Figure. Average characters per ambulatory progress note in U.S. and international health systems.

Column height represents number of organizations. Dark columns represent 13 organizations outside the United States (140 000 notes from Canada, the United Kingdom, Australia, the Netherlands, Denmark, the United Arab Emirates, and Singapore). Light columns represent 254 organizations in the United States (10 million notes). Downing e. al. Ann Intern Med 2018;169(1).

Problems with U.S. payment methodologies include administrative complexity, over-valuing procedures relative to cognitive tasks and services, gaming of documentation for payment, loss of physician control over their time due to payers requiring excessive unreimbursed documentation time and administrative tasks, inequities in patient access and quality of care, and financial incentives that often interfere with a physician’s intrinsic motivation and best judgment about patient care. None of these problems are inherent to fee-for-service, but are a result of the particular fee-for-service methodology used by Medicare and adopted by other payers in the U.S.

A 2016 “time and motion” study of 4 specialties (family medicine, internal medicine, cardiology, and orthopedics) found physicians were spending twice as much time on computer and administrative tasks as paying attention to their patients. Some computer time overlaps with billable patient visit time, but excessive uncompensated administrative tasks effectively cut fee-for-service hourly pay in half. This has become a major cause of demoralization for the physician workforce, especially in primary care specialties.

A single-payer system would be ill served by simply adopting current payment methodologies. This paper will focus on payment of independent physicians with fee-for-service, capitation and other variants of “value-based payment,” and combinations of these. We will propose an alternative fee-for-service payment system that would minimize interference with patient care and would be far less costly and cumbersome to administer. 

Fee-for-Service Payment as implemented by Medicare using RBRVS:

Since 1992, the Medicare fee schedule has been determined by the Resource Based Relative Value Scale (RBRVS), updated regularly by the resource-based relative value scale update committee (RUC), convened by the American Medical Association with representation from multiple specialties in a proprietary process. The RBRVS attempts to rationalize payment by calculating a fee for each procedure, with components based on “Relative Value Units” (RVUs) for work, practice expense, and malpractice liability, plus a geographic adjustment for cost of living. All of these components except the malpractice RVU have turned out to be problematic.

The methodology of the RBRVS has led to marked over-payment for procedures compared to cognitive services, and the RUC has never effectively corrected this imbalance or accurately updated practices expenses and the geographic practice cost index. Prior to 2015, the RUC was constrained by Congress’ budget neutrality rules and the Sustainable Growth Rate (SGR) formula, so that increasing primary care pay would have to be taken from specialist pay, while all fees were threatened with drastic reductions by the SGR formula. 

Since passage of the Affordable Care Act, correction of fee-for-service has been sidelined by efforts to move away from fee-for-service toward value-based payment, and by the replacement of the SGR with the MACRA law and MIPS in 2015. This new law imposes pay-for-performance and pushes doctors to join Alternative Payment Models such as Accountable Care Organizations that can accept insurance risk, along with substantial increases in data reporting required for “value-based” payment. The result has been steady increases in physician time required for documentation, data reporting, and administrative demands. 

Work RVU:

Criteria for determining the Work RVU include physician time, skill, judgment, stress, and amortization of physician education cost, but many of these are subjective and vulnerable to inter-specialty lobbying on the RUC, which is dominated by specialists. As a result the Work RVU is skewed toward procedures at the expense of cognitive services that predominate in primary care specialties. Primary care physicians are paid far less than many procedural specialists, and often not enough to cover medical school debt, practice startup costs, and leave enough for a reasonable middle-class lifestyle. 

The time element in calculating the Work RVU has been based on surveys of a very limited sample of physicians and has not been updated regularly, raising serious questions about its validity.

The Work RVU for cognitive services uses Evaluation/Management coding that has relied on a complex “pay-for-documentation” scheme that rewards over-documentation and has only paid for patient visits, leaving time for documentation and care coordination largely uncompensated. CMS is now planning to move away from reliance on details of documentation for E/M services and allow more use of time-based coding in 2021, which should help reduce documentation “bloat.”

