditures such as the Program of All-Inclusive Care for the Elderly (PACE) are of great importance. A substantial portion of health-care expenditures is out of pocket and on services not covered by Medicare. That said, Medicare comprises the major source of payments to physicians and other geriatrics professionals. If Medicare Advantage (Part C) spending is removed and assumed to be spent in the same manner as the remaining funds, Medicare spends approximately 40% of its funds on hospitals, 16% on physicians and professionals, 15% on Part D drugs, 5% on skilled nursing, 3% on home health, and the rest on a variety of services such as durable medical equipment, hospice, and Part B drugs [10]. These proportions reflect the $597 billion spent in 2014. For hospitals, the 40% is split into 31% inpatient and 9% outpatient. In 2014, approximately $150 billion was spent on long-term care (nursing care facilities and continuing care retirement communities) [11], and Medicaid funded half of the nursing facility expenditures. The reason to understand these figures and sources of payment is because financial models will be designed to facilitate shifts in fund distribution with an overall net savings or better return on investment in terms of functional or health status.

Insurance Basics

It is important to understand some general principles of insurance and to have a working knowledge of health-care finance terminology in order to consider models of care. It is becoming even more essential as most of the newer models of finance are pushing insurance risk onto providers. Therefore, to a degree, provider entities are becoming quasi-insurance companies. Of course, historically there have been insurance companies that were also providers, merging financing and delivery, such as the staff model prepaid health plan represented by Kaiser and other staff model health maintenance organizations.

Risk

Insurance is designed to pool risk across a group of people that contribute money (premiums), or have money contributed on their behalf, to pay for events that are generally predictable at the population level, but not at the individual level. The goal is to have enough money to pay for the cost of care of those persons who need care. Many individuals will have no costs and others will have very high costs. The law of averages keeps premiums affordable overall. By pooling resources, the risk is spread and no one person sustains financial calamity when a high-cost event occurs. This very basic principle is all too frequently forgotten. It is understandable that a fundamental concept can be overlooked when we use insurance to pay for predictable expenses such as preventive care. It also is the case that more and more high costs are not for random “events” such as leukemia or trauma; they are due to chronic high-cost conditions that will persist for years. Nonetheless, the implications of the basic principle of insurance are many. The first is that the risk pool needs to be large enough to blunt the effect of random events. Pooling the risk for 100 people together will not create financial protection from an expensive event, if one should occur. A one in a million very expensive event probably will not strike this small cohort, but if it does, the funds would never be enough, even if saved over many years of “dodging bullets.” So a model of financing that expects a provider to accept risk needs to also require that the provider has a sufficiently large population to do so.

The next key point relates to the profile of people in the risk pool. Something bad may happen in a group of young healthy prevention-minded people. Something bad will happen in a group of frail elderly. If the premiums need to support the whole pool, they need to reflect the risk profile of the pool. Of course, we all know that the monthly health-care costs for the average working population are lower than the monthly costs for the average aged population. But the impact of a small subset of high-cost patients/insureds can be more significant than one might presume. It is the case that approximately 50% of health-care expenditures are generated by payments for care of 5% of the population [12, 13]. Let’s label them “high-cost patients” and call this the 5/50 principle. Take two insurance companies, one has 5% high-risk patients, the other has 6% high-risk patients, a seemingly trivial difference. But, based on 5/50, 1% of the high-cost patients account for 10% of the plan cost, so the difference is not trivial. This is especially important to note because even excellent care management would be challenged to reduce overall costs of the entire population by 10%. To get back to the level of cost compared to the company that had 5% high-cost patients, medical management would need to reduce the cost on the extra 1% high-risk patients by 100% (an unlikely achievement) or on the 6% by 17% (one sixth). If that seems challenging, try finding the savings from a group of patients that already have no costs in a year or only costs related to appropriate preventive services. This is why there is a need for “risk adjustment” in premiums or risk assumption payment methodologies. Even with risk adjustment, there is usually still an incentive to “cherry pick” and engage in risk avoidance strategies, because the adjustment is partial or stated differently, undercompensates. A successful model will fairly risk adjust and not create incentives to avoid complex patients and the professionals who care for them.

Risk assumption can be modified. Even insurance companies often buy insurance, called “reinsurance ” for extreme cost outliers. Sometimes special “high-risk pools” are funded to encourage people to be insured or to encourage insurance companies to enter a market place. One way this happens in payer contracts with risk-assuming provider groups is that the payer truncates the maximum per-patient per-year cost at a threshold number like $150 thousand. The payer retains the risk for any additional costs beyond the cap. The lower the cap, the lower the risk assumed by the provider group, but because high-cost patients generate most of the health-care costs, low truncation levels like $25 thousand are not used as essentially no risk is shared at that point. In primary care, the best example of modification is narrowing or broadening the type of risk. If a group of primary care providers is at risk for the total costs of care, their risk far exceeds (by almost 20:1) their typical primary care revenue, but the risk could be limited to primary care costs. For illustration simplicity, assume primary care is exactly 5% of the total cost of care and it is paid to a primary care provider group. Therefore, a mere 5% total cost overage would eliminate 100% of the gross income (i.e., before the practice expenses are even paid) of the primary care group. On the other hand, if the group was only at risk for an amount not to exceed 10% of their income, while the risk would be significant at 20% of take-home pay assuming 50% of revenue overhead costs, it would be feasible. The primary care group would be at risk for only 0.5% of the total cost of care, yet this financial model may well drive primary care clinicians to take measures that are likely to reduce the total cost of care by more than 0.5% of the total cost, as they have a huge share of their income on the line. A financial model for effective primary care will motivate primary care providers to improve access, to seek creative cost management solutions, and to be good stewards of resources, but not subject them to unacceptable risk levels.

