Pay for performance in rheumatology: Will we get the carrot or the stick?†
The opinions expressed herein are solely those of the authors.
A number of developments suggest that large payers are placing increasing emphasis on value-based purchasing or pay-for-performance (P4P) programs for many medical services (see Table 1 for a glossary of terms used). This drive to tie provider compensation to objective measures of care processes and outcomes is fueled by the perception that payments are insufficiently related to value provided (1–3). No one can dispute that the US health care system is underperforming, especially with regard to chronic conditions such as the rheumatic diseases, or that fundamental changes in health care delivery processes will be required for improvement. A basic question is whether P4P programs will be designed to provide benefits to patients and reward providers for improving efficiency and outcomes, or will they be aimed exclusively at controlling payer costs—in other words, will we get the carrot or the stick?
Table 1. Glossary of terms used*
|P4P||Provider reimbursement plans that link payments more to value provided (improved clinical processes and/or reduced costs) and less to the volume of services billed.|
|TQM/CQI||Industrial programs used to continuously refine complex processes in order to optimize performance and products. These programs use plan-do-study-act cycles to evaluate positive alternatives in pilot tests before implementing them.|
|Six Sigma||A leading example of industrial CQI that seeks to reduce errors by 6 SDs, sigma being the symbol for 1 SD.|
|PQRIs or PQIs||Measurable indicators of provider and system adherence to clinical processes that support improved health care outcomes and/or decreased costs.|
P4P programs as now envisioned do not measure disease outcomes, but track the consistency with which providers follow processes that have been demonstrated to improve outcomes. For example, accepted osteoporosis P4P indicators assess the percentage of patients older than age 65 years who receive a bone density measurement (dual x-ray absorptiometry) and the percentage of patients with osteoporosis treated with a pharmacologic agent, rather than the fracture incidence in these high-risk populations.
This approach has antecedents in other industries where real-time quality control programs monitor manufacturing and service processes. Industry has long recognized that monitoring finished goods is inefficient in the extreme—by the time a single defective item is spotted, hundreds of defective items will have been produced. Furthermore, it may be difficult to establish a cause–effect relationship between process and outcome after the fact when processes are complex or the time from process to unfavorable outcome is prolonged.
To address this issue, total quality management (TQM)/continuous quality improvement (CQI) methods have been implemented in business and industry, based on the pioneering work of Deming (4). These methods involve an unending commitment to identify opportunities for improvement and to test alternative processes in small pilot cycles of change, called plan-do-study-act (PDSA) cycles. The Six Sigma movement has been one of the most successful applications of CQI, first adopted by Motorola in the 1980s and then made famous by General Electric's Jack Welch. Sigma is the symbol for 1 standard deviation; six Sigma suggests that by reducing manufacturing errors by 6 standard deviations through CQI, the number of defective products can be reduced to zero.
Health care systems, and especially academic health centers, have been slow to adopt CQI, but sufficient examples have been reported to document the value of this approach in improving health care delivery processes and patient outcomes. The most striking is a process improvement initiative started in the 1980s by anesthesiologists, hospitals, and the manufacturers of anesthesia drugs and equipment. Through standardization and CQI methods, complications and deaths from general anesthesia have been reduced from 1 in 500 to 1 in 200,000–300,000 cases (5).
More recently, the Geisinger Health System has developed a real-time quality control program called ProvenCare for monitoring and managing cardiac bypass surgery (6). A 40-item checklist is used to ensure optimal processes prior to admission, prior to and during surgery, postsurgery, prior to discharge, and during followup. ProvenCare has increased the number of patients discharged home (from baseline of 81% to 91%), reduced pulmonary complications (from 7.3% to 2.6%), reduced readmissions to the intensive care unit (from 2.9% to 0.9%), and reduced in-hospital mortality (from 1.5% to 0.0%). The program's confidence in its performance is reflected in a flat fee for all costs of surgery and followup care within 30 days.
Practice level CQI projects have reduced problems in rheumatology practice, including access to care (7, 8) and coordination of chronic disease management (9), and in other specialties as well, but few providers are trained in PDSA methods or engaged in continuous improvement initiatives. Changing this will be required for success in a P4P environment.