Practice Expense RVU:

The Practice Expense RVU has also relied on limited physician surveys, an unreliable source of data, and has not been updated nearly often enough to reflect actual practice expenses.

The Geographic Practice Cost Index (GPCI) uses a very complex mathematical formula and also relies on limited physician surveys, and has never been updated accurately. For example, cost-of-living in Hawaii is among the highest in the country, yet the Hawaii Medicare GPCI is mid-range.

The RBRVS is essentially a system for pricing procedures, instead of a system for paying physicians for their time and expertise, as is the case for all other highly trained professions and also for physicians who are salaried employees of hospitals and other care delivery institutions. 

Value-based Payment:

Current health care reforms in the U.S., collectively referred to as “value-based” payment, have been founded on the theory that the high cost of U.S. health care is caused by overutilization due to financial incentives inherent in fee-for-service payment. This ignores the fact that U.S. physician visits and hospital days per capita are at the low end of the range of utilization in other advanced countries that cover everyone, often use fee-for-service payment, and spend around half of what the U.S. spends per capita on health care. 

Although there are pockets of over-utilization driven by fee-for service, most notably in the hospital sector and with some procedural specialists, research finds much more under-utilization than over-utilization in the U.S., due to lack of access and financial and socio-economic barriers to care, often leading to costly complications and over-utilization of emergency rooms and preventable hospitalizations. There has never been any evidence of over-utilization of primary care services when paid with fee-for-service, yet “value-based” reform efforts have been focused primarily on primary care. 

The “value-based” solution to presumed over-utilization is to shift insurance risk onto doctors and hospitals in the form of pay-for-performance, capitation, and bundled payments, so that they make more money by providing fewer services, and hopefully only services that have “value.” Risk adjustment is required when insurance risk is shifted either onto capitated private plans within Medicare and Medicaid or onto doctors and hospitals under “value-based” payment reforms. The data demands of pay-for-quality and risk adjustment in turn require a marked increase in detail and complexity of documentation and diagnosis and procedure coding, adding ever more uncompensated physician time spent on documentation and administrative tasks.

Unfortunately, capitation and risk-shifting reward skimping on care and avoidance of sicker, more complex, and poorer patients and populations who are likely to need costly care. Proponents of “value-based” payment offer pay-for-quality or outcomes and risk adjustment to counter these perverse incentives, but both quality and risk are far too complex to measure accurately, and the crude measures in place divert physician time and effort to meeting metrics that often do not coincide with patient priorities. 

Crude risk adjustment also fails to deter avoidance of difficult patients and gaming of documentation. Capitated doctors soon find the keys to financial success are to maintain a large practice of relatively easy patients who don’t require too much of the doctor’s time, code their patients as intensively as possible to beat risk adjustment formulas, and try to avoid taking on more challenging patients who are likely to need a lot of the doctor’s time or require expensive care.

Administrative simplification is crucial to achieving a more cost-effective healthcare system, but risk shifting adds administrative complexity and cost. Due to their high administrative cost and exacerbation of disparities in care, we do not find “value-based” payment, pay-for-performance, shifting insurance risk onto doctors, or competition among risk-bearing entities such as Accountable Care Organizations and Medicare Advantage plans to be equitable or cost-effective strategies. We recommend eliminating all such provisions from payment for health care services under a single-payer system. Quality improvement should be founded on the intrinsic motivation of health professionals to improve patient care, not financial incentives and disincentives.

What happens when primary care physician pay is inadequate for the training, expertise, and practice costs required?

Third party payers generally control physician fees unilaterally. 

With fee-for-service based on the RBRVS, when fee increases fail to keep up with education debt, rising cost of living, and rising practice expenses, including uncompensated administrative burdens imposed by payers, then the only way primary care physicians can maintain their incomes is by increasing volume of patients seen. Since their time is limited, this means reducing time spent with each patient and limiting the number of problems addressed in a visit. Both primary care physicians and patients complain about this. Primary care physicians we know refer to this as the “fee-for-service treadmill,” and it accounts for the willingness of many of them to consider capitation, in which payment is not directly tied to the number of visits.