Risk Adjustment

The focus above was on accepting financial risk and insurance by pooling risk. Besides cutting off risk acceptance at a dollar cap, it may be possible to otherwise adjust risk. Payment or the pool of dollars budgeted for care may be a risk-adjusted sum. For example, assume a large integrated health system is paid by a health plan every month (per month) for every assigned Medicare patient (per member or patient) an amount of money. We call this the PMPM (per member per month) payment. Assume that the average PMPM based on average historical costs of the Medicare population in that region is $1200. The plan will keep $200 dollars for administrative costs and for reserves and gives the providers a $1000 PMPM budget. This hypothetical integrated system is renowned for caring for complex patients and serves a large segment of people that lack supports and never had access to preventive care when younger, and it also has a large group of nursing homes in its system. Its patients are more complex, without question to all observers. In fact, in the year prior, the average cost for its patients was $1400 PMPM. Clearly the system will fail if it were to receive $1000 PMPM. So there is a need for risk adjustment. But how much? One might wonder if the expected costs in the region for such a more complex group of patients would be $1500 PMPM (i.e., the system was efficient relatively) or $1300 PMP (i.e., the system was relatively inefficient).

Age, gender, institutional status (as compared to residing in the community), functional capacity, social supports, and diagnoses all may be analyzed for their correlation with Medicare-covered medical costs. What Medicare has in large databases and for every beneficiary are costs and diagnoses in addition to age, gender, Medicaid eligibility, and institutional status. It uses this data to create the Hierarchical Condition Categories (HCC) methodology to risk adjust total costs. Diagnoses treated and reported in claims are used to create a risk score for every patient. Not every diagnosis matters (as an HCC) as one can imagine many have little to no impact on costs. An example would be allergic rhinitis. Some conditions when together in the same patient have a combined adjustment that is bigger than the sum of the individual factors as there is an interaction, consistent with the health risks clinicians well understand in patients with multiple interacting conditions. An example would be chronic kidney disease and heart failure. This HCC risk adjustment is used for total cost of care. There may be risk adjustors for other events that are structured differently, such as the risk for readmission to a hospital.

It is important to understand that these adjustments are based on modeling on large databases and are valid only for the intended use. So, a score that determines that a patient is likely going to cost 150% of the average does not mean that the chances of going to the ER are 150% more than average, even if they probably are higher. It is well known that factors such as health literacy, social supports, educational level, and income affect the probability of a patient receiving certain services or attaining certain clinical end points. It is obvious that financial impediments to obtaining insulin might affect glycemic control as an example. Risk adjustment is not generally used for quality measures. While this may seem unfair to a clinician being “graded” or paid based upon a quality metric in a pay for performance program, the general philosophy is that it is not a good social goal to set lower standards for disadvantaged populations. A cursory familiarity with the controversies in testing in education (e.g., “no child left behind”) will serve to illustrate this conundrum.

Benefit Design

The design of a health plan can affect costs by direct financial impacts or by behavioral alteration or both. Assume a benefit design where the patient pays 10% of all bills and there is no ceiling. If nothing else changes, 10% is saved by the plan over what it would have had to pay at the 100% benefit level. But, it is likely that behavioral effects will also occur. Some patients will not get unnecessary or low-value services, especially if they are costly, such as an MRI for a few weeks of low-risk back pain. But they may still seek inexpensive antibiotics for a viral respiratory illness. Others may not accept or be able to afford important, effective high-cost care. Some patients may forego important and generally cost-effective preventive care. While most preventive care still has a positive net cost (i.e., there is a cost per life year saved, not a net savings), this could result in higher medical costs for some, such as needing chemotherapy and radiation versus a lumpectomy, due to breast cancer being diagnosed at a later stage. Medications may be skipped and some medications do have a net cost savings, such as ACE inhibitors for patients with diabetes [14]. Professionals will become more conscious of costs and may change ordering behaviors. It is not always predictable what will occur. Beliefs that are not supported by evidence, fear of uncertainty, wealth, education level, confidence, entitlement, and perception of life expectancy are all factors relevant to the behavior of an individual. These may outweigh the effects of evidence-based decision-making. Nonetheless, design is attempted so as to promote good care and good patient behavior while reducing costs. Low primary care co-pays, high ER co-pays, and no preventive care co-pays would be examples of such design.

There are some limitations of benefit design that warrant mentioning. The first point is to not confuse saving money on one patient with saving on a population. In geriatrics the best example would be a state Medicaid agency saving nursing home costs by paying for in-home services. Assume that adequate in-home support costs one-third of the nursing home costs and that if available would allow a patient to return to the community from a nursing home. It seems a no-brainer to pay for the services. That is a 3 to 1 return on investment and most people would rather stay in the community. Win win. But if allowing payment for these services means that three people already in the community are now eligible for the service and use it, even if the one nursing home discharge was accomplished, it would still be a loss financially because now four people are receiving the service. The second point is that it is generally the case that benefit design is not targeted, or practical to be targeted in a way that optimizes behaviors or settings of care. It may be desirable to promote prevention and primary care and to seek to reduce discretionary specialty care or avoidable emergency department use by loading significant patient costs on these services. But the fact remains that the highest cost people, those who most need and use insurance, even when care and behavior is optimal, will still need specialty care and need the emergency room and hospital. They may need drugs for which there is no generic or lower cost alternative. A benefit design that seems to promote more efficient behaviors can actually become just a mechanism to shift the costs of care to the chronically ill.

Providers also need to understand, or accept, that not everything should be or can be paid for by insurance, no matter how useful the service may be. The most prominent example in geriatrics is long-term nursing home care. It could be an entitlement (i.e., available to all regardless of wealth status) like Medicare, but it is not. Society (as represented by government) has determined that preserving intergenerational transfer of wealth, even if modest, and reducing caregiver duties are of lesser import than reducing the cost to taxpayers. Any clinician who practiced prior to the creation of Part D recalls when Medicare beneficiaries assumed the full cost of their drugs and more senior clinicians recall a time when no preventive services, including influenza vaccinations and screening mammography, were covered. (As an aside, some benefit design can seem so illogical that it promotes cheating. Prior to screening mammography being covered, almost all women somehow had something of concern on breast exam and required diagnostic mammography, a covered service.)

