Pay for Performance P4P and It s Impact on Health Disparities Network of Ethnic Physician Organizations Ethnic Physicia

1. Pay for Performance (P4P) and It�s Impact on Health DisparitiesNetwork of Ethnic Physician OrganizationsEthnic Physician Leadership Summit 2006Los Angeles, CaliforniaSeptember 30, 2006 Rodney G. Hood, MD President, Multicultural IPA CEO, Careview Medical Group San Diego, California

2. RadRod�s Disclosure Statements P4P was proposed and implemented more for cost containment rather than quality improvement. As currently being proposed P4P initiatives will exacerbate the financial reimbursement inequities in the health care system. P4P health care quality indicators do not always correlate with quality and give a false sense of quality improvement. In its current form P4P is a bad idea that can potentially worsen health disparities. No matter what we may believe about its inequities, P4P is here and we as ethnic health care providers must deal with it!

3. The Future of P4P �In the next 5 to 10 years pay-for performance-based compensation could account for 20% to 30% of what Medicare pays providers.� Mark McClellan, MD CMS Administrator (2004)

4. Quality Indicators and Health Disparities

5. Overview Utilization Trends in Racial and Ethnic Health Disparities IOM Unequal Treatment Report Overall Utilization Health Disparities Data Trends Suggest: � When more of a therapeutic modality may be beneficial, Blacks receive less but when less of a therapeutic modality may be beneficial, Blacks receive more.�Overall Utilization Health Disparities Data Trends Suggest: � When more of a therapeutic modality may be beneficial, Blacks receive less but when less of a therapeutic modality may be beneficial, Blacks receive more.�

6. Minorities Are Not All the SameNational Health Data by Race & Ethnicity Healthy People 2010 Target Goals�Deaths per 100,000 population Healthy People 2010 set target outcome goals to achieve by 2010. These target outcome goals are measured in deaths per 100,000 population except for Infant Mortality measured as deaths per 1000 live births. Findings: The overall death rate burden for Blacks is 34% greater than for Whites and 93% greater death burden than for Hispanics. Hispanics have met or surpassed the HP2010 outcome targets in 6 of 8 Asians/PI and Native Americans have met or surpassed HP2010 outcome targets in 5 of 8 categories; however Blacks and Whites meet none of the HP 2010 target disease goals. The overall death rates for all causes are highest for Blacks>Whites>Native Americans>Hispanics>Asians/PI Healthy People 2010 set target outcome goals to achieve by 2010. These target outcome goals are measured in deaths per 100,000 population except for Infant Mortality measured as deaths per 1000 live births. Findings: The overall death rate burden for Blacks is 34% greater than for Whites and 93% greater death burden than for Hispanics. Hispanics have met or surpassed the HP2010 outcome targets in 6 of 8 Asians/PI and Native Americans have met or surpassed HP2010 outcome targets in 5 of 8 categories; however Blacks and Whites meet none of the HP 2010 target disease goals. The overall death rates for all causes are highest for Blacks>Whites>Native Americans>Hispanics>Asians/PI

7. Quality of Care and Access to Care Comparisons by Selected Racial Groups 2000 � 2001 National Healthcare Disparities Report 2004 (AHRQ) Agency for Healthcare Research and Quality (AHRQ) DHHS annual National Healthcare Disparities ReportAgency for Healthcare Research and Quality (AHRQ) DHHS annual National Healthcare Disparities Report