With capitation, physicians get a set amount of money per attributed patient per month. If the capitation rate is inadequate and does not keep up with rising practice costs, administrative burdens, and cost-of living, the only way physicians can maintain their incomes is to increase their panel size, thereby limiting how much time they can spend with each patient. The Hawaii experience with primary care capitation found substantial reduction in primary care visits (which saves no money under capitation), small increases in referrals to urgent care or the ER and to specialists, and no cost savings overall. Physicians may have more freedom to prioritize, spending more time with patients who really need it and less time with stable patients, but overall the effect is to push primary care physicians to take on more patients than they can handle and still provide high quality care. Now they are stuck in the “capitated payment treadmill.” And now the incentive to avoid caring for sicker, poorer, and more complex patients is amplified. 

In both cases, primary care physicians are stripped of professional autonomy and control over their time, contributing to widespread demoralization and “burnout.”

A Proposal for Fee-for-Service Based on Time:

For independent physicians, we recommend that a U.S. single-payer system should return to a standardized fee-for-service payment system but replacing the RBRVS with a fee-for-service system designed to be as administratively simple, objective, and incentive-neutral as possible. 

This proposal would improve on the RBRVS and RUC fee-determination system in the following ways:

1. It would shift determination of relative values to the value of physician training and expertise instead of attempting to assign relative values to procedures themselves. It would use time as a common denominator linking the value of physician expertise to all services, cognitive and procedural. The actual value of a procedure depends heavily on the expertise of the physician performing the procedure and on each patient’s individual circumstances in any case.

2. It would improve the objectivity and accuracy of the practice expense component of fees and the geographic cost-of-living adjustment.

3. It would disconnect payment from the details of documentation and end over-documentation and gaming induced by pay-for-documentation.

4. It would compensate physicians for all professional time spent on patient care, including documentation, care coordination, alternative care delivery modes such as telephonic and telehealth based care, and administrative tasks, and would restore physician control over their time and schedules.

5. It would establish clear separation between the practice expense and work components of fees, and would require the single-payer to reimburse practice expenses at cost. This would create an incentive for the single-payer to minimize unnecessary administrative burdens and costs for physicians, and it would isolate the work component as nearly the entire determinant of physician personal income.

6. It would replace the RUC with collective bargaining at the national level to determine the work component of fees, which would determine physician personal income.  Collective bargaining is intended to keep fees reasonable for physicians and balance the interests of physicians with those of the single-payer and taxpayers. Nationally negotiated fees would be subject to regional cost-of-living adjustments.

7. In order to correct the procedural-cognitive pay gap, it would reduce negotiation of fees to only hourly rates based on years of post-graduate training required for a specialty or sub-specialty and set the same hourly rate for all specialties with the same required years of training, instead of negotiating the fee for each procedure.

A “Fee-for-Time” payment system was first proposed by Tom Wachtel and Michael Stein in a 1993 JAMA article, entitled, “Fee-for-Time System: A Conceptual Framework for an Incentive-Neutral Method of Physician Payment.” Our proposal is an updated version of the same concept.

Instead of a proprietary committee such as the RUC setting the fee scale using “relative value units,” we propose retaining fee-for-service but replacing the RUC and Congress’ budget neutrality rules with collective bargaining to keep fees reasonable for both Congress and physicians, and vastly simplifying payment procedures to reduce administrative costs. 

For practice expenses, we propose a more objective procedure by dividing the country into regions based on cost-of-living, each with a regional single-payer board that includes community representation. Instead of using surveys, these boards would determine the practice expense component of physician fees from regularly updated audits of full-time practice financial accounts by specialty or subspecialty within a region. Practice expense would be calculated by subtracting physician personal income from total practice costs. Malpractice Liability cost would be included in practice expenses based on these audits. The practice expense component of fees would then be standardized based on the average full-time practice cost for each specialty or sub-specialty within a cost-of-living region. Monthly specialty-specific average practice cost would be divided by 180 work-hours in a month to determine an hourly practice expense, and the single-payer would pay for this at cost, with no further conversion factor. Audits of practice expenses should be updated annually in the first 3 years and at least every 3 years thereafter to capture savings from reduced practice overhead as practices adjust to the simplicity of billing and collections with time-based coding and fees in a single-payer system. Physician personal income would then depend almost entirely on the work component of fees.