Benefits should be tailored to the population served. This is not just for sales and marketing purposes to niche groups, but for meeting needs and promoting better care and health behaviors. For example, a nominal drug co-pay designed to lower monthly insurance premiums probably is irrelevant for most people with employer-based coverage. But that may not be the case for a Medicaid recipient. The same co-pay on a highly cost-effective prescription drug may be a true cost barrier to care. Many of the costs for younger women and children are preventive and maternity-related services. The main costs in employer-based coverage may be driven by specialty drugs for single chronic conditions and a nonrecurring cost such as cancer treatment for a year. The main costs in Medicare are for those with multiple chronic conditions. Those dually eligible for Medicare and Medicaid under 65 living in the community often have high mental health-care costs and needs.

A good benefit design will promote cost-effective care, be transparent about cost sharing, and allow for some level of alternative benefit delivery, such as paying for social supports only for a subset of patients for whom net savings are very probable and be tailored to the needs of the population for whom it is provided. A good benefit design will promote and support the provider type that can deliver the highest value care. A good benefit design may need to have the capacity to merge payment programs, such as what occurs in the Programs of All-Inclusive Care for the Elderly (PACE).

Attribution

In order to care for or pay for the health-care cost of a population, there needs to be a mechanism to define the population. If an insurance company, it may be easy. It is everyone who has paid premiums. But if the population is a subset of all insureds, such as those assigned to a specific individual or group, it may be more complex. The process of matching patient to provider is called “attribution.” One method of attributing a patient to a provider is for a patient to pick or be assigned to a specific provider or group. The patient’s insurance card has a PCP (primary care provider) name right on it. Another method is to assign based on service utilization. The latter can be prospective, i.e., attribution is for the year ahead, based upon last year’s usage patterns, or retrospective, i.e., attribution for the year is based on services obtained during the year and is only known after the year is over. Many programs in Medicare use a primary care physician-based methodology for attribution to the PCP or to a larger entity in which the PCP is a member. Traditional Medicare almost always allows the Medicare beneficiary to go to whomever the beneficiary chooses. Attribution methodologies need to match the intended usage. For example, if a program is about improving primary care, the attribution would logically be to a primary care clinician. If the program is about saving costs and improving quality over a 90-day episode of care for joint replacement, attribution would be more logically to a hospital or orthopedist. Attribution can seem simple until one recognizes the number of physicians/clinicians a patient may see in a year.

  • An example may illustrate. A patient has a regular primary care physician who does only outpatient care and who has a nurse practitioner on the care team. The goal is to assign the patient to this physician (PCP), using claims (billing/payment records). In February the patient goes to the PCP with a set of papers about how in January she got screened for dangerous hidden conditions at a van that came to her church with a nurse practitioner. A review of the papers indicates an EKG, heel ultrasound, limited carotid ultrasound, and blood tests were done. A sheet labeled “annual wellness visit” guides the patient to get her flu shot and mammogram every year and colorectal cancer screening (if by colonoscopy) every 10 years until age 75. The PCP sees her for her COPD, hypertension, osteoarthritis, obesity, and chronic low back pain. Despite vaccination efforts, in September the patient is admitted with pneumonia and an exacerbation of COPD. The patient has a long stay and steroids induce hyperglycemia requiring a new therapy of insulin. A team of general internal medicine hospitalists and various specialists care for her. With limited supports at home, the patient goes to a skilled nursing facility (SNF) for 3 weeks and is seen regularly (five times) by a physician/NP team there. The SNF physician also has an outpatient primary care practice. Once back to baseline function and off insulin, she is discharged. At discharge, the PCP performs transitional care services, but just billed an office visit because the office was not notified of the discharge, despite checking in with the facility during the stay, until 3 days after it occurred, so the transitional care CPT code could not be reported. The PCP office sees the patient two more times that year, both by the NP. Now consider the possible attribution methodologies. The PCP attribution could be based upon services that are hallmarks of good primary care, the annual wellness visit, and the transitional care management visit. But then a van NP is the PCP, as the transitional care was billed as a regular established patient visit. It could go by the greatest number of evaluation and management (E/M) visits by a provider tax ID/NPI combination for a provider/group with a primary care specialty, but then the hospitalist group is the PCP. Maybe a better way is to define primary care providers as people in a primary care specialty who have 40% of their total Medicare payments based on primary care services, like office visits and exclude hospital visits as counting. The primary care provider with the most E/M becomes the assigned PCP. Getting closer, but it still is the nursing home doctor who is the PCP, who, by the way, does not want her attributed to him as she was expensive that year. In order to not keep long-term care-focused providers from being PCPs, the new and improved method of using E/M by a primary care specialty provider does count nursing facility CPT codes (so long-term care patients may be attributed), but not if the site of care (site of service code) is the skilled nursing facility as compared to the nursing facility. Finally, now the actual PCP is the attributed PCP unless it is the NP in the PCP office. And what specialty are nurse practitioners and can they be PCPs, you might now ask? To sum up, an effective system will correctly attribute patients the large majority of times. It will never be perfect. Attribution is relevant as it determines patient assignments for payment, quality measurement, risk assumption—all the types of newer “value-based” payments.

Key Terms

  • Accountable care organization (ACO): An entity that is accountable for cost and quality of care for a population. “Accountable”. generally means that there is financial risk at some level bourn by the ACO.

  • Allowance: The allowed fee or payment amount set by the payer for a participating or contracted provider. Amounts charged in excess of the allowance for a covered service cannot be charged to the patient.

  • Alternative payment model : A method of payment that entails upside and usually downside risk for the provider or provider organization. It could be for a population for comprehensive care or a set of services (such as joint replacement, breast cancer treatment, or management of diabetes over a year). The term has the potential to be used to describe such a wide variety of arrangements that there may not be a consistent definition at this point in time, except that it is an alternative to straight fee for service.

