Published online before print June 1, 2009, doi: 10.1161/CIRCULATIONAHA.109.869958
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Circulation. 2009;119:2965-2966.)
© 2009 American Heart Association, Inc.
Editorial |
Pay for Performance
A Work in Progress
From Partners Healthcare System, Brigham and Women’s Hospital (T.H.L.), Massachusetts General Hospital (T.G.F.), and Harvard Medical School.
Correspondence to Thomas H. Lee, MD, Partners Healthcare System, Prudential Tower, 11th Floor, 800 Boylston St, Boston, MA 02199. E-mail thlee{at}partners.org
Key Words: Editorials • hypertension • health policy and outcomes research
Any time money changes hands in health care, perverse consequences can result. This fact of financial life applies to the entire spectrum of payment options, including the 2 methodologies that define its extremes: fee for service potentially rewards clinicians for doing more than patients need, whereas capitation potentially rewards them for doing less. Pay for performance is an intermediate payment approach with some attractive features, but it carries its own set of potential problems. In the absence of any perfect payment model, the challenge for leaders in health care is to understand the risks of various payment methods and consider how to mitigate them.
Article see p 2978
This challenge has never been timelier. Interest in changing the payment system is approaching commitment because of the combination of rising healthcare costs, a historic economic downturn, and growing concern about the high proportion of Americans who are uninsured or underinsured.
Pay for performance is widely considered a step in the right direction for the fragmented US healthcare system because it introduces incentives to clinicians and hospitals for more than the performance of visits, tests, and procedures. In its more rudimentary forms, pay for performance provides small (eg, 1% to 2%) bonuses for reduction of errors of underuse (eg, failure to measure low-density lipoprotein cholesterol in patients with diabetes mellitus) in their patient population. Just asking providers to worry about a population rather than just the individual patients who present with acute complaints is a subtle game-changer for American medicine.
But pay for performance has limitations and raises concerns. At least in its early forms, it doesn’t seem to produce dramatic change. For example, hospitals that participated in a pay for performance demonstration project funded by the Centers for Medicare and Medicaid Services had 3% to 4% improvements in composite measures of quality for heart failure, acute myocardial infarction, and pneumonia compared with hospitals that engaged only in public reporting.1 And because most pay for performance programs are focused on reducing underuse, cost savings have not been demonstrated thus far.
An even greater concern about pay for performance is whether there could be unintended consequences that cause incentives for better quality to do more harm than good. Other approaches are likely to be invoked to control costs (eg, bundled payments, capitation, or fee reductions), and, in some markets, these cost-control strategies are already being combined with pay for performance bonus programs focused on quality. But what if clinicians focus only on the problems for which there are pay for performance incentives and ignore their patients’ other issues? Or will their patients’ other medical conditions cause clinicians to fall short of their pay for performance goals? Will physicians decide that the easiest way to achieve these bonuses is to avoid sick or complex patients? The ratio of speculation to research on these issues has thus far been high.
Petersen et al in the current issue of Circulation do not resolve these questions but provide some welcome data elsewhere in this issue.2 They studied the impact of 6 common coexisting conditions on quality of care for hypertension in 141 609 patients in the Veterans Affairs system. All of these patients had hypertension and were separated into 4 groups on the basis of their other medication conditions. Of this cohort, 49.5% had hypertension-concordant conditions (eg, diabetes mellitus, ischemic heat disease, and dyslipidemia) only; 9% had only hypertension-discordant conditions (eg, arthritis, depression, and chronic obstructive pulmonary disease); 26% had both; and 16% had neither.
Cynics will draw reassurance from the data in this study, which shows that patients with greater burden of illness received more medical attention than those with fewer problems. For example, patients with both types of comorbid conditions averaged 4.6 outpatient primary care visits and 2.0 specialist visits per year, compared with 3.0 and 0.1, respectively, for patients with neither. The proportion of patients whose hypertension was controlled at the index visit of the study period was 57% in those without either type of comorbid conditions, versus 63% in those with discordant comorbid conditions only, 65% in those with concordant comorbid conditions, and 69% in those with both. Among patients who did not have controlled blood pressure at their index visits, those with more comorbid conditions were more likely to get appropriate follow-up during the next 6 months. Multivariate analyses indicated that patients with discordant comorbid conditions received better care for their hypertension than those who had no comorbid conditions, and concordant comorbid conditions were associated with even better hypertension care. The best hypertension care of all was received by patients who had both types of comorbid conditions.
This analysis seems plausible because it is consistent with the interpretation that physicians were taking care of their patients, not just their patients’ blood pressures. In general, clinicians gave more attention to sicker patients, and high blood pressure was not ignored during visits for noncardiovascular problems. Hypertension is especially important in patients with other conditions that convey a high risk for cardiovascular problems, and the physicians in this study did a better job with hypertension in these populations.