P4P Programs in Rheumatology
The American College of Rheumatology (ACR) recognizes the intent of payers to establish P4P programs. If physicians accept P4P and implement CQI, it is likely that these programs will benefit patients and lower costs. There is a concern, however, that these programs might focus narrowly on cost savings to the disadvantage of rheumatology practices and the patients we serve. To avoid this real possibility, the ACR has moved proactively to identify quality measures that are evidence based and linked to patient benefits and that can be quantified and monitored in clinical practice, while maintaining or increasing reimbursement (1). Rheumatologists then must adopt these measures.
The ACR has already endorsed a set of physician quality reporting indicators (PQRIs) for rheumatoid arthritis (RA), osteoporosis, gout, and drug safety. Osteoporosis PQRIs have been adopted by the American Medical Association–led Physician Consortium for Performance Improvement (PCPI), the AQA Alliance (AQA), and the National Quality Forum (NQF) and implemented by payers such as Medicare (CMS). The ACR is presently revising its RA measures set to be consistent with the 2008 ACR recommendations for the use of disease-modifying antirheumatic drugs in RA (10); national approval by the PCPI, AQA, NQF, and CMS is anticipated by fall 2008.
The current PQRIs for RA specify that a standardized data set, including a joint examination, functional status assessment, acute-phase reactant test, measurement of pain, and physician and patient global assessments, be documented within 3 months of diagnosis and at least annually thereafter. In addition, treatment should include a disease-modifying antirheumatic drug (DMARD) unless contraindicated, and if there is evidence of disease activity and/or progression after initiation of DMARD treatment, the DMARD dose or route of administration should be changed, the DMARD changed, another DMARD added, or glucocorticoid therapy initiated (1). Failing to address uncontrolled RA will not be an option unless a contraindication to accelerating treatment or patient refusal is documented.
The need for objective measurement of RA disease activity to support this PQRI was best demonstrated in the Tight Control of Rheumatoid Arthritis (TICORA) study that compared sustained, tight, outpatient control of disease activity with routine outpatient care. Sixty-five percent of patients in the tight control group were judged to have achieved remission compared with 16% of patients in the routine care group (11).
These findings mirror tight control program results for other chronic diseases, such as diabetes and hypertension. Rewarding adherence to performance measures instead of work volume for diabetes treatment was tested in a collaboration of a managed care organization in San Francisco and an independent practice association. Performance scores on quality, patient satisfaction, and practitioner efficiency were evaluated. The program cost $1,150,000 annually; savings in the first and second years approached $2 million and $3 million, respectively. Patient outcomes were not reported; however, the program demonstrated that adherence to performance measures need not cost more, and in this case, changing the financial incentives for physicians actually saved money (12).
Issues for P4P Programs
A national survey found that approximately 75% of physicians support financial incentives for quality, but only approximately 33% support public reporting of quality scores. Physicians were concerned that if they were judged on patient outcomes, they might be penalized for treating more challenging patients. One physician identified “the poor, the unmotivated, the obese, and the non-compliant” as patients he would be unable to treat efficiently (13). Clinicians also are concerned that a proliferation of performance measures might lead to an emphasis on treatments that can be readily reduced to numerical scores, crowding out those that cannot be as easily measured. Not all treatment options benefit patients equally, and not all measures deserve equal priority (14).
A questionnaire regarding P4P programs was distributed to nearly 2,500 primary care physicians in New York and Massachusetts. From the responses, a group of constructs was identified that had a significant impact on physicians' attitudes toward P4P programs. These included clinical relevance, unintended consequences, and financial salience. Prospective P4P programs in rheumatology should consider these findings (3, 15).
The degree of physician engagement will depend heavily on positive P4P programs. The central idea in any acceptable P4P program should be rewarding clinicians for delivering excellent and efficient care, and initially for measurable improvements from baseline. This approach is imbedded in CQI, as long as the overriding goal is to provide the best possible care for every patient, all the time. Punishment for failing to meet predefined standards or, worse, peer group mean performance levels should not be the basis for such programs. If the Geisinger and San Francisco examples can be generalized, bonuses could be funded from the economic benefits derived from fewer hospitalizations, fewer adverse drug reactions, and less need for urgent treatments (2, 3, 5, 10).
