Linking Payment and Quality of Health Care in Germany and the U.S. – Lessons Learned and Future Perspectives

Increasing health care spending in the United States as well as in Germany raises a fundamental question of whether health care is of sufficiently high value, given its costs. The answer to this question depends on the definition of “value” in health care. One of the most prominent definitions derives from the six aims of quality stated by the Institute of Medicine in Crossing the Quality Chasm: Care should be safe, effective, efficient, patient-centered, equitable, and timely.[1] Concerns about rising costs paid for mediocre quality of care have initiated the development of different approaches to enhance “value for money” in health care either by improving quality of care, lowering or at least stabilizing costs, or both.[2]

Value for money – How to link payment and quality of care?

According to economic theory, volume-based payment thwarts high quality of care. The classic fee-for-service payment scheme incentivizes quantity rather than quality: “the more you do, the more you get.” This results in delivering services to patients from which they may not gain benefit. Alternatively, capitation-based reimbursement or salary fosters care delivery “at its minimum”: “do as much as needed for reimbursement.”[3] Therefore, health policy and payers have increasingly called for linking payment and quality of health care.

Different approaches link payment and quality of care. The most common approach named “pay-for-performance” (P4P) explicitly incentivizes health care providers to reach targets on predefined performance measures. P4P programs can generally be characterized following three dimensions: “What, whom, and how to incentivize?”[4]

P4P relies on measuring quality in health care. Therefore, valid performance measures are needed. Performance can either be measured by process indicators (e.g., number of mammograms done in specified patient population), structural indicators (e.g., installation of an electronic reminder system for offering mammograms to women over the age of fifty), or outcome measures (e.g., mortality rate for breast cancer). P4P programs may target different domains such as clinical care (chronic care, acute care, or prevention), practice management (e.g., team training, implementation of information technology systems), or patient experience. Depending on the priorities of the programs, some P4P programs address misuse, whereas others target underuse or overuse of services. Measures that qualify for P4P may either be nationally or locally approved. P4P implies attempting to achieve a distinct target value, but design choices for P4P programs include choosing relative or absolute thresholds or whether improvement and/or actual quality level should be incentivized.

The basic assumption of P4P is that financial incentives are able to motivate behavior change in individuals.[5] Incentives in P4P programs are given to physicians, organizations, and/or patients. P4P incentives differ in amount and direction: they can either be provided as additional payment, withholding of payment, or penalty fee. Depending on the measures used for P4P programs, more or less complex risk-adjustment may be performed prior to benchmark. As most outcomes in health care result from complex interactions between patients, their morbidity level, and the health care providers involved, particularly outcome measures are often risk-adjusted in P4P programs. Alternatively, exception reporting may be available for health care providers to exclude patients who could not be treated as recommended either because of contraindications or patients’ refusal of treatment.

Initiatives linking payment with quality of care: A comparison between Germany and the U.S.

Both health care systems mainly rely on volume, rather than value-based payment schemes. However, in the United States half of all commercial health maintenance organizations (HMOs) including contracts with ambulatory physicians, hospitals, and nursing homes have integrated P4P into their payment scheme.[6] In 2007, a total of 256 P4P programs existed in the United States. Most of them incentivize clinical care, followed by resource use, information technology adoption, and patient experience.[7] U.S. programs typically rely on a limited number of (process or structural) measures, typically ranging from ten to twenty-five. Outcome measures are rarely applied in established programs. In order to avoid the problems of small sample size, most U.S. programs target groups of physicians, not individuals, that care for sufficient numbers of patients. Incentives are commonly spent as bonus payments with an estimated average payment size of 7 percent for physicians and 2.5 percent for hospitals.[8] Seventy percent of the programs use thresholds rewarding actual quality levels whereas 25 percent pay for improvements comparing two time periods.[9]

In Germany, 90 percent of the population is insured by one out of 153 non-profit sickness funds. The remaining population is privately insured by any of the forty-three private insurers. Sickness funds are financed by pay-roll tax and tax subsidies all collected in a national central health fund with risk-adjusted payments to sickness funds. Most of the physicians in ambulatory care (general practitioners and specialists) work in solo practice (60 percent) and are collectively contracted with sickness funds by regional physician associations (KV). Following these contracts reimbursement is given by fee-for-service according to nationwide fee-schedule. Caps limit the amount of paid delivered services per practice. Since 2009, sickness funds have to offer additional selective care contracts (e.g., General Practitioner (GP)-centered care contracts). Sickness funds and local physician organizations (e.g., German Association of General Practitioners) directly negotiate these contracts. Beneficiaries and physicians can enroll for free. Enrolled patients select one primary treating GP, with mandatory requirements for referrals to specialists. Late-evening consultation hours and partly waived co-payments for medications are some of the incentives for patients. Enrolled physicians financially benefit from a capitation-fee system supplemented by additional premiums for offering special services like minor surgery or ultrasound that exceeds payments in collective contracts.