  • Beneficiary or member: The insured. ”. Providers refer to them as patients.

  • Capitation : A payment per head ”. or per member. It requires additional specificity to determine what service or services are being capitated or being paid on a per head basis.

  • Codes: When billing for care, it is necessary to use a procedure code and a diagnosis code. This is a terminology schema. CPT is Current Procedural Terminology and is the core of procedure (including office visits) coding. ”. Medicare and other payers also use Healthcare Common Procedure Coding System (HCPCS) Level II codes like the G codes. Diagnosis codes are ICD-10-CM, the International Classification of Diseases series.

  • Cost sharing: Cost sharing is when the insured pays a sum when consuming services. It is in addition to the insurance premium. It may be a deductible (an amount that must be paid prior to the insurance paying anything), a coinsurance (a flat percent of the allowed amount), or a co-pay (a fixed sum paid for a service). The Affordable Care Act disallowed cost sharing for many preventive services. In Medicare there is significant cost sharing in Parts A and B. However, it is typically unseen as almost all patients have supplemental insurance (Medigap) or Medicaid.

  • Covered services: This a service that is a benefit of the plan. Participating providers must accept allowances for a covered service as payment in full (after collecting any member cost share). Plan rules do not dictate payment for non-covered services which would most likely be an out-of-pocket expense for the patient. There are often rules about notifying patients that a service is non-covered prior to provision and charge.

  • Hierarchical Condition Categories: This is the Medicare risk adjustor for payments to ACOs and Medicare Advantage Plans. It is based upon age, gender, institutional status, and billed diagnosis codes. It has major economic impacts, potentially many thousands of dollars per patient, which can be much more significant than the effects of medical management.

  • IPA: Independent practice association. This is a mechanism to form an entity that is of greater mass, while retaining independent practices. The IPA agreement governs the degree of independence and mutual obligations between practices and between practices and the administrative structure formed. There are costs associated with operating the IPA and any centralized services. An IPA may form to create greater contracting influence or to pool risk or to simply share costs and services by creating economies of scale.

  • Medicare Advantage/Medicare Part C: Medicare contracts with insurance companies to provide benefits to Medicare beneficiaries. The plans are called Medicare Advantage Plans. They must provide actuarially equivalent benefits and usually provide enhanced services such as annual physicals. They do not have to pay contracted or participating providers in the same manner that Medicare does. They can have limited networks, typically have a higher level of utilization review and prior authorizations, and take other actions to control costs. They are subject to risk-adjusted payments from the government based on HCC scores. They are also subject to substantial quality payment adjustments in the “Five-Star Quality Rating Program.” Accordingly, they are very interested in providers meeting the quality metrics and advancing optimized risk adjustment.

  • Medicare Parts A, B, and D: Medicare has parts that are for different services and have different beneficiary enrollment and benefit rules. Part A includes inpatient hospital services and skilled nursing facility services, Part B is outpatient hospital and professional services, and Part D is for drugs. Parts A and B are sometimes called “Traditional Medicare.”

  • Network: The set of contracted/participating providers. Some payers control costs by limiting networks, e.g., by capitating a subset of the community of providers and requiring all patients to use that subset or by having lower cost sharing for preferred providers. Providers could be preferred because they share financial risk with the payer or because they are judged to be better by some measure, usually an efficiency measure.

  • Participating provider: A provider that agrees to a contract. In Medicare the “contract” is essentially the rules of Medicare and the fee schedule or allowances. Professionals may “participate” and receive direct payment from Medicare. They may be “nonparticipating” and charge and collect from beneficiaries, but the final charges to the beneficiary are still limited by law. The beneficiary pays the provider and then receives reimbursement from Medicare. A third provider option is to “opt out” and privately contract with the patients. In this case the provider is ineligible for any Medicare payment for any service to any beneficiary. It is not a patient-by-patient option; the provider leaves Medicare completely. Patients are not allowed to collect from Medicare, though services ordered are covered when performed by providers that have not opted out, meaning tests, consults, drug prescriptions, etc., are unaffected.

  • PMPM: Per member per month. A payment or cost allocation method that allows correction for population size and time.

  • Value-based payment: The concept of paying providers differentially based on performance.

Payment Basics

There are a variety of ways to pay for primary care [15, 16]. Anyone of them, in isolation or in combination, can be effective. or ineffective depending on how they are executed and depending on the vantage point of the evaluation. The goals, behavior, and ethics of the payer, recipients of care, and providers can be factors. Each method has its pros and cons and risks. In many cases the methods are somewhat interrelated. For example, a PCP may have a salaried position, but the salary is fundamentally based on productivity using a fee-for-service payment methodology. At this point in history, payment methods are being used to change behavior and transform delivery systems. In theory, they are also designed to recognize high performance, but the transformation goals appear more significant than paying more for “the best.” A first point to consider in evaluation is whether a payment model is relevant, available, or practical to the primary care practice. This could be because a provider elects to not accept any insurance, sets their own fees, and individually contracts with the patient. It could be because the payer has one method of payment that is not negotiable. It could be that the payer would like to have the providers manage populations and accept insurance risk, but the provider in question lacks sufficient numbers of patients or sophistication to be given the risk. It is useful to address some specific models and assess whether they promote effective primary care.