9. Health Care Quality Indicator DisparitiesAugust 2006 issue of the American Journal of Preventive Medicine  In 2000 � 2001, the overall biennial breast screening rates for women 40yrs and older were: 50.6 percent for non-Hispanic white women 40.5 percent for black women 34.7 percent for Asian-American women 36.3 percent for Hispanic women, and 12.5 percent for Native-American women. Therefore, 20% � 75% lower rates for minorities In California, women with insurance have an overall breast screen rate at 64% but approximately 70% for whites but less for Asians (Filipino & Chinese), immigrants, non-English speaking and other minority women. Self-reported cancer screening for PAPS and mammography for African Americans and Latinos are near or equal to whites but when documented by medical records the actual screening rates are significantly less. Maxwell and Bastani, �Breast cancer screening and related attitudes among Filipino-American women�, Cancer Epidemiology Biomarkers & Prevention, Vol. 6, Issue 9, American Association for Cancer Research 1997. (UCLA School of Public Health) Results � Filipino women 50 years and older in LA 66% had ever had a mammogram, 42% had one in the past 12 months and 54% in the past 2 years. These rates are 20% less than AA and white women as noted in the 1994 California Behavioral Risk Factor Survey. Goel, Wee, et al, Racial and Ethnic disparities in cancer screening: �The importance of Foreign birth as a barrier to care.� J. of General Internal Med., Dec. 2003. Results: Racial and ethnic groups comprised largely of foreign- born individuals have lower rates of cancer screening than other Americans. This study suggest that foreign birthplace may contribute to disparities as an independent factor in certain racial and ethnic groups. Haitt, et al, �Community-based cancer screening for underserved women: Design and baseline findings from the breast and cervical cancer intervention study.�, Preventive Medicine, Sept. 2001. Results: 66% of women ages 40 and older had at least one mammogram and most at least a clinical breast exam (88%) and PAPS smear (89%). However rates were significantly lower for non-English speaking Latinas and Chinese women (56% and 32% respectively for mammography and maintenance screening (3 mammograms in past 5 years) varied from 7% (non-English speaking Chinese) to 53% for Black women. Pap smear screening in the past 3 years was low among non-English speaking Latinas (72%) and markedly lower among non-English speaking Chinese women (24%). Chanpion, et al, �Validity of self-reported mammography in low-income African American women.�, Am. J. Prev. Medicine, Feb. 1998. Results: Although a few studies have shown a close correspondence between self-reported mammography and medical records, most had few minority participants. Small group of AA women (229) self-reported breast screening when compared to medical records showed that only 49% to 60% of reported mammography use could be verified by medical records. (Indiana School of Nursing)Maxwell and Bastani, �Breast cancer screening and related attitudes among Filipino-American women�, Cancer Epidemiology Biomarkers & Prevention, Vol. 6, Issue 9, American Association for Cancer Research 1997. (UCLA School of Public Health) Results � Filipino women 50 years and older in LA 66% had ever had a mammogram, 42% had one in the past 12 months and 54% in the past 2 years. These rates are 20% less than AA and white women as noted in the 1994 California Behavioral Risk Factor Survey. Goel, Wee, et al, Racial and Ethnic disparities in cancer screening: �The importance of Foreign birth as a barrier to care.� J. of General Internal Med., Dec. 2003. Results: Racial and ethnic groups comprised largely of foreign- born individuals have lower rates of cancer screening than other Americans. This study suggest that foreign birthplace may contribute to disparities as an independent factor in certain racial and ethnic groups. Haitt, et al, �Community-based cancer screening for underserved women: Design and baseline findings from the breast and cervical cancer intervention study.�, Preventive Medicine, Sept. 2001. Results: 66% of women ages 40 and older had at least one mammogram and most at least a clinical breast exam (88%) and PAPS smear (89%). However rates were significantly lower for non-English speaking Latinas and Chinese women (56% and 32% respectively for mammography and maintenance screening (3 mammograms in past 5 years) varied from 7% (non-English speaking Chinese) to 53% for Black women. Pap smear screening in the past 3 years was low among non-English speaking Latinas (72%) and markedly lower among non-English speaking Chinese women (24%). Chanpion, et al, �Validity of self-reported mammography in low-income African American women.�, Am. J. Prev. Medicine, Feb. 1998. Results: Although a few studies have shown a close correspondence between self-reported mammography and medical records, most had few minority participants. Small group of AA women (229) self-reported breast screening when compared to medical records showed that only 49% to 60% of reported mammography use could be verified by medical records. (Indiana School of Nursing)

10. California Integrated Health Association (IHA) A Pay for Performance Initiative in California

11. History of California Integrated Health Association (IHA) P4P Initiative In July 2000 a high level working group of California health care leaders from health plans, physicians, medical directors, etc. met to discuss a new statewide initiative for P4P. January 2002 six California health plans (Aetna, Blue Cross, Blue Shield, CIGNA, HealthNet and PacifiCare) launched this new initiative. A score card of common performance measures were agreed upon with clinical measures weighted at 50%, patient satisfaction weighted at 40% and Information Technology (IT) at 10%. Updates of this initiative began in 2003