Hospitals in Germany are reimbursed by prospective payment related to diagnosis related groups (DRGs). Thereby, each hospital stay is reimbursed due to the primary diagnosis and complicating co-morbidities. Referral is required for non-emergency admissions. Due to legal restrictions, outpatient services are rarely offered by German hospitals.

How is reimbursement in German ambulatory and hospital care linked to quality of care? P4P programs are not widely implemented in Germany. However, quality management strategies—mostly lacking a direct link to reimbursement—are integrated in both sectors. Collectively contracted physicians in ambulatory care require minimum qualification and equipment to get reimbursement for some of the delivered services like ultrasound or minor surgery. All ambulatory care providers have to introduce a quality management system (e.g., International Organization for Standardization (ISO), Qualität und Entwicklung in Praxen (QEP) or European Practice Assessment). Sickness funds can negotiate with the KV to offer bonus payments for quality targets refinanced by lowering payment to others. However, there have been few local P4P programs with incentives for performance based largely on structural indicators.

In addition to the fact that ambulatory care providers enrolled in GP-centered contracts also have to introduce a quality management system, they are asked to participate in quality circles where local (relative) benchmarking results regarding evidence-based pharmacotherapy is fed back and discussed among peers. In some GP-centered care contracts, prescription rates for generic drugs and influenza vaccination rates are measured and incentivized by P4P with absolute thresholds. Due to a current change of the legal framework for selective care contracts (§73b SGB V or GP-centered care) future GP-centered care contracts have to refinance payments above the capped reimbursement provided in collective contracts by savings due to more efficient care.

All hospitals have to publish a biannual structured hospital report including performance on nationally agreed quality indicators (structural, process, and outcome indicators). The introduction of quality management systems is also mandatory for German hospitals. German DRGs do have P4P embedded in them. Some complications that may occur as a result of suboptimal inpatient care are not reimbursed as they are excluded from the DRG complexity level measure. In addition to that, early re-admissions (within the maximum length of hospital stay expected due to the primary diagnosis) are not fully reimbursed as both inpatient cases will be grouped and paid together. Explicit P4P programs are rare in inpatient care. One example is the German Quality Inpatient Indicator set (G-IQI) introduced by one of the largest private hospital providers in Germany, Helios AG. Apart from process indicators, G-IQI also includes outcome indicators like mortality rates after major surgeries.

In comparison to the United States, P4P in Germany is in its early stages of development. Although quality measurement is widely implemented, it is only rarely linked to payment. Therefore, further design and implementation of P4P in Germany may benefit from the lessons learned from decades of P4P in U.S. health care.

Lessons learned from pay-for-performance programs

Although evidence on the effects of P4P has increased over the past decade results remain inconclusive.[10] However, some lessons could be learned from the evidence gained so far. Limited effectiveness is partially due to a flawed design of P4P programs including inadequate information technology support and informatics or inconsistent coding of diagnosis, procedures, and outcomes.[11]

Involving all stakeholders in P4P development, implementation, and evaluation strongly enhances the effects of P4P programs.[12] The design choices made for a P4P program should be based on its purpose and set into the context of the payment scheme. If P4P is going to be implemented in a fee-for-service system that at baseline tends to induce overuse of services, it should discourage overuse. Similarly, if P4P is implemented in a capitation-fee system, it should encourage the delivery of specified high quality services that go beyond “minimal care.” P4P programs designed to support minimal quality standards (e.g., as part of a capitation-fee system) work fairly well; if “optimal quality” is the target, however, P4P may face a number of obstacles. This observation could be explained by the goal-gradient theory: If a goal is perceived to be not immediately attainable, little effort will be made to achieve it.[13] And vice versa: The closer a person is to achieve the target, the higher his or her effort to achieve it. However, as soon as the target is achieved, motivation immediately decreases explaining the “ceiling effect” observed in P4P programs.[14] Therefore, program developers need a good understanding of baseline performance before designing P4P incentives to increase quality. Goal-gradient theory also supports the idea of combining incentives for both actual quality level and quality improvement over time thereby attracting high- and low-performers likewise.