Productivity and RVUs

Because so much of payment is based on some measure of productivity, it helps to understand how that may be measured. One method could just be number of visits. Another could be number of patients assigned to a provider without regard to how often the patient has a visit. Another could be hours worked. Another could be billed (not necessarily paid) services. Usually, productivity in primary care is based on visits that are adjusted for level of complexity or work. If the prices the practice charges are also based on that method, billings and work are equivalent. Every visit is assigned a “billing” code, a CPT (Current Procedural Terminology) code. Practitioners are familiar with these codes, such as 99214, an evaluation and management service in the office or other outpatient setting for an established patient that is of moderate complexity requiring a detailed history or exam. The specialty societies and the American Medical Association convene an expert body to help place services in a spectrum of relative work, and they make recommendations to Medicare. The recommendations are in units of work. The precise definition of a unit does not matter as all services are arrayed relative to one another, so an arbitrary service of 1.0 can be set and all services are a factor more (e.g., 1.74) or less (e.g., 0.68). Accordingly, these are called work relative value units or RVUs. Medicare sometimes sets the RVUs without advice and ultimately has authority to reject or accept any advice it receives. The methodology is inevitably imperfect, even if useful, and the code structure and fees/RVUs assigned are a source of debate. Though the word “value” is in the RVU term, it is important to understand that the term does not reflect utility, social value, or scientific value; it just reflects relative work. It is also very hard to effectively compare very different services such as an office visit, reading an MRI, and performing a total hip arthroplasty. But it is less difficult to compare different levels of office visits for established patients.

Salary

A salary can be an effective. financial model. However, as noted, the money for the salary must come from somewhere. Therefore, there are usually productivity requirements, with or without other incentives. A salary can be particularly effective in promoting high-quality care when it pools revenue streams so as to create an ability to provide care for time-consuming complex patients or for a population subset that is lower in income. For example, a 20-person primary group has five locations. One location is in a part of the community with high poverty rates. The other locations serve a well-insured population. The per-visit and per-patient revenue is lower in the lower socioeconomic status practice, even though the work RVUs could be more. The socially conscious group sets the salary without regard to location and actual revenue. Another good example may be patient population/panel size in a staff model HMO that has working aged and senior populations. A general internist will have some complicated patients, old or young. A geriatrician will have more complex patients, on average, and the practice of the HMO may even be to transfer complex patients from the general internists to a specialized geriatric team. Accordingly, the geriatrician’s panel size may be significantly smaller, if it is agreed that the internists and geriatricians all work at the same level of intensity, effort, and productivity. Therefore, the salaries may be the same. Salaries may not create a churn mentality of seeing more and more patients or performing marginally needed services. Conversely, there may be no productivity incentive. Of course, salaries can be coupled with incentives for retention, quality, productivity, and participation in activities that benefit the organization and its patients (sometimes labeled “citizenship”).

Fee for Service (FFS)

This is the “eat what you kill” financial model, though the terminology seems a little misplaced in health care. Like any business, revenue and expenses are key to having profit or take-home pay. They require careful monitoring and consideration. This model can be very successful financially and drive toward patient satisfaction as it is important to have satisfied customers to maintain business volume. There are many factors that determine the likelihood or ease of success. These include payer mix. A practice that is all Medicaid will be in dire condition. These include types of services. A practice that does everything, inpatient, nursing home, office, and occasional home care, may struggle with efficiency. In a given provider’s hands, some otherwise RVU equivalent services may be easier or more difficult, efficient or less efficient, and profitable or less so. There also needs to be a definition of success. Is an income of $100, $200, $300, or $400 thousand in a year a success? Is success defined as independence and self-determination—being your own boss? Is success defined as financial security or time with friends or family? Are there key attributes of care that if unable to be provided will cause career dissatisfaction?

To make this more concrete, let’s run some scenarios. In some cases, the scenario is somewhat artificial, but the overall example is valid.

Sample office practice: Goal income is $200,000 after expenses including health insurance, before reserves/savings, taxes, and retirement funding. 100% collection. Fee is 100% of the Medicare Physician Fee Schedule. 50% of visits coded 99213 ($74), 50% coded 99214 ($109). No other services are reported. Office overhead is 55% of gross revenue at a gross revenue of approximately $450,000. Because a large share is fixed costs like rent, utilities, staff (independent of volume), and professional liability and health insurance, a low patient volume will have a limited effect on reducing expenses.

  • Gross income will need to be $200,000/0.45 or $444,444.

  • The average visit is $91.5, so this will take 444,444/91.5 or 4860 visits annually.

  • The physician takes 4 weeks off a year for vacation, sick, and CME and works nine 4-hour sessions a week. This is 432 sessions annually. The number of patients seen each session must be (4860/432) or 11.25. The physician will book accordingly, considering cancelation rates.

  • That is a lot of hard work, but it may be possible. Alter any assumption and it can have a significant effect. Let’s look at a few:

Case 1: Fee Allowance of the Payer

The payer pays better than Medicare and the average visit is $100. That is not the 10% bonus Medicare once paid, but almost. This means $8.50 × 4860 or $41,310 annually just came to the physician. Expenses did not increase at all. Of course, if the fee was $8.50 less than Medicare, the lost income would be the same amount.

Case 2: Volume Effects

The physician wants to work less hard. The choices are eight sessions a week or nine patients a session.

Nine patients a session is a loss of 2.25 × $91.5 × 432 or nearly $89,000 in lost revenue. Maybe expenses decreased minimally (fewer disposable gowns and less exam table paper). Going to eight sessions is 48 fewer sessions annually, and this is a loss of 48 × $91.5 × 11.25 or $49,400 in revenue, and again expenses really do not fall. This illustrates how volume-dependent fee for service is. At some point, reductions in patients seen may mean fewer staff and increases may require more staff, but this mostly occurs at extremes or when there are multiple providers in a practice all making the same changes.

Instead the physician decides to make life easier by hiring another medical assistant (MA) and delegating some tasks appropriately. The physician works nine sessions and sees 11.25 patients per session, the MA costs $35,000 including all benefits, and the physician is much happier. It feels like seeing nine patients a session.

Case 3: Complete and Accurate Coding

The physician (physician A) compares billing patterns with another colleague (physician B) and notes that the colleague does not split the services 50/50 between 99213 and 99214; most are the higher level service. Both see their patients once a year for an annual visit to make sure all the bases are covered. Physician A usually codes the annual with 99214 because the patients usually have a few problems, but are not high complexity. Physician B usually does the same for the same reason, but also reports a variety of preventive medicine codes, since the services are being performed.

