- Email: [email protected]
Qualitative Insights Into How Pediatric Pay-for-Performance Programs Are Being Designed
Qualitative Insights Into How Pediatric Pay-forPerformance Programs Are Being Designed Alyna T. Chien, MD, MS; Matthew W. Colman, MD; Lainie Friedman Ross, MD, PhD pediatric performance measures and strategies for generally dealing with small sample sizes, and targeting the quality of care delivered to children with chronic health conditions, but they generally view these efforts to be effective.
Objective.—Pediatric pay-for-performance (P4P) programs are proliferating, and little is known about them. The goal of this study is to better understand how these programs began, and how they are designed and implemented from the perspectives of those with experience running pediatric P4P programs. Methods.—Cross-sectional semistructured interviews with named program directors and key supporting staff from 11 of 17 publicly described pediatric P4P programs that have been in operation for at least 1 year (commercial and Medicaid sponsored) regarding their program’s beginning, design, top challenges and recommendations, impact, and considerations regarding the needs of children with chronic health conditions. Results.—Eleven programs have allocated approximately $221 million toward pediatric P4P efforts by means of both bonus and penalty incentives, potentially affecting 4.3 million children. They struggle with involving pediatricians, desiring more vetted
Conclusions.—Those with experience running these early pediatric P4P programs show that pediatricians have not necessarily been involved in program design, face basic uncertainties of P4P program design, and generally do not target the care provided to children with chronic health conditions. They desire greater input from physicians who care for children and vetted pediatric measures, and they need help facing methodological challenges, such as small sample size and risk adjustment. KEY WORDS: pay for performance; pediatrics; quality Academic Pediatrics 2009;9:185–91
A
lthough much of the United States and United Kingdom is focused on adult pay-for-performance (P4P) programs,1,2 less attention has been paid to the 33 of nearly 100 P4P programs listed in The Leapfrog Compendium—the largest publicly available listing of American P4P programs—that use at least one performance measure applicable to children. Further, almost half of state Medicaid programs—the most important public payer of children’s health care—use P4P strategies, and 85% expect to do so in the next 5 years.3–5 Pediatric P4P programs are proliferating, but there is little information regarding how these programs are being designed or the challenges they face once implemented. Our knowledge of pediatric P4P is limited. The broader literature on the effectiveness of P4P in health care comprises little over 20 studies with control or comparison groups6,7; 5 of these address aspects of pediatric health care.8–13 Given the lack of information about P4P generally and with respect to pediatric P4P specifically, we hypothesize that pediatric P4P programs are in their infancy and that those experimenting with this incentive strategy may be contemplating,
designing, implementing, and evaluating programs in ways that may have common elements or themes. Insight into current programs may improve the design and implementation of future pediatric P4P efforts. P4P program directors represent an important type of informant regarding their programs.14 To provide insight into existing pediatric P4P efforts, we designed a qualitative study that interviewed program directors and key supporting staff about their pediatric P4P programs. We used an open-ended framework to ask program directors to describe the following: 1) how their programs began, 2) how their programs were designed, 3) their challenges and recommendations for future efforts, and 4) their perceptions of their program’s effectiveness. Because the broader P4P literature highlights a concern regarding the potential of P4P to yield negative unintended consequences for those with more severe or complex disease,15–18 we secondarily asked our informants how existing efforts address the differential health care needs of the average healthy child versus those of children with chronic health conditions.
From the Division of General Pediatrics, Children’s Hospital of Boston, Harvard Medical School, Boston, Mass (Dr Chien and Dr Ross); the Residency Program in Orthopaedic Surgery, University of Pittsburg Medical Center, Pittsburg, Pa (Dr Colman); and the MacLean Center for Clinical Medical Ethics, The University of Chicago, Chicago, Ill (Dr Ross). Address correspondence to Alyna T. Chien, MD, MS, 21 Autumn Street, Room 222, Boston, Massachusetts 02115 (e-mail: alyna.chien@ childrens.harvard.edu). Received for publication September 1, 2008; accepted January 24, 2009.
METHODS
ACADEMIC PEDIATRICS Copyright Ó 2009 by Academic Pediatric Association
Definition We defined a pediatric P4P program as one that explicitly linked financial or reputation-based rewards and/or sanctions to performance on measures of specific health care processes and/or outcomes,6,7,14 and that used at least 2 performance measures that specifically addressed the provision of pediatric health care (eg, hyperbilirubinemia check before
185
Volume 9, Number 3 May–June 2009
186
Chien et al
newborn discharge) or pertained to a health condition common in children and adolescents (eg, prescription of controller medications for persistent asthma). Semistructured Interview Sample Frame In August 2006, we compiled a purposive listing of pediatric P4P programs using the 2 main publicly available listings of P4P programs: the Leapfrog Compendium, a systematically updated listing posted by the not-for-profit quality improvement group, The Leapfrog Group, and the Centers for Medicare and Medicaid Service’s most up-todate survey of state Medicaid programs regarding their P4P efforts at that time.19,20 This assured that we included programs being sponsored by the 2 main payers of pediatric health care, commercial health plans and state Medicaid programs. These sources provided basic information about programs (eg, duration, performance measures being targeted) so that we could focus on those with child-relevant programs and at least 1 year of program experience (ie, launched by November 2005). We excluded programs not meeting our definitional and inclusion criteria, missing an identifiable program director, or with incorrect contact information and duplicates. Our final interview sample consisted of key informants (program directors, assistant program directors, and/or medical directors) from 17 pediatric P4P programs (9 sponsored by commercial health plans, 8 by state Medicaid programs). Recruitment From October 2006 through February 2007, we invited program directors from these 17 programs to participate in a 30-minute semistructured telephone interview by contacting our informants by e-mail and telephone (3 rounds each). At the time of recruitment, we confirmed that program directors understood that the purpose of the interview was to discuss the pediatric aspects of their P4P programs, and we reiterated that they could invite supporting personnel with more specific knowledge of the pediatric aspects of their program (usually the medical director, assistant program directors, and others) to join in the interview. The team of supporting personnel identified by the named program director, along with the program director, comprised our key informants for each program. We obtained University of Chicago Institutional Review Board approval for this study and oral consent from all participating informants. Interview Content We used the publicly available descriptions of these P4P programs and the broader literature on P4P in health care to develop this study’s conceptual model (Figure) and respective interview tool with 5 domains, described below. Program Initiation We queried informants on how their program began (eg, ‘‘How did your organization begin thinking about using P4P in terms of your pediatric population?’’). We included a probe regarding whether and/or how it involved pediatricians if this information was not spontaneously offered.