Economic theory suggests that the recipient of an incentive must at least be compensated for the incremental net costs of undertaking the desired action (including opportunity costs e.g., time spent on documentation rather than seeing patients).[15] As P4P requires measurement and measurement requires documentation, initial investment may be particularly high (e.g., for installing an electronic documentation system). Behavioral economics teaches us that small but frequent incentives may be better than large lump-sums.[16] Individuals rate a $10 gain  greater when it is a gain from $0 to $10 compared to a gain from $100 to $110, a phenomenon called “mental accounting.” This supports the idea that P4P incentives may better be paid apart from basic reimbursement in order to reduce mental accounting. Even more important than frequency and way of getting incentives is the timelag between performance and payment. As humans tend to “hyperbolic discount” (i.e., steep initial discounting), money received right away after performance is perceived as higher in value compared to money received in the future.[17] Therefore, it may be crucial to shorten the time lag between performance and incentive as far as possible. This may lead to a system of “fee-for-desired-services” directly rewarding services that indicate high quality of care (e.g., directly paying an additional payment of $5 for each mammogram offered to women above the age of $55 instead of measuring and incentivizing process performance annually). P4P program developers may also question if it is better to give benefits to high-performers or penalties to low-performers. It is known that a perceived loss of money is even a stronger motivator for desired behavior than receiving a benefit.[18] However, anger and suspicion regarding fair benchmark (including the question of sufficient risk-adjustment) may limit implementation and feasibility of penalizing P4P programs.[19] Extensive risk-adjustment requires comprehensive data to be gained leading to higher costs, thereby limiting cost-efficiency of P4P programs. In general, evidence on cost-efficiency of P4P remains scarce and inconclusive.[20]. However, as positive cost-effectiveness has been determined in a number of studies P4P may increase value for money rather than saving money.[21]

The overall assumption that P4P motivates health care providers to change behavior inherently requires measured indicators to be under provider control. Outcome indicators may insufficiently be controlled by provider behavior alone (e.g., depending on patients’ overall health status, co-morbidities, health seeking behavior, and adherence). This fact may explain why P4P for outcome indicators has been found to be less effective than P4P on process indicators.[22]

If P4P is going to be implemented, possible unintended consequences have to be considered and deliberately ascertained. This requires piloting indicators before implementing them in a P4P program.[23] “Teaching to test” is one of the major concerns against the implementation of P4P in health care: Providers’ intrinsic motivation to perform the complex tasks required for comprehensive patient care may be crowed out by introducing a strong extrinsic motive to perform a narrow number of tasks that are easy to measure. Although evidence supporting this concern is still limited[24] and there are few studies investigating this concern, it has to be kept in mind that P4P potentially worsens care. Particularly in case of caring for patients with multiple chronic conditions P4P programs may set perverse incentives hindering comprehensive and coordinated long-term care[25] by focusing on specific aspects of individual chronic diseases. Patients themselves may be unaware of P4P programs but recent work suggests that they can have concerns about P4P incentivizing basic measures that are perceived to be fundamental. Patients also emphasized concern about potential unintended consequences like reduced focus on non-incentivized areas.[26] Strategies against these unintended consequences are the introduction of large, comprehensive measure sets, rotating incentivized measures, and the inclusion of patient-centered outcome measures like patient satisfaction or continuity of care. It is strongly recommended to continuously evaluate P4P programs to detect unintended consequences as early as possible.