12. Integrated Health Association (IHA)Evidence based Pay for Performance Quality Measures IHA is a California leadership coalition consisting of health industry stakeholder groups, consultants, pharmaceutical companies, health plans (Aetna, Blue Shield of Calif, Blue Cross of Calif, CIGNA, HealthNet, and PacifiCare), physicians, academic institutions, purchasers and consumers.IHA is a California leadership coalition consisting of health industry stakeholder groups, consultants, pharmaceutical companies, health plans (Aetna, Blue Shield of Calif, Blue Cross of Calif, CIGNA, HealthNet, and PacifiCare), physicians, academic institutions, purchasers and consumers.

13. California HMO Report Card 2005

14. Pay for Performance Initiative in San Diego County Commercial HMO Products

15. MCIPA is a for profit Independent Physician Association (IPA) that was established in San Diego County California and was managed by the UCSD Health Network in 1994. Since 2003 MCIPA has been managed by SynerMed located in Los Angeles. MCIPA generates $6 million yearly from commercial, senior and Medicaid direct health plan contracts and composed of 50 PCPs and over 50 specialty health care providers. The MCIPA has 12,000 enrollees (8,000 commercial) with providers and enrollees that are ethnically diverse. Enrollees are mostly Latino and African American but include Asian, African and other Immigrants and those of European descent. MCIPA providers and enrollees are predominantly located in Central & South regions of San Diego County.

16. Physician Medical Group Practice Mix by Race and Ethnicity Group I � 3 AA PCPs and 1 Asian PCP � Ethnic patient population mix is majority African American (68% Black, 17% Latino, 8% Asian and 7% Caucasian). Group II � 2 Latino PCPs � Ethnic patient population mix is majority Latino. Group III � 1 Asian PCP � Ethnic patient population mix is majority Asian (Filipino). Group IV � 1 European PCP � Ethnic patient population mix is majority Caucasian. Group I (CVMG) � 3 AA PCPs and 1 Asian PCP � Ethnic patient population mix is majority African American (68% Black, 17% Latino, 8% Asian and 7% Caucasian). Group II (Borrero) � 2 Latino PCPs � Ethnic patient population mix is majority Latino. Group III (Razon) � 1 Asian PCP � Ethnic patient population mix is majority Asian (Filipino). Group IV (Sy Mallis) � 1 European PCP � Ethnic patient population mix is majority Caucasian. Group I (CVMG) � 3 AA PCPs and 1 Asian PCP � Ethnic patient population mix is majority African American (68% Black, 17% Latino, 8% Asian and 7% Caucasian). Group II (Borrero) � 2 Latino PCPs � Ethnic patient population mix is majority Latino. Group III (Razon) � 1 Asian PCP � Ethnic patient population mix is majority Asian (Filipino). Group IV (Sy Mallis) � 1 European PCP � Ethnic patient population mix is majority Caucasian.

17. San Diego County Demographics by Race, Ethnicity and Disease Burden Geographic regions include: North, North coastal, Central, Eastern, Inland and South regions. Latinos, African Americans and Immigrant populations are concentrated in the Central and South regions of San Diego County. SD County Health Needs Assessment Report (2004) shows the populations with the highest disease burdens and greatest obstacles to access health care are found in the Central and South regions with African Americans suffering the highest disease burdens and Latinos the worst access. The Central and South regions of San Diego County have the highest hospitalization and death rates from diabetes, asthma, CHD and cancer.

18. Physician Shortage Leads to High Patient Volumes San Diego County population is approximately 3 million with 8,700 physicians. Physician:population ratio in San Diego County is 1:350. Physician:population ratio for MCIPA service areas is approximately 1:1500. Therefore, MCIPA service areas have a physician shortage of 4 times fewer physicians than other parts of the county.