ACADEMIC PEDIATRICS
Program Design We asked informants to verify the publicly available information we had gathered before the interview: program age, types of incentives being used, pediatric metrics being incentivized, number of children potentially affected by the targeted metric, and the amount of money put into bonuses or rewards for the P4P program. Program Implementation We inquired after the top 3 challenges their programs faced, and we followed with questions about their recommendations for further designing these programs. Program Impact We asked our informants how they thought their programs were working and how they were assessing their program’s impact. We probed regarding the level of rigor being applied to these evaluations (ie, whether studies were informal pre/post evaluations, or more formal with quasi- and/or experimental design) if this information was not spontaneously offered. Issues Regarding Children With Chronic Health Conditions For 3 domains—program design, implementation, and impact—we asked about how these programs addressed the differential health care needs of the average healthy child versus those with chronic conditions that are: common (eg, asthma, attention-deficit/hyperactivity disorder), emerging (eg, obesity, type II diabetes), costly (eg, leukemia, congenital heart disease), or technology dependent (eg, children with special health care needs). For the purposes of this article, we refer to all of the above types of differentially diagnosed pediatric patients as ‘‘children with chronic health conditions.’’21 Qualitative Analysis Like other studies of this nature, we treated the interview as the unit of analysis because the goal of the study was to understand the pediatric P4P programs from the perspective of key administrative personnel, not individual perspectives.14 Two interviewers (A.T.C. and M.W.C.) conducted all semistructured interviews and independently transcribed handwritten notes. Shorter responses were recorded verbatim; long responses were paraphrased. We then compared our notes for precision and completeness, reconciled differences, and transcribed the documents into a single deidentified transcript. Two readers (A.T.C. and L.F.R.) subsequently coded the interviews by using the constant comparative method of qualitative analysis.22 This involved independently reading deidentified interview transcripts, characterizing the topics and themes raised, and creating and comparing coding sheets to verify similarities and to reconcile differences within each domain and across all interviews. Where necessary, we defined and used specific coding terms, and we used additional cycles of independent coding and reconciliation to characterize responses. We present our
ACADEMIC PEDIATRICS
Insights Into Pediatric Pay-for-Performance Programs
1. Initiation
How did your organization begin thinking about using performance incentives to target pediatric healthcare?
2. Design
3. Implementation
187
4. Impact
Program characteristics: Perceived effectiveness?
• Program age? • Number of children targeted? • Incentive type
(bonus, penalty)
Challenges? Recommendations?
Data upon which perception is based?
• Performance metric(s) being used?
5. Issues pertaining to the average healthy child versus “children with chronic health conditions”?
Figure. Conceptual model.
results according to the themes we identified within the 5 domains of our conceptual model.
RESULTS Respondents Key informants from 11 of 17 targeted programs (5 commercial, 6 state Medicaid) participated in the semistructured interviews; key administrative personnel from 6 (3 commercial, 3 Medicaid) did not. Key informants always included the named program director, but also included supplemental personnel (medical directors, assistant program directors, senior vice presidents and vice presidents of provider services, and statisticians) for 7 of 11 targeted programs (3 commercial, 4 state Medicaid). All but one set of key informants provided supplemental written information after the interview (eg, the total number of children included in their program, details regarding incentivized measures). Respondent and nonrespondent programs did not differ by type of sponsor, number of years in place, or program design. Program Initiation The effort expended in the development of pediatric P4P programs appeared to differ depending on whether programs were sponsored by commercial health plans or Medicaid ones. Key informants of programs sponsored by commercial health plans uniformly described focusing on the adult programs and thinking only incidentally about the pediatric aspects of their programs. ‘‘The pediatric population was folded into the program in a general way. We did not set out to target pediatric providers and their patients specifically.’’ —Program director of Commercial Health Plan 1
With one exception, key informants of programs originating from state Medicaid offices provided more detailed descriptions of the process surrounding the design of the pediatric component of their programs. They described choosing measures that they thought matched the needs of the pediatric population and gaining buy-in from various stakeholders, including health plans, county departments, advocates for child health, and state associations. ‘‘We certainly thought specifically about the pediatric population since that is what comprises the majority of [this state’s] Medicaid population. We tried to choose metrics which were appropriate to pediatrics and made sense to us, the managed care organizations, and key stakeholders.’’ —Program director of State Medicaid Program 2 Key informants spontaneously reported that pediatricians played little to no role in developing their P4P programs. Informants from 4 programs (3 commercial, 1 Medicaid) described their initial discussions with pediatricians as antagonistic. Three of these 4 programs were able to continue to develop working relationships with pediatricians. ‘‘We considered [pediatricians] dry holes..[In 2001] we met indifference at best and hostility at worst, then we spent months looking at pediatric measures but only found one.’’ —Program director of Commercial Health Plan Program 4 ‘‘[More recently] we have had universal buy-in from our provider network..This was quite a transition from [their being] hostile—they were quite literally throwing stones at us about 7 years ago–although they whispered to us that we were doing the right thing. Now they are cooperative in our meetings.’’ —Program director of Commercial Health Plan Program 5
188
Chien et al
Key informants from the remaining 7 programs, however, indicated that pediatricians were not at all involved in the development of their programs. Program Design Our informants verified that their programs had been in place from between 1 to 8 years and that, in total, these 11 programs had targeted care provided to 4.2 million children and adolescents (2.3 million in commercial-sponsored programs; 2 million in state Medicaid programs). All programs used financial incentives, 8 in the form of cash bonuses, and 3 used cash penalties. Many programs simultaneously employed reputation-based incentives in the form of report cards; no program relied solely on reputation-based incentives. Except for the use of penalty incentives among state Medicaid plans, there was no obvious difference in the use of financial versus reputation-based incentives when comparing commercial health plans with state Medicaid ones. Generally, commercial health plans directed their incentives at medical groups; state Medicaid programs targeted health plans. Among the 11 programs, 8 reported investing $221 million into funds for cash bonuses (5 commercial, 3 state Medicaid). Three state Medicaid pro-grams did not have bonus money and reconfigured their auto-assignment algorithms such that programs with better performance would gain more market share and/or designed ‘‘recoupment’’ strategies in which, for example, health plans return money at the end of year or accept lower reimbursement rates for relatively poor performance. Programs implemented pediatric-relevant measures in 5 main domains of pediatric care: 1) primary prevention (ie, care that aims to prevent diseases from occurring in the first place), 2) secondary prevention (ie, care that is designed to identify and/or treat asymptomatic persons at risk for a particular condition), 3) recommended utilization (ie, the number or timing of health care visits, not the content of the care provided), 4) prescribing practices (ie, the overuse or underuse of medications), and 5) chronic care for asthma (eg, controller medications) and mental health conditions (eg, posthospitalization follow-up). Although both commercially sponsored and state Medicaid programs incentivized up-todate immunization status and aspects of asthma care, state Medicaid programs additionally incentivized the provision of recommended utilization and secondary prevention. These findings are presented in the Table. Program Implementation When queried about the top 3 challenges faced while implementing their pediatric P4P programs, our informants cited the paucity of vetted and/or endorsed pediatric metrics, uncertainties about basic program design (ie, what to measure, how much to pay), and lack of pediatrician participation. The key informants of 8 programs (3 commercial, 5 Medicaid) cited the limited availability of pediatric performance measures as one of the top 3 challenges facing their program. Most emphasized their desire for the metrics to be nationally vetted by the National Quality Forum, the Healthcare Effectiveness Data and Information Set, and the Joint Commission on Accreditation of Healthcare
ACADEMIC PEDIATRICS
Organizations. A handful also specified that these measures should be evidence based and/or reflective of conditions prevalent in pediatrics. The key informants of 5 programs (2 commercial, 3 Medicaid) described significant difficulties navigating the uncertainties surrounding basic incentive program design. They wanted clarification on whether to use financial- or reputation-based incentives, how large to make bonuses, whether to implement penalty incentives, and how high to set the bar for performance targets (ie, low or easily achievable, or high or difficult to achieve). The key informants of 5 programs (3 commercial, 2 Medicaid) also declared the lack of pediatrician involvement to be a barrier to designing and/or implementing their P4P programs. None of the key informants spontaneously cited a concern regarding potential negative unintended consequences resulting from P4P programs or the need to better understand the role of risk adjustment in P4P schemes, but when prompted, several recognized the need to deal with inherent sample size limitations. For recommendations, our informants generally suggested that anyone interested in developing a pediatric P4P program work with vetted and/or endorsed metrics, and they emphasized the importance of gaining pediatrician buy-in. Two sets of informants suggested other tactics: carrots should be used instead of sticks; a health plan’s membership should be made very large or the sample size for individual measures will be too small; a health plan’s administration functions should run excellently and efficiently; and measures ought to be based on administrative records because otherwise the cost of running the P4P program will be too high. Program Impact Our informants generally perceived their respective pediatric P4P programs to be effective, and they based their perceptions on tracking incentivized components of care over time. The key informants of 9 of the programs (4 commercial, 5 Medicaid) noted improvements in incentivized components of care; 2 of these stated that the changes were statistically significant. Key informants of programs originating from state Medicaid offices had difficulty ascribing improvement solely to the P4P program given the number of concomitant strategies their offices had undertaken to improve the quality of care (eg, disease or case management programs). ‘‘Each year we have seen significant improvement across all of our metrics. On one hand it is tough to say whether this is the result of simply better data collection, or whether it is an actual improvement in quality. However, we have done extensive interviews for feedback with physicians who appear to have gained a real increase in awareness and energy for quality improvement which was not necessarily there before, so I think the effects are real.’’ —Program director of Commercial Health Plan 3 The key informants of the remaining 2 programs were unsure, or they stated that the program demonstrated little or no improvement.
