Pay-for-Performance - An option for Germany?

Germany is considered a developing country when it comes to “Pay for Performance” (P4P). P4P has also established itself in Germany as a name for strategic compensation systems in which part of the compensation is linked to the measured performance level. Two recently published reports (Veit et al. 2012 and Charlesworth et al. 2012) explore the possible uses of P4P in Germany. Against the background of these reports, this article presents the possibilities and limitations of P4P in Germany.

>> Two years ago, the OECD published a highly acclaimed study on the price-performance ratio of health spending (OECD 2010). Among the topics examined there was the use of quality incentives in compensation systems. These incentive systems usually go under the name “Pay for Performance” (P4P). The OECD created an overview of the use of P4P instruments in the OECD countries. While numerous areas of use have been documented for the United States - and most other OECD countries - not a single use has been proven for Germany. One had to get the impression that Germany is a developing country when it comes to P4P.
Two recent studies put this impression in a somewhat better light. On behalf of the Federal Ministry of Health, Veit et al. The BQS Institute prepared a report on Pay for Performance (Veit et al. 2012). This also included an overview of the status quo of P4P in Germany. Even if performance-oriented remuneration elements are not excluded in collective bargaining, the overview shows clearly: P4P takes place in selective contracts in Germany. However, the projects mentioned are mostly small, limited in terms of space and indications. The overview in the European comparative study by the Nuffield Trust in London is similarly meager (Charlesworth et al. 2012). Compared to the United States, where half of all insurers have established P4P in their compensation systems and over 250 different P4P initiatives were counted in 2007 (Boyd/Freund 2012), the reluctance in Germany is striking.
Why “Pay for Performance”?
It is a key challenge for every healthcare system to increase the quality of care. The need for quality improvements has been impressively demonstrated several times. In the United States, the Institute for Medicine documented this in two groundbreaking reports - first To Err is Human (Kohn et al. 2000) and then Crossing the Quality Chasm (IoM 2001). In Germany, the 2000/2001 annual report of the Advisory Council on Health Care with the programmatic title “Over-, Under- and Incorrect Care” plays a similar role.
Studies by the OECD show that healthcare costs in Germany are above average. However, the quality of care falls short of that in countries with far lower spending. Although there are definitely model areas, especially in acute care, the results for measures such as avoidable mortality or avoidable complications of chronic diseases are mixed (Nolte/McKee 2011). The rate of leg amputations due to diabetes is three times higher than in the OECD comparison (OECD 2011). This suggests that the German healthcare system has weaknesses in the area of structured treatment of chronic diseases, which account for an increasing share of the morbidity burden.
However, it cannot be said that no efforts are made to ensure quality in Germany. In Germany there is a wide catalog of quality assurance measures, which also include mandatory public quality reports. In particular, hospitals are required to submit structured quality reports every two years. These include 182 different quality indicators. Some health insurance companies have begun to enrich these indicators with additional key figures that they obtain from their own billing data or patient surveys. Some providers have gone a step further and have developed and voluntarily reported sophisticated outcome indicators. One example is the German Inpatient Quality Indicators G-IQI (Busse et al. 2009). These indicators were further developed together with health insurance companies and also applied to statutory health care (see http://www.qualitaetssicherung-mit-routinedaten.de/).
But the measures in Germany relate very much to quality measurement. This is very fragmented and strongly focused on structural aspects and certification. What is conspicuously missing - compared to the United States and other European countries - is the link with the compensation systems. Compensation systems often create negative incentives for quality. Compensation systems usually reward the treatment of illness, not the prevention of it. Avoidable complications trigger the need for additional compensable treatment. Remuneration is usually independent of the quality of the service. It is therefore not surprising that the focus of the quality debate quickly moved from reporting to remuneration systems: could at least part of the remuneration be tied to quality? Pay-for-performance systems usually leave the underlying compensation system - usually individual performance compensation - unchanged, but dedicate a fixed proportion of the compensation volume to performance-related bonuses. The bonuses can be awarded both when fixed quality goals are achieved and when concrete quality improvements are measured.
In the USA, the Medicare Payment Advisory Commission (MedPAC 2003) has developed the following criteria for developing criteria for pay-for-performance systems:
• Key figures must be evidence-based. They must also be accepted and understood by those affected.
• A majority of service providers must be able to achieve improvements in terms of the chosen metrics.
• The chosen metrics should not prevent service providers from also treating patients with riskier or more complex care needs.
• The key figures should be able to be determined without excessive effort.
Weak evidence
Pay for performance is a theoretically appealing concept. Nevertheless, there is little empirical evidence that these programs are effective. The evaluation of model projects in the USA has been almost entirely negative. The world's largest P4P project - the introduction of the "Quality and Outcomes Framework (QOF)" in the English NHS - was also not evaluated positively. After an overview of the relevant literature, the OECD came to the sobering conclusion in the report mentioned that Pay for Performance was very expensive and without major improvements in care (“very expensive […] without large increases in care” - OECD 2010, 148). Veit et al. give an overview of the disappointing evidence situation and state a “contradiction between evidence and everyday experience” (Veit et al. 2012, B43).