A proposal to establish a P4P program under Medicare has included the idea that 2% of the Medicare budget would be allocated initially to a pool for P4P bonuses; however, penalties for underperformance are also predicted, and the support for rewards would presumably further reduce payments for other services (16). Private payers have similar plans and seek to reduce costs through restricting important but more expensive treatments. Instead, payers should be encouraged to plan on higher profits accruing from a combination of short-term efficiencies and long-term reductions in disease morbidities and to invest in providers who accept the challenge to improve. This buy-in is most likely to happen if providers first document our commitment and additional CQI successes.
Off-setting provider costs
The time and effort required of providers to design and implement CQI-based P4P programs must be compensated, because these activities generally reduce traditional billable work, especially during the learning stage of the program and when an important goal is to replace unnecessary visits with more efficient care alternatives. An advantage of starting P4P programs with small PDSA pilots is their low cost and disruption relative to more complex, comprehensive, and untested approaches. Improvement work that reduces unnecessary provider visits must be rewarded separately rather than as a bonus to billings in order to break the dependence of practices on face-to-face provider–patient encounters. Success will require that provider incentives at least balance losses in incomes and cover practice overhead.
Any P4P program will no doubt require a bureaucracy to monitor performance and manage payments. It is essential that the costs of practice monitoring do not simply shift monies from providers and patient services to the managers or actually increase the health care bottom line. Payers must consolidate and standardize P4P programs instead of using them for competitive advantage if this is to be avoided. Presenting providers with multiple P4P processes and priorities, as drug precertification programs have done, will be viewed as a more unacceptable hassle. The same efficiency and effectiveness standards expected of the clinical community should be expected of and documented for the payers' P4P apparatus.
To be successful, those designing and implementing P4P programs must incorporate the CQI lessons from health care and other industries (17, 18). Change must be introduced in manageable increments over time and after testing in small PDSA cycles to reduce the risks of process breakdowns and unforeseen negative consequences. Physicians must become adept at using these methods in their practices, and payers must support such efforts, even, if necessary, by front-loading payments to share in the initial costs of less than optimal provider efforts to measure and improve performance.
The depth of change in physicians' practices required to measure disease activity, improve disease control and process efficiency, and move from reactive, individual patient visit–based care to proactive disease population team management should not be underestimated. A “step at a time” approach should be adopted by physicians and P4P programs. Initial performance measures should focus on high yield, realistically achievable changes—the multiple “low hanging fruit” opportunities. Successes and failures must be widely shared to accelerate the pace of improvement.
P4P will work best if the currently divergent interests in the health care system become aligned to provide optimal care at the lowest possible cost (5).
Benefits of P4P Programs
A well-designed P4P program for RA care emphasizing proven process improvements, CQI philosophy, and PDSA methods should result in numerous benefits for patients, providers, payers, society, and the US economy.
Patients would likely benefit from earlier and more accurate diagnosis and, when appropriate, more aggressive treatment to minimize irreversible joint damage and maintain function with fewer adverse drug reactions. The benefits of scientific discovery cannot be provided with existing delivery of care processes (19).
Practitioners would be rewarded not only financially, but also by contributing to the improved well-being of their patients and reduced health care costs. Greater efficiency and team care will help to maintain access to care in the face of projected rheumatology manpower losses (20). Without physicians' commitment to redesigning care, efforts by others cannot succeed.
The health care system as a whole would benefit by more appropriate and efficient use of medical resources.
A larger proportion of patients with arthritis who remain fully functional would reduce the indirect costs to society of these disabling diseases.
Providers' concern is that dollars saved, the easiest and most attractive measure of success for payers, will become the dominant metric of P4P. This possibility will be reduced if the rheumatology community and other providers become involved in creating and widely adopting P4P programs that use clear, evidence-based measures that lead to improved patient outcomes. Health care providers and professional organizations such as the ACR must take the lead in advocating for positive incentives because payers may be more drawn to the negative, choosing the stick instead of the carrot. Payers need to not only invest in improving the health system, but also share economic gains not just with insurance purchasers and shareholders, but with the providers whose efforts produce these gains. For all of these reasons, rheumatologists should anticipate and lead the effort to develop and implement P4P programs.
The research assistance of QD Healthcare Group (which was funded by Amgen and Wyeth) is gratefully acknowledged in the preparation of the manuscript.