Service

Dr. A

Dr. B

99214 established patient office visit

$109

$109

G0439 annual wellness visit

 

$118

G0442 annual alcohol misuse screening

 

$18

G0444 annual depression screening

 

$18

G0446 annual face-to-face behavioral therapy for cardiovascular disease

 

$26

99497 advance care planning

 

$83

99406 counseling to prevent tobacco use (3–10 min)

 

$15

Total

$109

$387

Physician B is very conscientious and knows all the requirements of each service. If a service is not medically necessary, it is not performed. For example, if the patient is not a smoker, 99406 is not reported, or if the physician asks if the patient is willing to discuss quitting and gets a quick negative response and stops there, the tobacco cessation counseling service is not reported. Physician A completes his annual in 30 min. Physician B takes 45 min and has a team-based system that involves a questionnaire and standardized instruments asking about depression, instrumental and other activities of daily living function, hearing, substance use, falls, physical activity, nutrition, use of preventive services, and advance care plans. His staff have been trained to do follow-up services such as a PHQ-9 if the PHQ-2 is positive and provide resources on a variety of items, like explaining the state-approved advance directive forms. They also complete a checklist for the doctor and patient on services already documented or needed such as colorectal cancer screening and immunizations. The physician reviews each item and provides assessments, advice, and arranges for needed assistance with the help of his staff. Aspirin and cardiac risk factor reduction is discussed. Of course, the usual medical treatment issues are addressed as well. The patient’s values and understanding of prognosis with potential medical events are reviewed, and the doctor checks to determine if the patient has had such discussions with a surrogate decision-maker. One other thing physician B does is to use a medical record with alerts that help to correctly code the complexity of care by ICD-10 and also by prompts that a high-risk diagnosis was not included in the assessment and provisional claim even though it is on the problem list. Physician B said that the practice used to code like Physician A, but they had a compliance program that determined they were under-coding many of the 99213 visits and not even billing the preventive screening and counseling services they performed. The practice felt strongly about advanced care planning and created a system to make sure it was addressed annually. They got organized and not only is the care better, it is more efficient and the practice income is much improved.

Case 4: The Care Manager

The practice always thought it would be great to have a nurse that could do triage and education and follow some of the patients by phone to monitor them and provide self-management support. One of the private payers would pay an office for a nurse that did diabetes education, and then Medicare began to pay for transitional care management and chronic care management. The practice read an economic model article [17] that suggested after staff and opportunity costs, each Medicare member enrolled in chronic care management would actually net the practice over $300, so long as 131 Medicare patients per nurse were enrolled. The practice assessed the number of patients with chronic conditions that would potentially benefit. Some of the doctors thought the patient cost sharing would be an issue or that it was wrong to charge for something that was done all along. Others in the practice pointed out that it was not being done, because it was not paid for and none of the doctors had unlimited time to do this, not to mention that nurses might do a better job in many cases. The patients did not pay the cost share out of pocket as that amount was usually picked up by “Medigap” insurance or Medicaid. If they did have to pay a small amount, why was that wrong? The practice has to be financially viable, and the practice has no issue charging for all the other services it provides, they argued. So they set up a specific program to identify the patients, comprehensively assess them (if not already done), create care plans, and get consents. Then, if the patient needed such services, everything was ready to go. When Medicare began paying for complex chronic care management and paid additional sums to do the assessment, if extra work was required, it was an even better decision, in retrospect.

What these cases show is that optimizing revenue or RVU generation requires a systematic approach of assessing costs and revenue. The examples focused on office-based practice, but the principles apply to a home visit or nursing facility practice. Success requires knowledge of coverage, coding, and billing rules. It requires putting a system and workflows in place that makes the activity work. It may require an electronic record or other forms to be created. These forms or templates may be available from Medicare or a specialty society as resources for a practice. The last few years have seen significant enhancements in describing and paying for services for those with chronic illness, including behavioral health problems. Payment has facilitated and is intended to drive providers toward creating teams and integrating behavioral health into primary care. However, many feel that these adjustments are the wrong model of paying for primary care because they are still within a volume-driven payment system that inadequately addresses payment disparity based upon complexity being undervalued. In a treadmill day at the office, a complex patient is a disruption, whereas a minor acute illness is a pleasant break. Effective primary care should focus on the complex patients and make room in the daily schedule for them, while handling minor illness through a portal or phone call. Fee for service also motivates providers to generate income through such practices as the “annual EKG” and having an in-office lab for revenue rather than clinical care reasons.

Capitation for Primary Care Services

Capitation for primary care services can be an effective model. Capitation is a payment per person (or head) for a defined period of time. A payer would want to be sure that a practice was not just creating access barriers or turfing every issue to the emergency department or a specialist, so it might be coupled with monitoring through quality measurement or service usage, i.e., encounter data may still need to be provided for tracking and to address cost-sharing issues. This type of capitation does not put the provider at risk for anything but their own time and overhead. While that is not inconsequential, it is to be distinguished from the total cost-of-care risk that an accountable care organization (ACO) may take on. It is critical to define what is primary care, however. Is it just office visits? Are vaccines included, which are a costly supply? Are any office labs, tests, and procedures bundled into the capitation? The capitation payment needs to have some level of risk adjustment, even if just age. A payment for a specific patient could be based upon historical costs for that patient. Payment by head requires attribution and assignment rules and these may affect adequacy of payment. In Medicare, fewer patients have no services than a population of young adult males, but the issue is still relevant. The payment for a population using an attribution method that is not based on service history, but is based upon selecting a PCP on enrollment, includes payment for persons who do not use services, whereas payment based upon patients attributed by service use history with a PCP does not include nonusers. If there is substantial cost sharing for primary care, the capitation can be complex. A fixed co-pay can be factored in with limited difficulty, but a deductible creates difficulty as it would vary based upon the timing of a primary care service in relation to other services. Capitation inherently allows payment for non-face-to-face care alternatives to the face-to-face visit. It may promote team-based services, such as care review and education by a team member such as a pharmacist, who otherwise could not be a billing provider. For this reason, the Comprehensive Primary Care Plus Medicare Innovation Center program has a track where payments are split capitation and fee for service [18, 19]. Primary care capitation is a form of population management, but for a specific set of services. It may promote a structure and skill set that promotes successful total cost of care population management. Capitation creates an incentive to manage more patients and grow panel size. As fewer physicians enter primary care, this may be an important societal goal. This may help drive the development of team-based care where every team member operates at the top of their license or skill set and efficiencies are created. A patient visit volume practice will need to adjust to this payment method, and it may be difficult to operate under capitation with one payer and visit volume system with another.

