Health care is the largest part of the national economy, and the costs of maintaining it are growing at an unsustainable rate. As a result, Medicare and other payers are exploring value-based payment models to better control both short- and long-term costs.
Value-based models shift payment from the services provided to the results achieved. These models take different forms: They may have requirements for specific conditions, such as bundled payments for hip or knee replacements, or they may take on different organizational models, such as accountable care organizations. Their common feature is that health care organizations assume more risk for the cost of patient care, making it imperative to identify high-risk patients and avoid costly complications, such as readmissions or health care–acquired infections.
Taking on risk for financial or clinical outcomes forces a provider to make better decisions regarding outcomes. But doing so requires access to better information that is personalized to the patient and localized to the health system. Successful organizations need to move beyond investments made simply to meet new reporting requirements; they must embrace continuous learning systems that will transform data, especially outcomes data, into readily available information geared toward the patients they manage and the decisions they make.
New Medicare requirements are a major factor in the shift to value-based care. In particular, the Medicare Access and CHIP Reauthorization Act of 2015 fundamentally changed how Medicare pays for care, from a fee-for-service model to payment models that emphasize value-based care.
According to the final rule issued by the Centers for Medicare & Medicaid Services, providers will be paid under MACRA via two pathways: the Merit-based Incentive Payment System or advanced Alternative Payment Models. Initially, most health care providers are expected to be paid under MIPS, which adds new outcomes-focused quality metrics to fee-for-service reimbursement. Over time, more and more payments will be made through APMs; CMS estimates that more than 50 percent of health care spending will fall under such models in the next two to three years.
One type of APM is the bundled payment model. Bundled payments are fixed payments for a defined service such as a hip replacement or care after a heart attack. In November 2015, CMS mandated approximately 800 hospitals to participate in a retrospective bundle payment model called the Comprehensive Joint Replacement bundle. This program fixes a payment level (e.g., $25,000) and phases in risk over four years until the provider organization is responsible for up to 20 percent of downside risk as well as upside opportunity. In July, CMS announced its intention to put in place a second mandated bundled payment program for coronary artery bypass grafting and acute myocardial infarction.
Under MACRA, outcomes measures, including patient-reported outcomes, are considered among the most important measurements for programs under regulation. Both CJR and the proposed cardiac bundles program tie payment rates much more closely to these outcomes measures than previous programs have.
While MACRA specifically applies to Medicare, private payers are also moving toward value-based care with new payment models, many of which mirror CMS programs. It is increasingly clear that, while new reporting requirements require new types of data collection, viewing the shift to value-based care only through this lens is shortsighted. Succeeding under value-based payment models will require measuring the outcomes that matter most to patients and the health care organization, as these will be the critical data not only for successful reimbursement and avoiding costly penalties but for providing good care at lower costs.
Although health care data are accumulating at an enormous rate owing to the widespread adoption of electronic health records and other related systems, these data are largely transactional and provide limited understanding of results that are relevant to patients. The term “outcomes management” was coined in 1988 by Paul Ellwood, M.D., in a memorable Shattuck Lecture for the Massachusetts Medical Society. Ellwood defined outcomes management as a "technology of patient experience” that estimates as best as possible the relation between medical interventions and health outcomes in a language understandable by patients.
These concepts, more recently repopularized by thinkers including Michael Porter at Harvard Business School and Nick Black at the London School of Hygiene & Tropical Medicine, are emerging as a new approach to achieving success with value-based care programs.
Continuous learning systems
While continuous learning systems may sound like an approach reserved for academic medical centers, nothing could be further from reality. All organizations hoping to succeed in value-based arrangements need access to actionable data for timely decision-making. Critical for value-based decision-making is the standardized, condition-specific information that Ellwood described. The traditional approach to collecting such data has been the patient registry.
The problem with patient registries is that they are resource intensive, costly and rarely standardized at a national level. Big data technologies, however, have emerged in the past two to three years that can automatically cull clinical outcomes data from a variety of sources and obtain outcomes data directly from patients. With appropriate modern technologies, these processes can be nearly fully automated so that the internal resources necessary to drive the programs are minimized.
With similar data from multiple organizations in continuous learning systems, health care organizations have access to comparative outcomes data on treatments, pathways, supplies, facilities and so forth. Furthermore, these large data sets enable highly accurate predictive analytics for personalizing decisions. For example, in bundled payment programs, the data can identify patients who are most likely to experience high-cost events before they occur, allowing organizations to target resources to avoid problems.
With private payers rapidly joining the value-based care movement, health care organizations need to start thinking beyond the data collection and reporting requirements of today. They need to consider how to scale for the ever-evolving requirements of the future.
Only by implementing outcome management tools that employ existing systems — and allowing for automation, integration and analytics — can health care organizations transform into learning systems that not only understand more than population health but also deliver highly personalized care focused on value and quality. Organization leaders who understand and use their patient outcomes for timely decision-making will be the long-term winners in value-based care.
Richard Gliklich, M.D., is the founder and CEO of OM1 in Cambridge, Mass., and is a chaired professor at Harvard Medical School.
The opinions expressed by the author do not necessarily reflect the policy of the American Hospital Association.