Future prospects

Increasing value for money spent in health care will stay one of the major goals of health policy and management in the United States as well as in Germany. Linking payment to quality of care may be approached by introducing P4P. Due to the lessons learned from prior experiences with P4P, future programs should be developed in collaboration with multiple stakeholders including payers, providers, and patients. The U.S. National Quality Forum recently reported a Multiple Chronic Conditions Framework emphasizing that future quality improvement initiatives should promote collaborative care among providers and across settings at all levels of the health care system, assess care longitudinally, be as inclusive as possible regarding patients with multiple chronic conditions, sparsely use risk-adjustment (only in case of outcome measures), and include patient-important outcomes.[27]

In Germany, the era of cross-sectoral quality measurement and improvement has recently begun with commissioning the AQUA Institute Göttingen (Institut für angewandte Qualitätsförderung und Forschung im Gesundheitswesen). This institution will develop methods to measure quality across the sectors of inpatient and ambulatory care thereby changing the focus from sectoral performance to longitudinal quality of care (§137 SGB V). Reimbursement may in the future be linked to cross-sectoral quality of care thereby demanding collaboration among providers from different sectors. This may also require new approaches to define “shared accountability.”

Selective care contracts like GP-centered care contracts (§73b SGB V) or specialist care contracts (§73c SGB V) provide new opportunities for linking payment with quality of care like P4P or “pay-for-desired-services” as these contracts allow the negotiation of additional value-based premiums that have to be covered by savings due to increased quality and efficiency of care. However, collective contracts in German ambulatory care may also expand successful local quality improvement initiatives to large-scaled nationwide strategies, but may face barriers regarding fixed budgets.

In the United States, the Affordable Care Act provides several opportunities for value-based payment models.[28] The Medicare Shared Savings Program provides payment for accountable care organizations if they introduce quality of care measures including care coordination, patient and caregiver experience, as well as hospital admissions for ambulatory care sensitive conditions. In addition to that, a new “value-based payment modifier” will be implemented in 2015 providing differential payment based on quality and costs of care. As this provision is planned to be budget neutral it may combine bonuses and penalties for providers. Existing P4P programs such as in hospital care are planned to be expanded. Starting in fiscal year 2013 with the Hospital Value Based Purchasing Program, the Centers for Medicare & Medicaid Services (CMS) will make value-based incentive payments to acute care hospitals based on performance indicators including patient experience.

Against the background of lessons learned from P4P so far, future initiatives to enhance value for money in health care should keep in mind that P4P programs are only one arrow in the quiver of quality improvement strategies that should be flanked by complementing strategies and critically evaluated in order to serve the overarching aim of high quality of care at reasonable costs.

 

Tobias Freund, M.D., University Hospital Heidelberg, Department of General Practice and Health Services Research, Vossstrasse 2, D-69115 Germany. Email: tobias.freund@med.uni-heidelberg.de

Cynthia M. Boyd, M.D., M.P.H, The Johns Hopkins University, 5200 Eastern Avenue, COAH, 7th Floor Center Tower, Baltimore, MD 21224 cyboyd@jhmi.edu



[1] Institute of Medicine, Crossing the Quality Chasm: A New Health System for the Twenty-first Century (Washington, DC: National Academy Press, 2001).

[2] Robert A. Berenson, “Moving Payment from Volume to Value: What Role for Performance Measurement?” Policy Brief (Washington, DC: Urban Institute, 2010).

[3] Meredith B. Rosenthal and R. Adams Dudley, “Pay-for-performance. Will the latest payment trend improve care?” Journal of the American Medical Association 297:7 (2007):740-744.

[4] Ibid.

[5] Gerd Flodgren, Martin P. Eccles, Sasha Shepperd, Anthony Scott, Elena Parmelli, Fiona R. Beyer, “An Overview of Reviews Evaluating the Effectiveness of Financial Incentives in Changing Healthcare Professional Behaviors and Patient Outcomes,” Cochrane Database of Systematic Reviews 2011, Issue 7. Art. No.: CD009255. DOI: 10.1002/14651858.CD009255

[6] Meredith B. Rosenthal and R. Adams Dudley, “Pay-for-performance. Will the latest payment trend improve care?” Journal of the American Medical Association 297:7 (2007):740-744.

[7] Frank Eijkenaar, “Pay for Performance in Health Care: an International Overview of Initiatives,” Medical Care Research and Review (2012).

[8] Meredith B. Rosenthal and R. Adams Dudley, “Pay-for-performance. Will the latest payment trend improve care?” Journal of the American Medical Association 297:7 (2007):740-744.

[9] Meredith B. Rosenthal, Bruce E. Landon, Sharon-Lise Nommand, Richard G. Frank, Arnold M. Epstein, “Pay for Performance in Commercial HMOs,” The New England Journal of Medicine 355 (2006):1895-1902.

