The system for paying health care providers is extremely fragmented. In response, both the United States and the Netherlands are now experimenting with bundled-payment models, whereby a single prospective payment is made for all services for a patient with a given condition, even when multiple providers deliver that care. I believe that the ongoing Dutch experience with bundled payments has unique lessons for U.S. policymakers.
Bundled-payment efforts in both countries shift accountability to a single provider-led entity that must ensure quality, thereby emphasizing value over volume of care. The U.S. Centers for Medicare and Medicaid Services’ Comprehensive Care for Joint Replacement (CCJR) program, for example, uses a mandatory bundled payment for total-hip and total-knee replacement surgeries. In 2007, the Netherlands initiated a bundled-payment model for type 2 diabetes care and, subsequently, for chronic obstructive pulmonary disease and vascular-risk management. (A similar model for pregnancy and childbirth is underway.)
The orientations of the U.S. and Dutch models differ somewhat. The CCJR model has a downstream focus: improving care during inpatient stays and the 90-day post-discharge period, to limit the need for hospital readmissions. The Dutch model, by contrast, has an upstream focus: improving primary care to prevent expensive outpatient-specialist care and hospitalizations. I will describe how the Dutch bundled-payment model for diabetes care works, as well as its successes and challenges.
In our model in the Netherlands, insurers pay a bundled payment to a principal contracting entity — the care group — to cover a full range of diabetes-care services for a fixed period of 365 days. The care group, a new legal entity in the Dutch health care system, comprises multiple providers, often exclusively general practitioners. By signing the bundled-payment contract, the care group assumes both clinical and financial accountability for all diabetes patients assigned to its care program. The contract is limited to general diabetes care (services to manage the underlying disease and reduce risk for complications) and does not include services to address complex complications that may arise. Therefore, the model focuses on primary care.
General decisions about services covered in the diabetes-care bundle were made at a national level and, in 2007, codified in a Health Care Standard for type 2 diabetes. For the various components of diabetes care, the care group either delivers services or subcontracts with other providers. Insurers and care groups negotiate the price of the bundle, and the care group negotiates with the subcontracted care providers about fees for specific services. All services are covered under the basic benefit package for all Dutch citizens. The Dutch bundled-payment model is consistent with the principles of Michael E. Porter and Thomas H. Lee’s strategic value agenda for health care.
Keys to Success
In the four years since the Dutch bundled-payment model for type 2 diabetes was introduced, patient mortality rates and costs have dropped significantly. (My colleagues and I expect to report the specific numbers in a journal article in the next few months.) The model has had success for three key reasons:
- It was codified. The Dutch Diabetes Federation Health Care Standard (DFHCS), agreed on by all national provider and patient associations, specifies the minimum requirements for optimal diabetes care and sets the criteria for improvements. By law, the bundled-payment contract must include all services described in the DFHCS, which identifies what services to provide but not who delivers those services or where and how they are delivered. In addition, the DFHCS specifies a standardized minimum data set of quality measures, thereby giving care groups an incentive to adopt innovations and to reallocate tasks so that providers each do the work that best matches their qualifications.
- It fostered transparency through use of electronic health records. By 2010, three years after bundled payments were introduced, 66% of the care groups had web-based electronic health records (EHRs) where subcontracted providers were required to record their data. The EHR system made patient data available to primary care providers in real time and helped to reduce duplicated services. Web-based EHRs also enabled care groups to benchmark the performance of care providers, who could then learn from one another. In addition, the EHRs were used to generate accountability reports for insurers and to inform the public about care groups’ achievements. In interviews conducted by the National Institute of Public Health and the Environment, most providers said that they perceived this greater transparency as the main success of the reform.
- It optimized the value of clinical expertise. Care groups are led by providers, who use their clinical knowledge directly in decisions to achieve efficient, high-quality care. Therefore, fewer low-value services are purchased, and both overuse of unnecessary services and underuse of high-value services are avoided. For instance, after bundled payments were introduced, the number of routine check-ups went down for diabetes patients with well-controlled blood-glucose levels but went up for patients who needed more-intensive monitoring. Also, diabetes patients who had no abnormalities on their annual eye exam were switched to a biannual eye-exam schedule, consistent with Dutch clinical-practice guidelines.
Despite its successes, the Dutch bundled-payment model faces three main challenges that are relevant to U.S. policymakers:
The model is limited to primary care. Outpatient specialist care and inpatient care are still paid via existing hospital-payment systems. This distinction was probably wise in the early stages of implementation, as general practitioners (GPs) were being urged to adopt bundles. However, it potentially encourages GPs to refer the more-complex (and more costly) patients to specialists. Currently, some care groups are exploring whether to extend the care bundle to outpatient specialist care and inpatient care.
Quality measures should focus more on outcomes. Despite the initial goal of improving patient outcomes, most DFHCS quality measures still focus on process metrics, such as the percentage of diabetes patients whose HbA1c levels were measured in the past 12 months. I expect current measures to be replaced by measures that matter more to patients, such as those outlined by the International Consortium for Health Outcomes Measurement.
Better payment models are needed. Having provider-led care groups assume financial risks has been an important step in payment reform, but the Dutch health care system must move toward more-disruptive payment models that focus on caring for patients rather than merely treating disease. Models like global payments, analogue to the Alternative Quality Contract, are receiving scrutiny in the Netherlands. It might even be wise just to scale up the number of bundled-payment contracts for the most prevalent chronic conditions before introducing more-disruptive payment models, but that option has not yet been actively considered. We may simply have to wait for a willing Dutch insurer to take a chance. I am hopeful that the next step will come soon, so that Dutch providers that are willing to take the lead are not discouraged.
For now, as U.S. policymakers aim to strengthen the primary-care orientation of payment models, they should consider “going Dutch” — not by splitting the bill, but by bundling it.