Capitation could also be for a subset of services or “infrastructure.” Several PCMH programs have a PMPM for care management. This is a form of capitation that helps fund the PCMH and team-based care. Because this payment is often used to advance primary care, it is reasonably viewed as an effective financial model.

Fee for Service Linked to Quality and Value

Medicare is now using this payment methodology with the onset of MACRA (Medicare Access and CHIP Reauthorization Act of 2015). This type of payment is FFS, but the fees are adjusted based upon performance. It could be a multiplier being applied to fees (less than or greater than 1.0). It typically would only be applied to certain fees, for example, payment for an administered chemotherapy drug that costs $1000 a dose would neither be cut nor enhanced. Different payers may use different criteria, but usually quality and cost are factors. The payment could be presented as a bonus, but if fees are held down so as to fund the bonus, then it really is just an opportunity to regain ground and possibly surpass what would have been otherwise paid. If it truly is new money to primary care, then it can help promote the discipline. It remains to be seen whether this is a good model for effective primary care. On the one hand, it promotes quality measurement and improvement, processes that were rare in most practices until the advent of meaningful use and PCMH programs. It rewards those who invest and succeed in improving primary care. On the other hand, the record of performance programs improving outcomes and health is uncertain [20]. Driving focus on specific measures may not drive overall health improvements and measures of member experience, while important may not correlate with professional assessment of technical quality [21]. It is also the case that performance on outcomes type measures is affected by patient socio-demographics [22] and that conclusions based on small numbers are suspect. For example, efficiency on total cost of care, even with risk adjustment, or efficiency based upon inpatient admission rates would be unreliable at the individual provider or small group level. Some primary care providers care for atypical populations and are more significantly affected either by lack of appropriate measures or by factors that skew results. There is also a significant cost for the practice and payer to collect, report, and assess data and then modify fees or process bonuses. This approach seems to address a desire to pay for performance, but it may also be so inherently random or unfair that what it does most of all is to stimulate practices to move on to alternative payment models.

Alternative Payment Models Built on Fee for Service

These types of payments rely upon a fee-for-service event to trigger the payment model or population definition. It is more than an adjustment to the fees based on quality or efficiency. It usually requires downside risk, i.e., the potential for the provider to lose money and therefore be stimulated to manage the patient efficiently. This does not necessarily mean the provider is at risk for the total cost of care. An example of a more limited risk APM would be that the provider receives infrastructure support, such as a capitation payment for a nurse care manager. The provider also gets an advance on a quality and efficiency bonus. If the quality is inadequate or the costs do not suggest effective management, the funds for the nurse and prepaid bonus must be returned. Practices can do well in such an arrangement. In fact, primary care geriatrics may clamor for such arrangements as the Independence at Home and CPC+ demonstration. The model can drive and support effective primary care. The problem with any such risk or quality-based payment system is that it could also serve to drive a threatened provider segment into extinction. If the quality metrics and thresholds are not set properly, failure can be all but assured. Likewise for the financial targets. For this reason, it is important that any model have testing and protections for unexpected events. A basic example may serve to illustrate. Our goal is to design a payment system that promotes efficient high-quality care.

  • Approach A: A risk-adjusted budget for the population of concern is set, based upon national spending patterns for Medicare beneficiaries. There are corrections applied to account for payment amount differentials regionally, but not for regional utilization differences. In other words, if a DRG is paid more in New York City than in a rural Idaho hospital, that payment differential is accounted for, but no adjustment is made for the possibility that more New York City beneficiaries receive that DRG per thousand persons than is the case for the Idahoans. This system rewards efficient regions. It may even promote the redistribution of health-care providers to serve less urban areas which often use less services because profits are higher in Idaho than New York. There is still an impetus for the Idaho providers to be more efficient, though ironically if more providers were to move into the region, costs would predictably rise because provider supply has an enormous effect on service usage variation. This method is pay for performance with a lesser potential to be paid for improvement. A higher cost region would be at very significant risk for losses if it does not dramatically improve. If the threshold to break even after investment costs is too high, this essentially guarantees losses and is a funds redistribution program more than a program to drive performance. Urban safety net hospitals could be wiped out.

  • Approach B: This is the same as approach A, except the benchmark or goal is based upon historical costs for the region. In this case a very efficient region that has a year with higher costs is penalized, whereas the region that improves from ridiculously overly costly to just overly costly is rewarded. This method does reward improvement. It does account for factors in regional spending variation that may be beyond the control of any provider entity and take a generation to change.

As in all financial model s, there are pros and cons and incentives and disincentives, so execution matters greatly. But, if the model is well designed, an APM can be very positive for primary care. It can stimulate better primary care and promote recognition (and thus proper payment) for primary care. If the existence of a home visit service and digital medicine will prevent usage of an emergency room and exacerbation of chronic disease due to inattention, then they will be built or sustained. These services are not being promoted in a FFS system. The geriatric hip fracture comanagement system with a coordinated post-acute care service will prevent readmissions and drive down the complications that are costly for the hospital. It is no longer a pilot research project, but becomes an essential care element. The palliative medicine team that helps assure patient-centered care and in doing so reduces readmissions is now more than a nice-to-have service. But, if the model is poorly designed, the pioneers in transformation will be burned and future innovation stifled.