Type of Pediatric Metric† Primary Prevention
Incentive Type
Payor
State Medicaid #4 State Medicaid #5 State Medicaid #6
Program Age, y
Cash Cash Report Bonus Penalty Card
Immunization Status
Recommended Utilization
WellPrenatal Well-Child Adolescent Newborn Child Adolescent Care Visit Visit
Prescribing Practices
Chronic Care
Use of Antibiotics or Generic Medication
Asthma Mental Health Care Follow-up
Medical groups
4
40 000
Medical groups
6
230 000
Medical groups
3
1 500 000
Medical groups - Patients - Hospitals
1
340 000
7
150 000
-
-
-
-
Health plans
1
170 000
-
Health plans
7
300 000
Medical groups - Case managers - Health plans
1
190 000
1
990 000
-
Health plans
8
80 000
-
Health plans
5
220 000
-
-
*A dot ¼ present; blank ¼ absent. †Example metrics are as follows: Immunization Status indicates percentage of patients with up-to-date immunization status by 15 months of age; Newborn, percentage of newborns screened for hyperbilirubinemia before hospital discharge; Child, percentage of children screened for high blood lead levels; Adolescent, percentage of adolescents aged $15 years screened for chlamydia infection; Prenatal Care, percentage of women who start prenatal care in the first trimester; Well-Child Visit, percentage of children who have 6 well-child visits by 15 months of age; Well-Adolescent Visit, percentage of well-adolescents with at least one visit in 3 years; Use of Antibiotics or Generic Medication, percentage of patients not provided antibiotics for upper respiratory infections; Asthma Care, percentage of children with persistent asthma who were prescribed controller medications; and Mental Health Follow-up, percentage of patients with psychiatric hospitalization with outpatient follow-up within 1 week.
Insights Into Pediatric Pay-for-Performance Programs
Commercial health plan #1 Commercial health plan #2 Commercial health plan #3 Commercial health plan #4 Commercial health plan #5 State Medicaid #1 State Medicaid #2 State Medicaid #3
Payee
No. of Children in Program
Secondary Prevention
ACADEMIC PEDIATRICS
Table 1. Basic Pediatric Pay-for-Performance Program Characteristics and Incentivized Aspects of Pediatric Care*
189
190
Chien et al
When asked about more formal research techniques (ie, experimental or quasi-experimental study designs) to evaluate their programs, our informants denied having access to personnel or academic partners with whom they could partner. Issues Related to ‘‘Children With Chronic Health Conditions’’ Asthma and mental health conditions were the 2 most commonly incentivized chronic health conditions. None of our informants thought that their P4P programs could be used to tackle emerging health conditions because of the lack of availability of corresponding performance metrics. Informants of 3 programs questioned whether P4P should even be used as a strategy to improve care for children with chronic health conditions because of the great variety of conditions affecting these children and the correspondingly low volume of children with each condition. ‘‘These conditions are too low volume and it is hard to get a solid diagnosis for many of these conditions.’’ —Leader of Commercial Health Plan 2 Informants considered the care for children with mental health conditions to be beyond the reach of the P4P program for several reasons. First, the funds for these conditions came from insurance plan carve-outs—that is, specific sums of money made available to children with these diagnoses up to a certain annual or lifetime limit. Second, they reiterated the problem of small sample size. Third, leaders of 3 programs (all state Medicaid) explained that P4P may not be a necessary incentive for these issues because existing case management systems would identify and coordinate care for these children irrespective of P4P. ‘‘There is a mechanism for helping these types of kids. [We have a tool that] provides an opportunity to ask relevant questions on patient health needs and status.. It is done at the point of enrollment and may alert HMOs and caseworkers to special and immediate interventions they may need to carry out.’’ —Leader of State Medicaid Program 6 Our informants’ advice for using P4P to improve health care for children with chronic health conditions focused on developing ways to aggregate measures across conditions, and to incentivize aspects of care that are not condition specific but that can be applied across disparate conditions (eg, maintenance of up-to-date problems and medication lists, care coordination, or disease management). ‘‘It is very tough to incentivize small numbers of kids for a given disease or condition because it is hard to measure if the prevalence is low.. If there were standardized measures which took into account multiple rare conditions from more of a case management standpoint, that might be a good idea.’’ —Program director of State Medicaid Program 2 ‘‘We are doing poorly in these categories. We need to look at incentivizing coordination of care, additional
ACADEMIC PEDIATRICS
supports for these children, and incentivizing some more complex therapies for their conditions.’’ —Program director of Commercial Health Plan 1 DISCUSSION This study shows that early pediatric P4P programs are underway and are already affecting the care provided to our nation’s children, even though pediatricians have not routinely engaged in the design of these programs. It also shows that those designing and implementing these programs are grappling with basic program design problems, and they are paying less attention to subtle, albeit crucial, challenges, such as the potential for negative unintended consequences, risk adjustment, and overcoming inherent sample size limitations. Informants generally view their programs as effective, suggesting that their P4P efforts will probably continue. Our study is limited by our participant sample. Although our response rate matches what is described in the literature,23,24 informants who did not respond to our survey may have answered our questions differently. Our study only includes named program directors and their identified supporting staff, so their responses may differ from that of others with different roles in the P4P program. However, the aim of our study was to solicit and analyze the perspectives of those who currently have decision-making capacity or who hold leadership roles within these programs, not of those in rankand-file positions. Our study could also be limited by social desirability bias. The direction of this bias, if significant, is unclear. We made every effort to mitigate these effects by phrasing questions in a neutral fashion and by asking for supporting data where appropriate and available. This study provides a unique opportunity to compare the features of early pediatric P4P programs to the qualities endorsed by researchers, professional societies, and regulatory agencies.25–28 This literature emphasizes the importance of providing reward rather than penalty incentives,29 including physicians in program design, and selecting measures that reflect evidence-based clinical processes that health care organizations and/or their physicians’ control. It is of concern that in contrast to the cautious approach advocated above, pediatric P4P programs have begun instituting penalties, have been unable to engage pediatricians, and have introduced measures that may not necessarily reflect evidence-based clinical processes controlled by health care organizations or their physicians. It appears that future pediatric P4P efforts need to understand the limitations of our current P4P knowledge, and that researchers and physicians who care for children need to help build the empirical basis for rational pediatric P4P program design. Pediatric P4P is proliferating. Insights from those with experience designing and implementing these programs suggest that there is considerable room to improve our nation’s pediatric P4P efforts. ACKNOWLEDGMENTS We thank Charlie Homer and Stephen Muething for their input when this study was being conceived.