[10] Anthony Scott, Peter Sivey, Driss Ait Ouakrim, Lisa Willenberg, Lucio Naccarella, John Furler, Doris Young, “The Effect of Financial Incentives On the Quality of Health Care Provided by Primary Care Physicians,” Cochrane Database of Systematic Reviews Sep 7:9 (2011):CD008451.

[11] Stephen M. Campbell, Anthony Scott, Rhian Parker, Lucio Naccarella, John S. Furler, Doris Young, Peter Sivey, “Implementing Pay-for-Performance in Australian Primary Care: Lessons from the United Kingdom and the United States,” Medical Journal of Australia 193:7(2010): 408-411.

[12] Pieter Van Herck, Delphine De Smedt, Lieven Annemans, Roy Remmen, Meredith B. Rosenthal, Walter Sermeus, “Systematic review: Effects, Design Choices, and Context of Pay-for-Performance in Health Care,” BioMed Central Health Services Research 10 (2010):247.

[13] I Siva, “Using the Lessons of Behavioral Economics to Design More Effective Pay-for-Performance Programs,” American Journal of Managed Care 16:7 (2010): 497-503.

[14] Stephen M. Campbell, David Reeves, Evangelos Kontopantelis, Bonnie Sibbald, Martin Roland, “Effects of Pay for Performance on the Quality of Primary Care in England,” New England Journal of Medicine 361:4 (2009):368-78.

[15] Meredith B. Rosenthal and R. Adams Dudley, “Pay-for-performance. Will the latest payment trend improve care?” Journal of the American Medical Association 297:7 (2007):740-744.

[16] I Siva, “Using the Lessons of Behavioral Economics to Design More Effective Pay-for-Performance Programs,” American Journal of Managed Care 16:7 (2010): 497-503.

[17] Ibid.

[18] Ibid.

[19] Ibid.

[20] Martin Emmert, Frank Eijkenaar, Heike Kemter, Adelheid Susanne Esslinger, Oliver Schöffski, “Economic Evaluation of Pay-for-Performance in Health Care: A Systematic Review,” The European Journal of Health Economics (10 June 2011), Epub ahead of print.

[21] Ibid.

[22] Pieter Van Herck, Delphine De Smedt, Lieven Annemans, Roy Remmen, Meredith B. Rosenthal, Walter Sermeus, “Systematic review: Effects, Design Choices, and Context of Pay-for-Performance in Health Care,” BioMed Central Health Services Research 10 (2010): 247.

[23] Stephen M. Campbell, David Reeves, Evangelos Kontopantelis, Bonnie Sibbald, Martin Roland, “Effects of Pay for Performance on the Quality of Primary Care in England,” New England Journal of Medicine 361:4 (2009):368-78.

[24] See: Ibid.; Robert M. Wachter, Scott A. Flanders, Christopher Fee, Peter J. Pronovost, “Public Reporting of Antibiotic Timing in Patients with Pneumonia: Lessons from a Flawed Performance Measure,” Annals of Internal Medicine 149:1 (1 July 2008):29-32; and Adam A. Powell, Katie M. White, Melissa R. Partin, Krysten Halek, John B. Christianson, Brian Neil, Sylvia J. Hysong, Edwin J. Zarling, Hanna E. Bloomfield, “Unintended Consequences of Implementing a National Performance Measurement System into Local Practice,” Journal of General Internal Medicine 27:4 (April 2012):405-12, Epub 13 October 2011.

[25] Cynthia M. Boyd, Jonathan Darer, Chad Boult, Linda P. Fried, Lisa  Boult, Albert W. Wu. “Clinical Practice Guidelines and Quality of Care for Older Patients with Multiple Comorbid Diseases: Implications for Pay for Performance,” Journal of the American Medical Association 294:6 (10 August 2005):716-24.

[26] Kerin L. Hannon, Helen E. Lester, Stephen M. Campbell. “Patients’ Views of Pay for Performance in Primary Care: a Qualitative Study,” The British Journal of General Practice 62:598 (May 2012):322-8.

[27] National Quality Forum, “Multiple Chronic Conditions (MCC) Measurement Framework,” 2012.

[28] Robert A. Berenson, “Moving Payment from Volume to Value: What Role for Performance Measurement?” Policy Brief (Washington, DC: Urban Institute, 2010).

 

The views expressed are those of the author(s) alone. They do not necessarily reflect the views of the American-German Institute.