An interesting feature of any payment system that puts providers at risk is that the providers tend to adopt many techniques that payers have long used and providers had railed against. However, it may not just be a different perspective that justifies this change of heart; the application may be considerably more deft.

Any payment system that has financial gainsharing must have quality metrics and other oversight programs. No payment system is without fraud or abuse. While most providers are of high integrity, not all issues are about gross misconduct. Hospitals need their beds filled to break even financially as they have huge fixed costs. That does not mean that the community-spirited members of a not-for-profit hospital board want their neighbors to be sick, but they do want margin for mission. So too will every system have some degree of perverse incentives. Accordingly, it is essential to have measures of quality of care, access, and patient experience. Public trust, common sense, and lessons learned demand this.

Alternative Payment Models of Population Management

This is the attainment of providers being accountable to manage a population. There is no tie to a fee-for-service event. There is no disease-specific payment and silo by condition. Provider organizations must work together across specialties, professions, and institutions to achieve the triple aim. The promotion of or risk to effective primary care is essentially the same as in the APM based on FFS. Where the money goes matters even more in this model. A primary care capitation will go to primary care. A population payment could go to a provider organization that marginalizes primary care. While such an attitude may ultimately be an impediment to the long-term success of that organization, the short-term effects may be a complete loss of control by primary care. Protections could be applied, such as a minimum amount of payments being for primary care services or a requirement that organizational leadership must include a set percentage of primary care providers. It would be logical for Medicare programs to require organizational leadership by professionals with geriatric care competencies. Again, if well executed, this model can promote effective primary care and reward primary care providers.

Concierge Medicine

While there is great discussion, policy, law, and regulation on increasingly complex financial model s, some clinicians have turned back the clock to a time before insurance for primary care. These providers contract with their patients through a variety of mechanisms. They may be able to reduce overhead associated with billing or quality measurement that seems to miss the mark. Medicare allows some forms of concierge practice without opting out. This is typically by not having a retainer per se, but by having an annual fee for non-covered services. While a concierge provider may be more practical for an upper income patient, the model can allow for charitable care and should not be dismissed as socially unconscious. A concern about this model, beyond the major concern of equity, relates to keeping the customer satisfied. If the patients who elect a concierge practice have a sense of entitlement or the provider anticipates this situation, there is a risk that medically unnecessary services will proliferate. Every test will be done, and entry to the practices of the best specialist for whatever ails you will be facilitated, needed or not. This is a potential outcome that serves neither the patient nor society. On the other hand, this payment method can promote personalized service and time with the patient and caregiver for discussions that really matter. It was a mechanism to pay for chronic care management before such a benefit and payment were created in Medicare. It remains a mechanism to pay for case management, the additional nonmedical supports many patients or families need.

Practice/Provider Organization Size and Structure

There is no inherently ideal practice size or structure that promotes high-quality, efficient, and effective primary care [23]. A micropractice may have intense focus on access and patient empowerment and be highly conscious of cost [24]. They may deploy resources external to the practice, such as a community health team, and outperform groups and organizations. They may measure what really matters [25] and risk stratify their patients for effective management better than others. However, most providers benefit from a greater support structure and need scale to develop that structure. The present value-based payment system is also potentially averse to smaller practices. Medicare estimated that smaller practices would receive more penalties under MACRA [26]. This may relate to infrastructure needed to succeed in the measurement paradigm, e.g., smaller practices are more likely to also not use electronic records or to be able to measure quality, and not be due to actual performance on quality of care or efficiency. Where organizational size is legitimately critical is risk assumption and investment capacity. No single provider can assume total cost of care risk. It takes sufficient numbers of patients to make the basic principles of insurance work. Payment models that entail risk typically pay the reward (if achieved) more than a year after the investment period begins. Capitalizing an ACO requires millions of dollars of investment [27] in infrastructure and care managers. There must be an ability to withstand a loss of more than investment, but have an actual payback capacity. Therefore, the organization must have strong financial reserves. Organizational size does not require a common employer. It can be achieved with an independent practice association (IPA) or other mechanisms. Success in any payment model and success in delivering effective primary care do depend on shared mission, goals, and consistent processes. Success in newer models of payment requires more than is typically provided within the four walls of the primary care office. It requires planning, effective execution, maintenance, and support systems. It requires analytics and more sophisticated financial management structures. It requires an ability to invest in new services that may not immediately generate revenue. The ideal is to achieve the intense patient focus of the micropractice with the organizational support (clinical and administrative) and financial resilience of a larger organization.

Preparing for the Future

Predicting the future, especially in a chaotic political environment, is risky or hubris, but planning is important. It may make sense to seek to be prepared to meet the needs of our society and to keep fundamentals in mind. Health care is too costly. Quality is too variable. The solution does not lie in tweaking insurance company or Medicare rules, but in delivery system change. Providers know their patients and know the system best. Therefore, providers are the ones who can transform health care and address the needs of the population. They will be limited by a national focus on health care rather than health, but there is great opportunity within their locus of control. With this opportunity will come accountability. This means financial risk acceptance, quality measurement, and regulation. Primary care has been and will continue to be the cornerstone of any successful system of care. Therefore, any successful model will need to support effective primary care.

Summary

A small portion of total health-care expenditures goes to primary care, but primary care is an essential element to safe, timely, effective, equitable, efficient, and patient-centered care. In communities and countries where primary care is a higher proportion of resources and expenditures, care is of higher quality and lower cost. There are many ways to finance primary care and many business models for primary care providers. Some can operate simultaneously and be complementary; others may be in conflict and create confusion, such as simultaneous “value”- and “volume”-based payments. All models will depend upon the local environment (infrastructure, readiness, competing market forces) and the quality of execution. Productive and hard workers will always be needed. Providers and payers need to exist in the current world, while they prepare for and stimulate change to a more effective system of total care and improved primary care. Flexibility, creativity, self-assessment, big picture understanding, and preparation for change are required.