ACADEMIC PEDIATRICS
Insights Into Pediatric Pay-for-Performance Programs
This study was internally funded by the Department of Pediatrics, The University of Chicago, while Dr Chien was a member of the faculty there.
11. Fairbrother G, Siegel MJ, Friedman S, et al. Impact of financial incentives on documented immunization rates in the inner city: results of a randomized controlled trial. Ambul Pediatr. 2001;1:206–212. 12. Felt-Lisk S, Gimm G, Peterson S. Making pay-for-performance work in Medicaid. Health Aff. 2007;26:w516–w527. 13. Mandel KE, Kotagal UR. Pay for performance alone cannot drive quality. Arch Pediatr Adolesc Med. 2007;161:650–655. 14. Chien AT, Chin MH, Davis AM, Casalino LP. Pay-for-performance, public reporting and racial disparities in health care: how are programs being designed? Med Care Res Rev. 2007;64:285S–304S. 15. Glied S. Getting the incentives right for children. Health Serv Res. 1998;33(4 pt 2):1143–1160. 16. Freed GL, Uren RL. Pay-for-performance: an overview for pediatrics. J Pediatr. 2006;149:120–124. 17. Profit J, Zupancic JAF, Gould JB, Petersen LA. Implementing payfor-performance in the neonatal intensive care unit. Pediatrics. 2007;119:975–982. 18. Chien AT, Dudley RA. Pay-for-performance in pediatrics: proceed with caution. Pediatrics. 2007;120:186–188. 19. The Leapfrog Group. The Leapfrog group fact sheet. Available at: http://www.leapfroggroup.org/leapfrog-factsheet. November 2005. Accessed February 11, 2009. 20. Center for Health Care Strategies Inc. Centers for Medicare and Medicaid Services. Descriptions of selective performance incentives programs. Available at: http://www.cms.hhs.gov/smdl/downloads/ StatePerformanceIncentiveChart040606.pdf. Accessed February 11, 2009. 21. van der Lee JH, Mokkink LB, Grootenhuis MA, et al. Definitions and measurement of chronic health conditions in childhood: a systematic review. JAMA. 2007;297:2741–2751. 22. Miles M, Huberman A. Qualitative Data Analysis: An Expanded Sourcebook. 2nd ed. Thousand Oaks, Calif: Sage Publications; 1994. 23. Chien AT. The Potential Impact of Performance Incentive Programs on Racial Disparities in Health Care. Totowa, NJ: Humana Press; 2007. 24. Scanlon DP. Evidence for pay for performance: hope for the US health care system? Manag Care.. 2005;14(12 suppl):6–10. 25. American College of Physicians. Linking Physician Payments to Quality Care. Philadelphia, Pa: American College of Physicians; 2005. Position paper. Available at: http://www.acponline.org/ advocacy/where_we_stand/policy/link_pay.pdf. Accessed February 11, 2009. 26. American Medical Association. Guidelines for pay-for-performance programs. Available at: http://www.ama-assn.org/ama1/pub/upload/ mm/368/guidelines4pay62705.pdf. Accessed February 11, 2009. 27. American Medical Association. Principles for pay-for-performance programs. http://www.ama-assn.org/ama1/pub/upload/mm/368/princi ples4pay62705.pdf. Accessed February 11, 2009. 28. Van Amringe M. The JCAHO guidelines for pay for performance. Manag Care. 2005;14(12 Suppl):11–12. 29. Shen Y. Selection incentives in a performance-based contracting system. Health Serv Res. 2003;38:535–552.
REFERENCES 1. Centers for Medicare and Medicaid Services. Medicare pay-forperformance demonstration shows significant quality of care improvement at participating hospitals. May 3, 2005. Available at: http://www.cms.hhs.gov/apps/media/press/release. asp?Counter¼1441&intNumP erPage¼10&checkDate¼&checkKey¼2&srchType¼2&numDays¼0&srchOpt¼0&sr chData¼part þd&keywordType¼All&chkNewsType¼1%2Cþ2%2Cþ3%2Cþ4 %2Cþ 5&intPage¼&showAll¼1&pYear¼&year¼0&desc¼false &cboOrder¼date. Accessed February 11, 2009. 2. Centers for Medicaid and Medicare Services. Pay for performance initiatives. Herb Kuhn testimony for the Senate Committee on Finance. June 27, 2005. Available at: http://www.cms.hhs.gov/apps/ media/press/testimony.asp?Counter¼1537. Accessed February 11, 2009. 3. The Leapfrog Group. The Leapfrog compendium. Available at: http:// www.leapfroggroup.org/compendium2. Accessed February 11, 2009. 4. Kurhmerker K, Hartman T. Pay-for-performance in State Medicaid Programs: a survey of state Medicaid directors and programs. April 12, 2007. Available at: http://www.commonwealthfund.org/ publications/publications_show.htm?doc_id¼472891. 2007. Accessed February 11, 2009. 5. Kurhmerker K, Hartman T. Pay-for-performance in State Medicaid Programs. Appendix B: State pay-for-performance program summaries. Commonwealth Fund Publication 1018. Available at: http:// www.commonwealthfund.org/usr_doc/Kuhmerker_P4Pstate MedicaidprogsappendixB.pdf?section¼4039. Accessed February 11, 2009. 6. Dudley R, Frolich A, Robinowitz D, et al. Strategies to Support Quality-Based Purchasing: A Review of the Evidence. Summary, Technical Review 10. Prepared by the Stanford–University of California San Francisco Evidence-Based Practice Center under contract 290-02-0017. AHRQ Publication 04-0057. Rockville, Md: Agency for Healthcare Research and Quality; July 2004. 7. Petersen LA, Woodard LD, Urech T, et al. Does pay-for-performance improve the quality of health care? Ann Intern Med. 2006;145: 265–272. 8. Chien AT, Conti RM, Pollack HA. A pediatric-focused review of performance incentive literature. Curr Opin Pediatr. 2007;19: 719–725. 9. Hillman AL, Ripley K, Goldfarb N, et al. The use of physician financial incentives and feedback to improve pediatric preventive care in Medicaid managed care. Pediatrics. 1999;104(4 pt 1):931–935. 10. Fairbrother G, Hanson KL, Friedman S, Butts GC. The impact of physician bonuses, enhanced fees, and feedback on childhood immunization coverage rates. Am J Public Health. 1999;89:171–175.