We propose that the value of the work component be based on years of training required for each specialty or sub-specialty, with hourly rates linked to payment for specific procedures based solely on the time associated with each service or procedure. Specialties with the same number of years of required training would be paid at the same hourly rate, so as to correct the discrepancy between payment for procedures compared to cognitive services. The work component must include everything the physician does on behalf of their patients that requires their time and expertise, including documentation and care coordination time, telehealth, etc. To avoid encouraging inefficiency or over-documentation, paid time for documentation should be standardized and in proportion to the time of the visit. 

For reasons of simplicity and to avoid disputes between patients and their doctors about how many minutes were actually spent versus billed, we recommend a prospective method (incremental, standardized times chosen prior to each visit) for assigning time billed for most services and procedures, but with the option of switching to a retrospective method (counting actual minutes spent) in case of complications and non-typical circumstances requiring more time from the physician than usual. 

Cost-of-living adjustments to the work component would be applied at the regional level based on publicly available cost-of-living statistics instead of using the GPCI formula. The practice expense component of fees would not need any adjustment for cost-of-living because it would be based on audits of actual expenses that include cost-of-living.

The Fee-for-Time proposal requires assignment of a time for all physician services, both cognitive and procedural. The fee for a given service would be determined by applying the hourly rate for a given specialty or sub-specialty to the time associated with each service.

For cognitive services, the doctor, or doctor and patient together, would choose from an array of incremental times, such as in 15-minute increments. The rules governing payment would specify that the time increment billed includes reasonable documentation time, and if necessary, care coordination time, writing a letter for the patient, etc. Documentation requirements would follow the general principles for documentation of Evaluation/Management services from the CMS Evaluation/Management Services Guide, but the criteria for levels of service based on the details of documentation would be replaced with time-based codes. Since payment would depend on time only and not the details of documentation, the focus of documentation should return to patient care priorities and quality improvement instead of “pay-for-documentation.” For most office visits for cognitive services, a simple “SOAP” note (Subjective: Objective: Assessment: Plan:) should suffice. For unusually complicated patient situations, the doctor could document the nature of the complication and switch to the retrospective method of counting minutes actually spent. 

For procedural services, uncomplicated procedures would be paid prospectively for a standardized time usually scheduled for each procedure. For most commonly used procedures, the times would be determined objectively based on audits of physician practice schedule books and electronic record time stamps, and the time associated with a given procedure would be the median time scheduled for the physician performing the procedure. It may not be possible for auditors to determine usual times for rarely used procedures, and in such cases determination of time would revert to the retrospective method of adding up time actually spent. The times for each uncomplicated procedure would be adjusted to include reasonable standardized times for documentation. As with cognitive services, in case of time-consuming complications the physician could document the complication and switch to retrospectively adding up actual time spent. 

Setting hourly rates for each procedure: 

For the work component of fees, we propose establishing a base hourly rate for a general practitioner with one year of post-MD training and a standardized incremental increase for each additional year of training required to practice a given specialty or sub-specialty. It may be reasonable to establish a higher incremental increase for the first 3 years beyond internship and a lower value for subsequent years. However, all specialties with the same number of years of required post-graduate training would be paid at the same hourly rate, with the exception of adjustments to correct maldistribution of specialties, which would be negotiated separately based on needs assessment supplied by the single-payer regional Boards.

A sub-specialist may treat some patients who only require the physician expertise and training of a lower level of specialization. For example, a cardiologist, who had to complete internal medicine training before sub-specializing in cardiology, may have a practice that is half cardiology and half general internal medicine. We suggest paying this physician at the higher cardiologist hourly rate for all their patients with a diagnosis relevant to cardiology. CMS, in consultation with their regional Carrier Advisory Committees, already establishes medical necessity criteria that include relevant diagnoses required for each procedure. For this cardiologist’s patients without a cardiology relevant diagnosis, we recommend paying at the internal medicine hourly rate.