This study does not truly answer the question posed in its title ("Will hypertension performance measures used for pay-for-performance programs penalize those who care for medically complex patients?"). After all, it was not a test of a pay-for-performance program for hypertension, and the patients were in a prepaid healthcare system without disincentives for patients to seek care. Virtually all of their care was delivered by salaried physicians who used electronic records and other systems to promote good quality for chronic diseases. The physicians were not actually being paid incentives based on their blood pressure control success, but, at least in this setting, sicker patients got more and better care for their hypertension.
These findings conflict with some data that have aroused concern that preventive and chronic condition care might suffer in patients with more comorbid conditions. For example, Turner et al found that patients with more unrelated conditions were less likely to have uncontrolled hypertension addressed at visits in 6 primary care practices affiliated with an academic medical center,3 and other researchers have found that patients with chronic diseases are less likely to receive treatment for unrelated disorders.4 But other data suggest that quality of care seems to improve as the number of medical conditions increases, perhaps because the number of opportunities for physicians to address clinical issues is higher in sicker patients.5
One reasonable conclusion from these conflicting data is that the outcome is in doubt: A greater number of comorbid conditions could hurt or could help quality of care, and it is up to healthcare leaders to protect against the former and increase the likelihood of the latter. An important first step is to support and strengthen the primary care system. Recent data indicate that the strength of the connection between primary care physicians and patients correlates with the likelihood that patients will receive guideline-consistent care.6
A second conclusion is that quality improvement for hypertension and other common chronic diseases is likely to be enhanced considerably when organized clinicians use systems like electronic medical records. In the Veterans Affairs system and other provider organizations that use electronic records, all clinicians can see all of the blood pressure readings and all of the office notes, making it easier for primary care physicians and specialists to intervene when hypertension is poorly controlled. Electronic medical records are used by a minority of US physicians outside of this organized delivery system, and, among new users, clinicians accustomed to dictating notes have to learn new habits such as entering blood pressures into the appropriate data field. But as electronic records become more widely used and as clinicians learn to use them effectively, we will have new opportunities to make real inroads in the historically poor rates of hypertension control.
The third conclusion is that, as pay for performance becomes more common as an alternative to traditional fee-for-service payment methods, it will be important to evaluate its effects carefully. These observational data from the Veterans Affairs system indicate that a greater number of comorbid conditions does not have to compromise the quality of hypertension care, but when the laboratory is changed and new variables are introduced, the findings might be different. This call for vigilance is consistent with the principles and recommendations of the Reimbursement, Coverage, and Access Policy Development Workgroup of the American Heart Association.7
That said, physicians should not shy away from payment systems that introduce accountability for clinical outcomes. The alternatives could be worse, and physicians have the chance to use this period of openness to new payment models to change health care for the better. We should not fight off pay for performance. We should make it our own by adopting it, studying it, correcting its faults, and constantly seeking its improvement.
 |
Acknowledgments |
|---|
Disclosures
None.
|
Footnotes |
|---|
The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.
|
References |
|---|
 Top References |
|---|
1. Lindenauer PK, Remus D, Roman S, Rothberg MB, Benjamin EM, Ma A, Bratzler DW. Public reporting and pay for performance in hospital quality improvement. N Engl J Med. 2007; 356: 486–496.
2. Petersen LA, Woodard LD, Henderson LM, Urech TH, Pietz K. Will hypertension performance measures used for pay -for-performance programs penalize those who care for medically complex patients? Circulation. 2009; 119: 2978–2985.
3. Turner BJ, Hellenbeak CS, Weiner M, Have TT, Tang SS. Effect of unrelated comorbid conditions on hypertension management. Ann Intern Med. 2008; 148: 578–586.
4. Redelmeier DA, Tan SH, Booth GL. The treatment of unrelated disorders in patients with chronic medical disease. N Engl J Med. 1998; 338: 1516–1520.
5. Higashi T, Winger NS, Adams JL, Fung C, Roland M, McGlynn EA, Reeves D, Asch SM, Kerr EA, Shekelle PG. Relationship between number of medical conditions and quality of care. N Engl J Med. 2007; 356: 2496–2504.
6. Atlas SJ, Grant RW, Ferris TG, Chang Y, Barry MJ. Patient-physician connectedness and quality of primary care. Ann Intern Med. 2009; 150: 325–335.
7. Bufalino V, Peterson ED, Burke GL, LaBresh KA, Jones DW, Faxon DP, Valadez AM, Brass LM, Fulwider VB, Smith R, Krumholz HM, Schwartz JS; American Heart Association’s Reimbursement, coverage, and Access Policy Development Workgroup. Payment for quality: guiding principles and recommendations: principles and recommendations from the American Heart Association’s Reimbursement, Coverage, and Access Policy Development Workgroup. Circulation. 2006; 113: 1151–1154.
Related Article:
Circulation 2009 119: 2978-2985.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||