191
Objective.—Pediatric pay-for-performance (P4P) programs are proliferating, and little is known about them. The goal of this study is to better understand how these programs began, and how they are designed and implemented from the perspectives of those with experience running pediatric P4P programs. Methods.—Cross-sectional semistructured interviews with named program directors and key supporting staff from 11 of 17 publicly described pediatric P4P programs that have been in operation for at least 1 year (commercial and Medicaid sponsored) regarding their program’s beginning, design, top challenges and recommendations, impact, and considerations regarding the needs of children with chronic health conditions. Results.—Eleven programs have allocated approximately $221 million toward pediatric P4P efforts by means of both bonus and penalty incentives, potentially affecting 4.3 million children. They struggle with involving pediatricians, desiring more vetted
Conclusions.—Those with experience running these early pediatric P4P programs show that pediatricians have not necessarily been involved in program design, face basic uncertainties of P4P program design, and generally do not target the care provided to children with chronic health conditions. They desire greater input from physicians who care for children and vetted pediatric measures, and they need help facing methodological challenges, such as small sample size and risk adjustment. KEY WORDS: pay for performance; pediatrics; quality Academic Pediatrics 2009;9:185–91
A
lthough much of the United States and United Kingdom is focused on adult pay-for-performance (P4P) programs,1,2 less attention has been paid to the 33 of nearly 100 P4P programs listed in The Leapfrog Compendium—the largest publicly available listing of American P4P programs—that use at least one performance measure applicable to children. Further, almost half of state Medicaid programs—the most important public payer of children’s health care—use P4P strategies, and 85% expect to do so in the next 5 years.3–5 Pediatric P4P programs are proliferating, but there is little information regarding how these programs are being designed or the challenges they face once implemented. Our knowledge of pediatric P4P is limited. The broader literature on the effectiveness of P4P in health care comprises little over 20 studies with control or comparison groups6,7; 5 of these address aspects of pediatric health care.8–13 Given the lack of information about P4P generally and with respect to pediatric P4P specifically, we hypothesize that pediatric P4P programs are in their infancy and that those experimenting with this incentive strategy may be contemplating,
designing, implementing, and evaluating programs in ways that may have common elements or themes. Insight into current programs may improve the design and implementation of future pediatric P4P efforts. P4P program directors represent an important type of informant regarding their programs.14 To provide insight into existing pediatric P4P efforts, we designed a qualitative study that interviewed program directors and key supporting staff about their pediatric P4P programs. We used an open-ended framework to ask program directors to describe the following: 1) how their programs began, 2) how their programs were designed, 3) their challenges and recommendations for future efforts, and 4) their perceptions of their program’s effectiveness. Because the broader P4P literature highlights a concern regarding the potential of P4P to yield negative unintended consequences for those with more severe or complex disease,15–18 we secondarily asked our informants how existing efforts address the differential health care needs of the average healthy child versus those of children with chronic health conditions.
From the Division of General Pediatrics, Children’s Hospital of Boston, Harvard Medical School, Boston, Mass (Dr Chien and Dr Ross); the Residency Program in Orthopaedic Surgery, University of Pittsburg Medical Center, Pittsburg, Pa (Dr Colman); and the MacLean Center for Clinical Medical Ethics, The University of Chicago, Chicago, Ill (Dr Ross). Address correspondence to Alyna T. Chien, MD, MS, 21 Autumn Street, Room 222, Boston, Massachusetts 02115 (e-mail: alyna.chien@ childrens.harvard.edu). Received for publication September 1, 2008; accepted January 24, 2009.
METHODS
ACADEMIC PEDIATRICS Copyright Ó 2009 by Academic Pediatric Association
Definition We defined a pediatric P4P program as one that explicitly linked financial or reputation-based rewards and/or sanctions to performance on measures of specific health care processes and/or outcomes,6,7,14 and that used at least 2 performance measures that specifically addressed the provision of pediatric health care (eg, hyperbilirubinemia check before
185
Volume 9, Number 3 May–June 2009
186
Chien et al
newborn discharge) or pertained to a health condition common in children and adolescents (eg, prescription of controller medications for persistent asthma). Semistructured Interview Sample Frame In August 2006, we compiled a purposive listing of pediatric P4P programs using the 2 main publicly available listings of P4P programs: the Leapfrog Compendium, a systematically updated listing posted by the not-for-profit quality improvement group, The Leapfrog Group, and the Centers for Medicare and Medicaid Service’s most up-todate survey of state Medicaid programs regarding their P4P efforts at that time.19,20 This assured that we included programs being sponsored by the 2 main payers of pediatric health care, commercial health plans and state Medicaid programs. These sources provided basic information about programs (eg, duration, performance measures being targeted) so that we could focus on those with child-relevant programs and at least 1 year of program experience (ie, launched by November 2005). We excluded programs not meeting our definitional and inclusion criteria, missing an identifiable program director, or with incorrect contact information and duplicates. Our final interview sample consisted of key informants (program directors, assistant program directors, and/or medical directors) from 17 pediatric P4P programs (9 sponsored by commercial health plans, 8 by state Medicaid programs). Recruitment From October 2006 through February 2007, we invited program directors from these 17 programs to participate in a 30-minute semistructured telephone interview by contacting our informants by e-mail and telephone (3 rounds each). At the time of recruitment, we confirmed that program directors understood that the purpose of the interview was to discuss the pediatric aspects of their P4P programs, and we reiterated that they could invite supporting personnel with more specific knowledge of the pediatric aspects of their program (usually the medical director, assistant program directors, and others) to join in the interview. The team of supporting personnel identified by the named program director, along with the program director, comprised our key informants for each program. We obtained University of Chicago Institutional Review Board approval for this study and oral consent from all participating informants. Interview Content We used the publicly available descriptions of these P4P programs and the broader literature on P4P in health care to develop this study’s conceptual model (Figure) and respective interview tool with 5 domains, described below. Program Initiation We queried informants on how their program began (eg, ‘‘How did your organization begin thinking about using P4P in terms of your pediatric population?’’). We included a probe regarding whether and/or how it involved pediatricians if this information was not spontaneously offered.