19. California HMO Report Card 2005Medical Groups in San Diego County

20. P4P Inequities for High-Risk Populations

21. Reasons for Low Quality Performance with High-Risk Populations Inequities Encountered with Disproportionate Enrollment of High-Risk Populations Low financial reimbursements P4P administrative cost Group quality indicator disparities Incomplete encounter data Unfair quality indicator comparisons Tiered networks and physician economic profiling De facto racial and ethnic discrimination The Ultimate Inequity UltimateUltimate

22. P4P Inequity #1 - Reimbursement Physicians� health services are reimbursed based upon the average costs that assumes the enrolled population has a bell-shaped curve �risk� distribution with an equal balance of low and high-risk populations. If the served population has an adverse risk selection based upon race, ethnicity, geographic location or SES the average service costs are expected to be higher. If a group�s adverse high-risk population is reimbursed at the lower rates for the average-risk population they will receive less compensation for their high-risk population.

23. Population Disease Burden and Risk Distribution Utilized in Managed Care Reimbursement Formulas

24. Medical Group Managed Care Reimbursement Formula Assumptions for Commercial Product If the managed care average-risk population has a professional services cost of $50 pmpm, and If the managed care low-risk population has a professional service cost of $40 pmpm, and If the managed care high-risk population has a professional service cost of $60 pmpm, but no matter the actual risk; The contracting physicians or medical groups are reimbursed at the average-risk cost minus HMO administrative withhold then reimbursement is more or less depending upon the number of services contracted and the groups negotiating strength or weakness. Therefore, a medical group with a disproportionate high-risk population enrollment and a weak negotiation position due to small enrollment will likely receive a rate between the low-risk vs average-risk rates.

25. P4P Inequity #2 - Cost The HMO withholds $3 to $4 pmpm from participating physician groups to cover P4P incentive cost �not extra money. The physician group P4P quality improvement program cost $1 pmpm to implement. A fee is charge to the Group ($2000) to cover costs of the patient survey portion. Therefore, the incentive withhold, the group program costs, and other fees further diminishes physicians� reimbursements.

26. P4P Inequity #3 Disparities with Quality Indicators The groups serving populations having the highest health disparities and the greatest disease burdens such as Blacks, Latinos and Asians have lower average baseline quality indicator levels than the general population. Therefore, P4P quality indicator criteria based upon low-risk groups will establish goals that are disproportionately higher when compared to the high-risk groups. Therefore, groups with high disease burden (high-risk) populations will receive little or no financial benefit from the P4P incentive withhold and in fact may be penalized with even less reimbursement.

27. Cancer Screening in CaliforniaUCLA Center for Health Policy Research Health Interview Survey Self-Reported PAPS Tests � December 2003 NHOPI = Native Hawaiian and other Pacific Islanders, AI/AN = American Indian and Alaska Native Champion, Menon, �Am.J.Prev. Med Feb. 1998 � �Validity of self-reported mammography in low-income African American women.� � Results showed that comparison with self-report showed that only 49% to 60% of the reported mammography use could be verified with medical record documentation. Paskett, Tatum,, Cancer Epidemiology Biomarkers & Prevention, Issue 9, Vol 5, 1996. �Validation of self-reported breast and cervical screening test among low-income minority women.� Results showed that of the self-reported rates of breast and cervical screening for mammography, 77% of self-reports were correct and for PAPS it was only 67% verified.NHOPI = Native Hawaiian and other Pacific Islanders, AI/AN = American Indian and Alaska Native Champion, Menon, �Am.J.Prev. Med Feb. 1998 � �Validity of self-reported mammography in low-income African American women.� � Results showed that comparison with self-report showed that only 49% to 60% of the reported mammography use could be verified with medical record documentation. Paskett, Tatum,, Cancer Epidemiology Biomarkers & Prevention, Issue 9, Vol 5, 1996. �Validation of self-reported breast and cervical screening test among low-income minority women.� Results showed that of the self-reported rates of breast and cervical screening for mammography, 77% of self-reports were correct and for PAPS it was only 67% verified.