From the perspective of economic theory, there is a lot to be said for pay for performance. The information asymmetries between service providers and payers lead to the usual problems of delegated actions (agency problems). These in turn lead to suboptimal quality results. In this context, the task of Pay for Performance is to reduce information asymmetries by providing quality indicators on which service providers and payers can base their remuneration contracts. In practice, however, this depends on the quality of the quality indicators.
In this context, the economic literature on delegation theory (principle-agent theory) emphasizes that measurability and incentives are complementary in nature (e.g. McClellan 2011): If the desired results are fully measurable, strong incentives can be set. However, if the desired results can only be measured incompletely, the incentives should also be weaker. In relation to the healthcare system, incentives that are too weak lead to inefficient care (and, if the costs are borne by third parties), too expensive. However, if strong incentives are linked to indicators that only incompletely reflect the desired quality and quantity of care, simultaneous over- and under-supply can occur.
Measurability is therefore the Achilles heel of pay for performance. Veit et al. state: “Most of the problems with P4P arise from problems of quality measurement” (Veit et al. 2012, A11). Do the chosen indicators actually measure what they are supposed to measure? Indicators that are too broad can confuse service providers, while indicators that are too narrow can lead them to have the wrong focus while neglecting other aspects. Unintended side effects are a clear danger of Pay for Performance.
In addition to measurability, two other problems limit the effectiveness of P4P programs: the coordination problem and the attributability problem. The coordination problem arises in health systems with a fragmented payer structure - such as in Germany or the United States. If a single payer runs a P4P program in such a fragmented landscape, the impact of the program is naturally limited. From a service provider's perspective, the performance-related fee components only make up a small proportion of the total remuneration. It is possible that the incentives set by one payer are even counteracted by different incentives that come from programs of other payers (cf. Rosenthal 2006).
The problem of accountability arises when a patient has contact with different service providers during a treatment episode. The quality incentives may not apply to all providers in the care chain. But then the influence of a change in the behavior of an individual service provider on the overall result is negligible (cf. McClellan 2011). But even if all service providers involved (family doctor, specialist, hospital, physiotherapist, etc.) are rewarded based on success: to whom should the quality improvement be attributed?
Prerequisite for success for P4P
How can Pay for Performance be a success? How can the problems of measurability, coordination and attributability be best solved?
First of all, it is necessary to develop meaningful quality indicators for a successful P4P implementation. A uniform data basis is also required for this. In particular, it is necessary to track patients' treatment processes across service provider boundaries throughout the entire treatment episode. This is only possible with the data from the cost bearers, not the service providers, because only this data allows the necessary population reference. In addition, to ensure sustainable and consistent incentives, all payers should, if possible, use the same indicators. To quote McClellan:
“Just as all healthcare players today rely on standard codes for thousands of procedures and diagnoses in their fee-for-service payments, consistent methods of measuring quality and cost performance would enable more accurate measurement and thus a greater ability for providers and payers to compete on quality and cost. Such consistent measures are also a necessary feature of reforms that seek to increase consumer pressure for greater efficiency” (McClellan 2011, 85).
Porter and Guth (2012, 41) go one step further: they demand quality indicators that are not only uniform nationally but, if possible, also internationally.
Measurability and coordination therefore require action across all cost bearers, while a solution to the attribution problem requires cooperation on the part of the service providers. But this is not about eliminating competition as a regulatory feature on both sides. There are many design options, especially with P4P - you shouldn't forego competition as a way to discover which options lead to the best success. Rather, it is about balancing the relationship between cooperation and competition. However, providing information about the quality of care is a collective good that is unlikely to be successful in the face of competition.
Finally, collective learning from experience must also be guaranteed. This includes not only enabling and supporting model projects in this area, but also making the results and experiences available to a wide range of payers and service providers. Above all, it should be possible to establish successful model projects on a large scale in the short term. The CMS Innovation Center in the United States offers an example of how this can be anchored institutionally (Göpffarth 2012).
Conclusion
It's a paradox: In Germany there are long and ongoing discussions in the area of P4P, but hardly any activity. In the United States, on the other hand, there is an overwhelming amount of activity in this area, but it has remained without resounding success. Germany, with its collective agreement system and the cooperation between health insurance companies anchored in SGB V, has better conditions for making Pay for Performance successful than the fragmented and hyper-competitive system of the United States. What one country lacks in terms of innovation and willingness to experiment, another country lacks in the ability to implement. Strangely enough, the situation with P4P in Germany is similar in many ways to the situation in the USA: implementation, especially within the framework of selective contracts, keeps the projects small, unrepresentative and without impact. Above all, the contracts remain trade secrets - collective learning is almost impossible.
The United States is establishing an innovation center to institutionalize learning from experience. In Germany, too, an institution that develops key performance indicator systems based on a uniform data base and makes these available to contractual partners would be helpful. Because, as both newly presented reports show, quality-oriented remuneration systems will not continue without political support.
Despite all the euphoria, one should not forget: Pay for performance is not an all-purpose weapon. Rather, it is an isolated instrument for achieving selected quality goals. P4P often leaves the underlying compensation systems, with all their disincentives, unchanged. P4P can therefore only represent one element of a wider reform of compensation systems - which also includes episode compensation, shared savings and capitation. The same problem of experimentation, but also learning and implementation, applies here too. <<