ACADEMIC PEDIATRICS
Program Design We asked informants to verify the publicly available information we had gathered before the interview: program age, types of incentives being used, pediatric metrics being incentivized, number of children potentially affected by the targeted metric, and the amount of money put into bonuses or rewards for the P4P program. Program Implementation We inquired after the top 3 challenges their programs faced, and we followed with questions about their recommendations for further designing these programs. Program Impact We asked our informants how they thought their programs were working and how they were assessing their program’s impact. We probed regarding the level of rigor being applied to these evaluations (ie, whether studies were informal pre/post evaluations, or more formal with quasi- and/or experimental design) if this information was not spontaneously offered. Issues Regarding Children With Chronic Health Conditions For 3 domains—program design, implementation, and impact—we asked about how these programs addressed the differential health care needs of the average healthy child versus those with chronic conditions that are: common (eg, asthma, attention-deficit/hyperactivity disorder), emerging (eg, obesity, type II diabetes), costly (eg, leukemia, congenital heart disease), or technology dependent (eg, children with special health care needs). For the purposes of this article, we refer to all of the above types of differentially diagnosed pediatric patients as ‘‘children with chronic health conditions.’’21 Qualitative Analysis Like other studies of this nature, we treated the interview as the unit of analysis because the goal of the study was to understand the pediatric P4P programs from the perspective of key administrative personnel, not individual perspectives.14 Two interviewers (A.T.C. and M.W.C.) conducted all semistructured interviews and independently transcribed handwritten notes. Shorter responses were recorded verbatim; long responses were paraphrased. We then compared our notes for precision and completeness, reconciled differences, and transcribed the documents into a single deidentified transcript. Two readers (A.T.C. and L.F.R.) subsequently coded the interviews by using the constant comparative method of qualitative analysis.22 This involved independently reading deidentified interview transcripts, characterizing the topics and themes raised, and creating and comparing coding sheets to verify similarities and to reconcile differences within each domain and across all interviews. Where necessary, we defined and used specific coding terms, and we used additional cycles of independent coding and reconciliation to characterize responses. We present our
ACADEMIC PEDIATRICS
Insights Into Pediatric Pay-for-Performance Programs
1. Initiation
How did your organization begin thinking about using performance incentives to target pediatric healthcare?
2. Design
3. Implementation
187
4. Impact
Program characteristics: Perceived effectiveness?
• Program age? • Number of children targeted? • Incentive type
(bonus, penalty)
Challenges? Recommendations?
Data upon which perception is based?
• Performance metric(s) being used?
5. Issues pertaining to the average healthy child versus “children with chronic health conditions”?
Figure. Conceptual model.
results according to the themes we identified within the 5 domains of our conceptual model.
RESULTS Respondents Key informants from 11 of 17 targeted programs (5 commercial, 6 state Medicaid) participated in the semistructured interviews; key administrative personnel from 6 (3 commercial, 3 Medicaid) did not. Key informants always included the named program director, but also included supplemental personnel (medical directors, assistant program directors, senior vice presidents and vice presidents of provider services, and statisticians) for 7 of 11 targeted programs (3 commercial, 4 state Medicaid). All but one set of key informants provided supplemental written information after the interview (eg, the total number of children included in their program, details regarding incentivized measures). Respondent and nonrespondent programs did not differ by type of sponsor, number of years in place, or program design. Program Initiation The effort expended in the development of pediatric P4P programs appeared to differ depending on whether programs were sponsored by commercial health plans or Medicaid ones. Key informants of programs sponsored by commercial health plans uniformly described focusing on the adult programs and thinking only incidentally about the pediatric aspects of their programs. ‘‘The pediatric population was folded into the program in a general way. We did not set out to target pediatric providers and their patients specifically.’’ —Program director of Commercial Health Plan 1
With one exception, key informants of programs originating from state Medicaid offices provided more detailed descriptions of the process surrounding the design of the pediatric component of their programs. They described choosing measures that they thought matched the needs of the pediatric population and gaining buy-in from various stakeholders, including health plans, county departments, advocates for child health, and state associations. ‘‘We certainly thought specifically about the pediatric population since that is what comprises the majority of [this state’s] Medicaid population. We tried to choose metrics which were appropriate to pediatrics and made sense to us, the managed care organizations, and key stakeholders.’’ —Program director of State Medicaid Program 2 Key informants spontaneously reported that pediatricians played little to no role in developing their P4P programs. Informants from 4 programs (3 commercial, 1 Medicaid) described their initial discussions with pediatricians as antagonistic. Three of these 4 programs were able to continue to develop working relationships with pediatricians. ‘‘We considered [pediatricians] dry holes..[In 2001] we met indifference at best and hostility at worst, then we spent months looking at pediatric measures but only found one.’’ —Program director of Commercial Health Plan Program 4 ‘‘[More recently] we have had universal buy-in from our provider network..This was quite a transition from [their being] hostile—they were quite literally throwing stones at us about 7 years ago–although they whispered to us that we were doing the right thing. Now they are cooperative in our meetings.’’ —Program director of Commercial Health Plan Program 5
188
Chien et al
Key informants from the remaining 7 programs, however, indicated that pediatricians were not at all involved in the development of their programs. Program Design Our informants verified that their programs had been in place from between 1 to 8 years and that, in total, these 11 programs had targeted care provided to 4.2 million children and adolescents (2.3 million in commercial-sponsored programs; 2 million in state Medicaid programs). All programs used financial incentives, 8 in the form of cash bonuses, and 3 used cash penalties. Many programs simultaneously employed reputation-based incentives in the form of report cards; no program relied solely on reputation-based incentives. Except for the use of penalty incentives among state Medicaid plans, there was no obvious difference in the use of financial versus reputation-based incentives when comparing commercial health plans with state Medicaid ones. Generally, commercial health plans directed their incentives at medical groups; state Medicaid programs targeted health plans. Among the 11 programs, 8 reported investing $221 million into funds for cash bonuses (5 commercial, 3 state Medicaid). Three state Medicaid pro-grams did not have bonus money and reconfigured their auto-assignment algorithms such that programs with better performance would gain more market share and/or designed ‘‘recoupment’’ strategies in which, for example, health plans return money at the end of year or accept lower reimbursement rates for relatively poor performance. Programs implemented pediatric-relevant measures in 5 main domains of pediatric care: 1) primary prevention (ie, care that aims to prevent diseases from occurring in the first place), 2) secondary prevention (ie, care that is designed to identify and/or treat asymptomatic persons at risk for a particular condition), 3) recommended utilization (ie, the number or timing of health care visits, not the content of the care provided), 4) prescribing practices (ie, the overuse or underuse of medications), and 5) chronic care for asthma (eg, controller medications) and mental health conditions (eg, posthospitalization follow-up). Although both commercially sponsored and state Medicaid programs incentivized up-todate immunization status and aspects of asthma care, state Medicaid programs additionally incentivized the provision of recommended utilization and secondary prevention. These findings are presented in the Table. Program Implementation When queried about the top 3 challenges faced while implementing their pediatric P4P programs, our informants cited the paucity of vetted and/or endorsed pediatric metrics, uncertainties about basic program design (ie, what to measure, how much to pay), and lack of pediatrician participation. The key informants of 8 programs (3 commercial, 5 Medicaid) cited the limited availability of pediatric performance measures as one of the top 3 challenges facing their program. Most emphasized their desire for the metrics to be nationally vetted by the National Quality Forum, the Healthcare Effectiveness Data and Information Set, and the Joint Commission on Accreditation of Healthcare
ACADEMIC PEDIATRICS
Organizations. A handful also specified that these measures should be evidence based and/or reflective of conditions prevalent in pediatrics. The key informants of 5 programs (2 commercial, 3 Medicaid) described significant difficulties navigating the uncertainties surrounding basic incentive program design. They wanted clarification on whether to use financial- or reputation-based incentives, how large to make bonuses, whether to implement penalty incentives, and how high to set the bar for performance targets (ie, low or easily achievable, or high or difficult to achieve). The key informants of 5 programs (3 commercial, 2 Medicaid) also declared the lack of pediatrician involvement to be a barrier to designing and/or implementing their P4P programs. None of the key informants spontaneously cited a concern regarding potential negative unintended consequences resulting from P4P programs or the need to better understand the role of risk adjustment in P4P schemes, but when prompted, several recognized the need to deal with inherent sample size limitations. For recommendations, our informants generally suggested that anyone interested in developing a pediatric P4P program work with vetted and/or endorsed metrics, and they emphasized the importance of gaining pediatrician buy-in. Two sets of informants suggested other tactics: carrots should be used instead of sticks; a health plan’s membership should be made very large or the sample size for individual measures will be too small; a health plan’s administration functions should run excellently and efficiently; and measures ought to be based on administrative records because otherwise the cost of running the P4P program will be too high. Program Impact Our informants generally perceived their respective pediatric P4P programs to be effective, and they based their perceptions on tracking incentivized components of care over time. The key informants of 9 of the programs (4 commercial, 5 Medicaid) noted improvements in incentivized components of care; 2 of these stated that the changes were statistically significant. Key informants of programs originating from state Medicaid offices had difficulty ascribing improvement solely to the P4P program given the number of concomitant strategies their offices had undertaken to improve the quality of care (eg, disease or case management programs). ‘‘Each year we have seen significant improvement across all of our metrics. On one hand it is tough to say whether this is the result of simply better data collection, or whether it is an actual improvement in quality. However, we have done extensive interviews for feedback with physicians who appear to have gained a real increase in awareness and energy for quality improvement which was not necessarily there before, so I think the effects are real.’’ —Program director of Commercial Health Plan 3 The key informants of the remaining 2 programs were unsure, or they stated that the program demonstrated little or no improvement.