28. Cancer Screening in CaliforniaUCLA Center for Health Policy Research Health Interview Survey Self-Reported Mammography - December 2003 Champion, Menon, �Am.J.Prev. Med Feb. 1998 � �Validity of self-reported mammography in low-income African American women.� � Results showed that comparison with self-report showed that only 49% to 60% of the reported mammography use could be verified with medical record documentation. Paskett, Tatum,, Cancer Epidermiology Biomarkers & Prevention, Issue 9, Vol 5, 1996. �Validation of self-reported breast and cervical screening test among low-income minority women.� Results showed that of the self-reported rates of breast and cervical screening for mammography, 77% of self-reports were correct and for PAPS it was only 67% verified.Champion, Menon, �Am.J.Prev. Med Feb. 1998 � �Validity of self-reported mammography in low-income African American women.� � Results showed that comparison with self-report showed that only 49% to 60% of the reported mammography use could be verified with medical record documentation. Paskett, Tatum,, Cancer Epidermiology Biomarkers & Prevention, Issue 9, Vol 5, 1996. �Validation of self-reported breast and cervical screening test among low-income minority women.� Results showed that of the self-reported rates of breast and cervical screening for mammography, 77% of self-reports were correct and for PAPS it was only 67% verified.

29. Relationship Among Race, Ethnicity, SES, Foreign Birth and Non-English Speaking on Cancer Screening Rates Champion,�Validity of self-reported mammography in low-income African American women.�, Am. J. Prev. Med. Feb. 1998: Results showed AA women self-reported mammography with only 49% - 60% that could be verified with medical record documentation. Paskett, �Validation of self-reported breast and cervical screening test among low-income minority women.�, Cancer Epidemiology Biomarkers & Prevention, 1996: Results showed that low-income minority women self-reported mammography rates were 77% correct and 67% for self-reported PAPS. Maxwell, AE, �Breast cancer screening and related attitudes among Filipino-American women.� Cancer Epidemiology Biomarkers & Prevention, 1997: Results showed Filipino women 50 years and older residing in Los Angeles with 66% never having a mammogram, 42% had had one in the past 12 months, and 54% in the past 2 years. Goel, MS, �Racial and Ethnic Disparities in Cancer Screening: The Importance of Foreign Birth as a Barrier to Care.�, J. General Internal Med., Dec. 2003: Results show foreign born women in US (Latino, Asian and Pacific Islanders) were significantly less likely to report cancer screening than US born counterparts.

30. P4P Inequity #4 Incomplete Encounter Data Physicians� services encounter data is utilized to measure physician groups� levels of compliance for quality improvement measures. Physicians with less information technology (IT) capacity tend to submit incomplete encounter data at higher rates. Therefore, incomplete collection of encounter data results in lower quality indicator scores.

31. P4P Inequity #5 Unfair Quality Comparisons Each physician group�s quality data are published as a quality report card. Physicians serving disproportionate high-risk populations will be perceived as giving poor quality and therefore negatively affect enrollment.

32. P4P Inequity #6Tiered Physician Networks and Physician Economic Profiling Tiered Physician Networks: Physicians are partitioned into different tiers utilizing physician economic profiling that will classify a physician or group into levels of cost efficiency. Physician Economic Profiling: Those select physician groups that are deemed cost-efficient are placed into a select network tier that offer patients lower co-pays and a more enriched benefit plan which gives incentives for greater enrollment. Traditional High-Risk Providers: Physicians or groups with high-risk populations (SES, geographic location, high disease burdens or co-morbidities and race) are deemed less cost-efficient and further penalized by being forced into lower tiered plans that offer higher co-pays, fewer benefits and encourage lower-risk patients not to enroll with traditional high-risk providers.

33. P4P Inequity #7De facto Racial & Ethnic Discrimination P4P creates disincentives for physicians and medical groups to not enroll high-risk patients that are disproportionately ethnic minorities therefore encouraging de facto racial, social and economic discrimination. Thus, high-risk patients default to traditional health care providers further worsening disease burdens and quality indicator data for these high-risk populations.

34. P4P Ultimate Inequity #8Worsening Health Disparities P4P programs that do not fairly and equitably compensate for high-risk populations and then utilize unfair quality indicator comparisons will not enhance the elimination of health disparities but may actually worsen health disparities.