Collective negotiation of hourly rates:

Standardized payment for the work component for doctors and other healthcare professionals requires a mechanism to keep payment in proportion to the training and expertise required for each profession, best accomplished with collective negotiation between the single-payer and organizations representing each profession. We propose requiring all physicians to join a professional organization, tied to licensure, that would be authorized to negotiate with the single-payer to establish the base hourly rate and the incremental increase for each additional years of training required for specialty or sub-specialty practice, with re-negotiation of these rates every 1-2 years, and at least annually for the first 3 years. Negotiation of fees would occur at the national level, and each professional organization would include all specialties and sub-specialties with the same professional degree. The regional boards would be responsible for determining community needs, and the work component could be adjusted by negotiated agreement to correct maldistribution of physicians and specialties. 

Promoting professional ethics and quality:

Mandatory membership in a professional organization would enable this same organization to be charged with peer review and enforcement of professional ethics, deterring fraud and abuse, continuing medical education, and facilitating communication and referrals between physicians and other healthcare professionals. Peer review and discipline of unethical physicians would be much more effective if failure of remediation meant expulsion from the organization and loss of licensure. An effective peer review system should mean much reduced need for government agencies charged with investigation and punishment of fraud and abuse by physicians. All these functions would be covered by professional dues at no cost to the single-payer system beyond the practice expense component of hourly fees. 

The professional organizations, along with independent publicly funded quality improvement organizations, could contribute to quality improvement projects, responding to demonstrated problems with quality or consistency of care and using metrics developed and modifiable by project participants, similar to the Intermountain model of quality improvement. This would require nominal block grant funding from the single payer. Quality improvement should be founded on the intrinsic motivation of professionals to improve patient care, not financial incentives and penalties.

Implications of single-payer reform using the Fee-for-Time proposal:

We believe that time-based payment of physicians within a single-payer system would be simple, objectively determined, and would require minimal administrative cost for both the physician and the single payer. 

Billing and collections would become straightforward, because there would be a single payer with standardized fees and no patient cost-sharing. At the end of each day, the physician or office staff would enter the codes for procedures and services performed, the single payer would add up the time associated with all of them, and payment would be based on a standardized hourly rate. Physician offices should not need billers, coders, and scribes, or contracts with billing services, collection agencies, and revenue cycle management companies, almost completely eliminating any need for the extensive industry that has grown up around the complexities of billing and collections in our current multi-payer system. Only office staff directly assisting in patient care should be needed. Reduced office staffing and overhead means the practice expense component of physician fees could be reduced accordingly without affecting physician personal income.

Physician control of their time: The Fee-for-Time proposal specifies that the sole link between the standardized hourly rate for a given specialty and payment would be the time associated with each service or procedure. Since time-based coding would include documentation and care coordination time, there would be no uncompensated services or uncompensated physician time spent on behalf of patients. For cognitive services, the physician or physician and patient together would choose the time needed for each visit. For both cognitive and procedural services, the physician would have the option of retrospectively billing for time spent in case of time-consuming complications or crisis situations. A physician could schedule more time for a difficult patient knowing they would be paid the same hourly rate with no financial penalty for scheduling the time needed. Physicians, not the single payer administration, would have control of their schedules and time.

No pay-for-documentation: With the value of a physician’s time determined by the training required for their specialty, payment for all procedures and services would depend on time only, not the details of documentation, freeing documentation from “pay-for-documentation” that has been used in Evaluation/Management coding. The focus of documentation would return to patient care and quality improvement priorities. This should eliminate the pervasive problems now seen with over-documentation, cloned notes, and documenting things not actually done, resulting in improved accuracy and relevancy of documentation and markedly reducing information technology needs for billing purposes. Based on comparison with other countries that do not use pay-for-documentation, disconnecting payment from the details of documentation could reduce length of medical notes by 50-75% while keeping them focused on information truly relevant to patient care.

Incentive-neutrality and intrinsic motivation:  Payment based on time would provide no financial advantage to perform procedures compared to cognitive services. Pay-for-performance schemes have been shown to be harmful to intrinsic motivation in health care, whereas the incentive-neutrality of time-based payment would leave the best interest of the patient as the primary motivation for choice of service. 