Type of Pediatric Metric† Primary Prevention
Incentive Type
Payor
State Medicaid #4 State Medicaid #5 State Medicaid #6
Program Age, y
Cash Cash Report Bonus Penalty Card
Immunization Status
Recommended Utilization
WellPrenatal Well-Child Adolescent Newborn Child Adolescent Care Visit Visit
Prescribing Practices
Chronic Care
Use of Antibiotics or Generic Medication
Asthma Mental Health Care Follow-up
Medical groups
4
40 000
Medical groups
6
230 000
Medical groups
3
1 500 000
Medical groups - Patients - Hospitals
1
340 000
7
150 000
-
-
-
-
Health plans
1
170 000
-
Health plans
7
300 000
Medical groups - Case managers - Health plans
1
190 000
1
990 000
-
Health plans
8
80 000
-
Health plans
5
220 000
-
-
*A dot ¼ present; blank ¼ absent. †Example metrics are as follows: Immunization Status indicates percentage of patients with up-to-date immunization status by 15 months of age; Newborn, percentage of newborns screened for hyperbilirubinemia before hospital discharge; Child, percentage of children screened for high blood lead levels; Adolescent, percentage of adolescents aged $15 years screened for chlamydia infection; Prenatal Care, percentage of women who start prenatal care in the first trimester; Well-Child Visit, percentage of children who have 6 well-child visits by 15 months of age; Well-Adolescent Visit, percentage of well-adolescents with at least one visit in 3 years; Use of Antibiotics or Generic Medication, percentage of patients not provided antibiotics for upper respiratory infections; Asthma Care, percentage of children with persistent asthma who were prescribed controller medications; and Mental Health Follow-up, percentage of patients with psychiatric hospitalization with outpatient follow-up within 1 week.
Insights Into Pediatric Pay-for-Performance Programs
Commercial health plan #1 Commercial health plan #2 Commercial health plan #3 Commercial health plan #4 Commercial health plan #5 State Medicaid #1 State Medicaid #2 State Medicaid #3
Payee
No. of Children in Program
Secondary Prevention
ACADEMIC PEDIATRICS
Table 1. Basic Pediatric Pay-for-Performance Program Characteristics and Incentivized Aspects of Pediatric Care*
189
190
Chien et al
When asked about more formal research techniques (ie, experimental or quasi-experimental study designs) to evaluate their programs, our informants denied having access to personnel or academic partners with whom they could partner. Issues Related to ‘‘Children With Chronic Health Conditions’’ Asthma and mental health conditions were the 2 most commonly incentivized chronic health conditions. None of our informants thought that their P4P programs could be used to tackle emerging health conditions because of the lack of availability of corresponding performance metrics. Informants of 3 programs questioned whether P4P should even be used as a strategy to improve care for children with chronic health conditions because of the great variety of conditions affecting these children and the correspondingly low volume of children with each condition. ‘‘These conditions are too low volume and it is hard to get a solid diagnosis for many of these conditions.’’ —Leader of Commercial Health Plan 2 Informants considered the care for children with mental health conditions to be beyond the reach of the P4P program for several reasons. First, the funds for these conditions came from insurance plan carve-outs—that is, specific sums of money made available to children with these diagnoses up to a certain annual or lifetime limit. Second, they reiterated the problem of small sample size. Third, leaders of 3 programs (all state Medicaid) explained that P4P may not be a necessary incentive for these issues because existing case management systems would identify and coordinate care for these children irrespective of P4P. ‘‘There is a mechanism for helping these types of kids. [We have a tool that] provides an opportunity to ask relevant questions on patient health needs and status.. It is done at the point of enrollment and may alert HMOs and caseworkers to special and immediate interventions they may need to carry out.’’ —Leader of State Medicaid Program 6 Our informants’ advice for using P4P to improve health care for children with chronic health conditions focused on developing ways to aggregate measures across conditions, and to incentivize aspects of care that are not condition specific but that can be applied across disparate conditions (eg, maintenance of up-to-date problems and medication lists, care coordination, or disease management). ‘‘It is very tough to incentivize small numbers of kids for a given disease or condition because it is hard to measure if the prevalence is low.. If there were standardized measures which took into account multiple rare conditions from more of a case management standpoint, that might be a good idea.’’ —Program director of State Medicaid Program 2 ‘‘We are doing poorly in these categories. We need to look at incentivizing coordination of care, additional
ACADEMIC PEDIATRICS
supports for these children, and incentivizing some more complex therapies for their conditions.’’ —Program director of Commercial Health Plan 1 DISCUSSION This study shows that early pediatric P4P programs are underway and are already affecting the care provided to our nation’s children, even though pediatricians have not routinely engaged in the design of these programs. It also shows that those designing and implementing these programs are grappling with basic program design problems, and they are paying less attention to subtle, albeit crucial, challenges, such as the potential for negative unintended consequences, risk adjustment, and overcoming inherent sample size limitations. Informants generally view their programs as effective, suggesting that their P4P efforts will probably continue. Our study is limited by our participant sample. Although our response rate matches what is described in the literature,23,24 informants who did not respond to our survey may have answered our questions differently. Our study only includes named program directors and their identified supporting staff, so their responses may differ from that of others with different roles in the P4P program. However, the aim of our study was to solicit and analyze the perspectives of those who currently have decision-making capacity or who hold leadership roles within these programs, not of those in rankand-file positions. Our study could also be limited by social desirability bias. The direction of this bias, if significant, is unclear. We made every effort to mitigate these effects by phrasing questions in a neutral fashion and by asking for supporting data where appropriate and available. This study provides a unique opportunity to compare the features of early pediatric P4P programs to the qualities endorsed by researchers, professional societies, and regulatory agencies.25–28 This literature emphasizes the importance of providing reward rather than penalty incentives,29 including physicians in program design, and selecting measures that reflect evidence-based clinical processes that health care organizations and/or their physicians’ control. It is of concern that in contrast to the cautious approach advocated above, pediatric P4P programs have begun instituting penalties, have been unable to engage pediatricians, and have introduced measures that may not necessarily reflect evidence-based clinical processes controlled by health care organizations or their physicians. It appears that future pediatric P4P efforts need to understand the limitations of our current P4P knowledge, and that researchers and physicians who care for children need to help build the empirical basis for rational pediatric P4P program design. Pediatric P4P is proliferating. Insights from those with experience designing and implementing these programs suggest that there is considerable room to improve our nation’s pediatric P4P efforts. ACKNOWLEDGMENTS We thank Charlie Homer and Stephen Muething for their input when this study was being conceived.