35. Health Disparities Math Assume quality gradient of 1 10 (best): Whites = 6 and minorities = 4 Disparity difference = 2 Goal: Improve quality to 9: We need to achieve a 50% (6 to 9) increase for whites and 125% (4 to 9) increase for minorities in order to achieve equity. If we achieved a 50% equal improvement for all: Whites = 6 to 9 minorities = 4 to 6 Disparity difference = 3 Therefore we have a worsening quality disparity of 50%. use of Evidence-based medicine (EBM) without special attention to minorities may increase quality for all but widen minority disparities: use of Evidence-based medicine (EBM) without special attention to minorities may increase quality for all but widen minority disparities:

36. Early Experience with Pay-for-Performance in CaliforniaRosenthal, et al, JAMA, Oct. 2005 (Harvard School of Public Health) Finding: For all 3 measures (cervical cancer screening, mammography and hemoglobin A1c), physician groups with baseline performance at or above the performance threshold for receipt of a bonus improved the least but garnered the largest share of the bonus payments ($3.4 million). Conclusion: �Paying clinicians to reach a common, fixed performance target may produce little gain in quality for the money spent and will largely reward those with higher performance at baseline.� Rosenthal, et al, JAMA, 2005 (Harvard School of Public Health), Early Experience with Pay-for-Performance. Study compared P4p initiative in California with large health plan in Pacific Northwest and followed 3 measures � cervical cancer screen, mammography and hemoglobin A1c � Compared with physician groups in the Pacific Northwest the California network demonstrated greater quality improvement after p4p intervention only in cervical cancer screening (3.6% improvement) after awarding $3.4 million or 27% of the amount set aside in bonus payments. Reports from Oct 2001 to April 2004 and bonus given for 2003-2004.Rosenthal, et al, JAMA, 2005 (Harvard School of Public Health), Early Experience with Pay-for-Performance. Study compared P4p initiative in California with large health plan in Pacific Northwest and followed 3 measures � cervical cancer screen, mammography and hemoglobin A1c � Compared with physician groups in the Pacific Northwest the California network demonstrated greater quality improvement after p4p intervention only in cervical cancer screening (3.6% improvement) after awarding $3.4 million or 27% of the amount set aside in bonus payments. Reports from Oct 2001 to April 2004 and bonus given for 2003-2004.

37. Lessons and Recommendations Health care disparities are quality issues that came about because of health care inequities. Recommendation: Cautiously adopt the concept of P4P as a tool to address health disparities as a quality issue and begin to eliminate barriers causing the inequities. P4P is a potential tool to monitor and improve health disparities. Recommendation: P4P has the potential to worsen health disparities. We must not continue to allow the system to ignore and minimize the inequities of performance criteria that do not address risk factors such as population disease burdens, access disparities, geographic disparities and race as independent variables for health risks. Baseline reimbursements should reflect the population�s risk levels. Recommendation: Reimbursements must reflect the population�s level of risk and P4P incentive payments should be based upon percent improvement of the actual groups� baseline quality measures rather than set levels that are based upon lower risk populations.

38. Lessons and Recommendations Physician groups associated with larger networks perform better probably because of access to better management tools. Recommendation: Medical practice integration and embracing information technology will be imperative for success. Independent physicians and small physician groups must find ways to integrate their practices with larger entities in order to take advantage of cost efficiencies and access to IT. Health Policy advocates should prioritize to bring about programs and legislation at both the state and national levels that promote these changes by: Obtaining grants to establish P4P Quality Improvement pilot programs with physicians and medical groups serving high-risk populations. Advocate with our allies for California DMC and CMS to monitor and mandate that HMOs establish P4P reimbursement formulas that will ensure appropriate reimbursement rates for high-risk populations.

39. The Danger of Perceived Quality Improvement with P4P in Worsening Health Disparities �Pay-for-Performance� programs that have been initiated do not adequately take into account the documented baseline quality performance variances found within and between different racial and ethnic groups.� ����� �Not considering specific health disparities risk variables (ses, geographic location, race and level of disease burdens) creates the real potential to economically penalize and cause unintended disincentives for individual physicians, medical groups and health institutions that have traditionally provided health services for these high-risk populations. Providers treating high-risk populations are penalized even if they obtain significant quality improvement. These inequities will further worsen quality of care in high-risk populations and widen healthcare disparities.� Comments by Rodney G. Hood, MD, NMA Representative to the AMA HOD at the AMA Forum on P4P on December 5, 2004

40. The Challenge Like it or not, P4P is a reality that is now being utilized and presumed to monitor and measure health quality � We must therefore become engaged and ensure P4P will improve health quality for all of America�s diverse populations.