Reduced fraud and abuse: Time is objective and inherently much less susceptible to gaming than pay-for-documentation, risk adjustment formulas, or determining prices of procedures based on assessment of physician skill, judgment, and stress. Other countries such as Canada, where the physician payment system is perceived as fair and consistent, report few problems with billing fraud and abuse by physicians.

Correction of disparities. Under a universal single-payer system using time-based fee-for-service, physicians would be compensated fairly and consistently for time spent on behalf of their patients, with no financial advantage or disadvantage to influence provision of services to any particular category of patients:  complicated or straightforward, rich or poor, succinct or rambling in communication style. Uncompensated and under-compensated care would be eliminated. A physician could schedule more time for complex or difficult patients without financial penalty. Reduced practice overhead cost and equalized payment for all patients, with no uncompensated care, would markedly reduce barriers to establishing medical practices in underserved communities. Any maldistribution of physicians and specialties not corrected by these changes in how physicians are compensated could be addressed with negotiated adjustments to hourly rates to incentivize practice in underserved communities, based on needs assessment recommendations from the regional single-payer boards.

Simplified diagnosis coding. With elimination of “value-based payment” and no shifting of insurance risk onto competing health plans, hospitals, and doctors, there would be no need for risk adjustment, so maximum detail in diagnosis coding would not be required. Diagnosis coding for medical necessity and quality improvement should require only 3-4 digits of diagnosis codes, as is the case in other countries that use the ICD coding system.

Audits vs surveys. Regional boards within a single-payer system would commission audits instead of surveys to accurately determine the practice expense component of fees and the times to be associated with common uncomplicated procedures. Audits of times scheduled for cognitive (E/M) services would not be needed because the times would be chosen by the doctor. Regional audits would be objective and not subject to biased surveys and the inter-specialty rivalries that now influence the RBRVS Update Committee (RUC). The up-front cost of audits would be greater than surveys, but this cost would be offset by elimination of the audits of E/M code documentation now being done by CMS, and by eliminating the cost of paying the AMA to run the RUC. Once the standard times for uncomplicated procedures have been determined, the simplicity and objectivity of time-based billing and collections and reduced opportunities for gaming and fraud should result in overall reduction in administrative costs. 

Collective bargaining to keep physician payment reasonable. Based on international experience, unilateral control of physician fees by government under a single-payer system would likely lead to under-payment of physicians over time, as has happened in countries such as Italy, Japan, Taiwan, and New Zealand. Collective bargaining has worked well in Canada for both physicians and for keeping the Canadian single-payer system cost-effective. We believe collective bargaining would be essential to achieving widespread physician buy-in and for the sustainability of a single-payer system in the U.S.

Reduced malpractice liability cost. With a universal single-payer system, the medical care component of malpractice liability could be removed from litigation and placed under the universal system. This would markedly reduce both the size and frequency of liability claims, because injured parties would not have to sue to assure coverage of their injury related medical bills. The liability component of physician fees could be reduced accordingly without affecting physician personal income.

Sustaining physicians during a pandemic: A federally funded single-payer system would not depend on state tax revenues and employment status, so federal funding could fall back on deficit spending during an abrupt economic contraction. The Fee-for-Time proposal would not restrict payment to office visits only and would pay for everything physicians do on behalf of their patients, including telephone and telehealth-based care, documentation time, and care coordination time. Cognitive services could easily switch to other modes of delivering care during a pandemic without abruptly cutting off physician fee-for-service income. Office visits and elective procedures might be delayed during a pandemic but could be rescheduled when safe without the threat of loss of employer-based health insurance.

Conclusion:

We believe fee-for-service payment of independent physicians using the Fee-for-Time proposal would be substantially simpler, more objective, and less costly to administer than either fee-for-service based on RBRVS and Evaluation/Management coding or “value-based” payment with pay-for-performance and risk adjustment. A universal single-payer system that paid hospitals with global budgeting and physicians based on time and expertise, with standardized salaries for employed physicians and time-based payment for independent physicians, would likely achieve far greater administrative savings and substantially lower total cost than any proposal based on existing Medicare fee-setting policies or shifting insurance risk onto doctors and hospitals.