ACADEMIC PEDIATRICS
Insights Into Pediatric Pay-for-Performance Programs
This study was internally funded by the Department of Pediatrics, The University of Chicago, while Dr Chien was a member of the faculty there.
11. Fairbrother G, Siegel MJ, Friedman S, et al. Impact of financial incentives on documented immunization rates in the inner city: results of a randomized controlled trial. Ambul Pediatr. 2001;1:206–212. 12. Felt-Lisk S, Gimm G, Peterson S. Making pay-for-performance work in Medicaid. Health Aff. 2007;26:w516–w527. 13. Mandel KE, Kotagal UR. Pay for performance alone cannot drive quality. Arch Pediatr Adolesc Med. 2007;161:650–655. 14. Chien AT, Chin MH, Davis AM, Casalino LP. Pay-for-performance, public reporting and racial disparities in health care: how are programs being designed? Med Care Res Rev. 2007;64:285S–304S. 15. Glied S. Getting the incentives right for children. Health Serv Res. 1998;33(4 pt 2):1143–1160. 16. Freed GL, Uren RL. Pay-for-performance: an overview for pediatrics. J Pediatr. 2006;149:120–124. 17. Profit J, Zupancic JAF, Gould JB, Petersen LA. Implementing payfor-performance in the neonatal intensive care unit. Pediatrics. 2007;119:975–982. 18. Chien AT, Dudley RA. Pay-for-performance in pediatrics: proceed with caution. Pediatrics. 2007;120:186–188. 19. The Leapfrog Group. The Leapfrog group fact sheet. Available at: http://www.leapfroggroup.org/leapfrog-factsheet. November 2005. Accessed February 11, 2009. 20. Center for Health Care Strategies Inc. Centers for Medicare and Medicaid Services. Descriptions of selective performance incentives programs. Available at: http://www.cms.hhs.gov/smdl/downloads/ StatePerformanceIncentiveChart040606.pdf. Accessed February 11, 2009. 21. van der Lee JH, Mokkink LB, Grootenhuis MA, et al. Definitions and measurement of chronic health conditions in childhood: a systematic review. JAMA. 2007;297:2741–2751. 22. Miles M, Huberman A. Qualitative Data Analysis: An Expanded Sourcebook. 2nd ed. Thousand Oaks, Calif: Sage Publications; 1994. 23. Chien AT. The Potential Impact of Performance Incentive Programs on Racial Disparities in Health Care. Totowa, NJ: Humana Press; 2007. 24. Scanlon DP. Evidence for pay for performance: hope for the US health care system? Manag Care.. 2005;14(12 suppl):6–10. 25. American College of Physicians. Linking Physician Payments to Quality Care. Philadelphia, Pa: American College of Physicians; 2005. Position paper. Available at: http://www.acponline.org/ advocacy/where_we_stand/policy/link_pay.pdf. Accessed February 11, 2009. 26. American Medical Association. Guidelines for pay-for-performance programs. Available at: http://www.ama-assn.org/ama1/pub/upload/ mm/368/guidelines4pay62705.pdf. Accessed February 11, 2009. 27. American Medical Association. Principles for pay-for-performance programs. http://www.ama-assn.org/ama1/pub/upload/mm/368/princi ples4pay62705.pdf. Accessed February 11, 2009. 28. Van Amringe M. The JCAHO guidelines for pay for performance. Manag Care. 2005;14(12 Suppl):11–12. 29. Shen Y. Selection incentives in a performance-based contracting system. Health Serv Res. 2003;38:535–552.
REFERENCES 1. Centers for Medicare and Medicaid Services. Medicare pay-forperformance demonstration shows significant quality of care improvement at participating hospitals. May 3, 2005. Available at: http://www.cms.hhs.gov/apps/media/press/release. asp?Counter¼1441&intNumP erPage¼10&checkDate¼&checkKey¼2&srchType¼2&numDays¼0&srchOpt¼0&sr chData¼part þd&keywordType¼All&chkNewsType¼1%2Cþ2%2Cþ3%2Cþ4 %2Cþ 5&intPage¼&showAll¼1&pYear¼&year¼0&desc¼false &cboOrder¼date. Accessed February 11, 2009. 2. Centers for Medicaid and Medicare Services. Pay for performance initiatives. Herb Kuhn testimony for the Senate Committee on Finance. June 27, 2005. Available at: http://www.cms.hhs.gov/apps/ media/press/testimony.asp?Counter¼1537. Accessed February 11, 2009. 3. The Leapfrog Group. The Leapfrog compendium. Available at: http:// www.leapfroggroup.org/compendium2. Accessed February 11, 2009. 4. Kurhmerker K, Hartman T. Pay-for-performance in State Medicaid Programs: a survey of state Medicaid directors and programs. April 12, 2007. Available at: http://www.commonwealthfund.org/ publications/publications_show.htm?doc_id¼472891. 2007. Accessed February 11, 2009. 5. Kurhmerker K, Hartman T. Pay-for-performance in State Medicaid Programs. Appendix B: State pay-for-performance program summaries. Commonwealth Fund Publication 1018. Available at: http:// www.commonwealthfund.org/usr_doc/Kuhmerker_P4Pstate MedicaidprogsappendixB.pdf?section¼4039. Accessed February 11, 2009. 6. Dudley R, Frolich A, Robinowitz D, et al. Strategies to Support Quality-Based Purchasing: A Review of the Evidence. Summary, Technical Review 10. Prepared by the Stanford–University of California San Francisco Evidence-Based Practice Center under contract 290-02-0017. AHRQ Publication 04-0057. Rockville, Md: Agency for Healthcare Research and Quality; July 2004. 7. Petersen LA, Woodard LD, Urech T, et al. Does pay-for-performance improve the quality of health care? Ann Intern Med. 2006;145: 265–272. 8. Chien AT, Conti RM, Pollack HA. A pediatric-focused review of performance incentive literature. Curr Opin Pediatr. 2007;19: 719–725. 9. Hillman AL, Ripley K, Goldfarb N, et al. The use of physician financial incentives and feedback to improve pediatric preventive care in Medicaid managed care. Pediatrics. 1999;104(4 pt 1):931–935. 10. Fairbrother G, Hanson KL, Friedman S, Butts GC. The impact of physician bonuses, enhanced fees, and feedback on childhood immunization coverage rates. Am J Public Health. 1999;89:171–175.
191