1. Heart Rhythm 2008 San Francisco May 15, 2008 Stephen Hammill, MD Heart Rhythm 2008 San Francisco May 15, 2008 Stephen Hammill, MD CP1262561-1 Using the ICD Registry to Evaluate Practice Patterns and Improve Quality Overview of the Development of the ICD Registry

2. ICD Registry Highlights to Date CP1262561-2 1,490 hospitals >260,000 implants (+10,000/month) 55% are primary prevention CMS patients 88% of implants from hospitals entering all patients (1° and 2° prevention, all ages) Version 2.0 in development Longitudinal registry developed (CED) Research and publications in progress 1,490 hospitals >260,000 implants (+10,000/month) 55% are primary prevention CMS patients 88% of implants from hospitals entering all patients (1° and 2° prevention, all ages) Version 2.0 in development Longitudinal registry developed (CED) Research and publications in progress Getting to this point and moving forward

3. ICD Registry Developing the Baseline Registry CP1262561-3 9/28/04: CMS published proposed NCD Following SCD-HeFT release National data base proposed HRS asked to chair the Working Group to develop the registry 9/28/04: CMS published proposed NCD Following SCD-HeFT release National data base proposed HRS asked to chair the Working Group to develop the registry

4. National ICD Registry Working Group HRS, chair ACC HFSAAHA MedtronicBiotronik GuidantSt Jude BCBSNCDR United HealthCareAetna Am Hlth Ins PlansAHRQ FDAAm Hosp Assoc CMSAt large members HRS, chair ACC HFSAAHA MedtronicBiotronik GuidantSt Jude BCBSNCDR United HealthCareAetna Am Hlth Ins PlansAHRQ FDAAm Hosp Assoc CMSAt large members

5. ICD Registry Developing the Baseline Registry CP1262561-3 9/28/04: CMS published proposed NCD Following SCD-HeFT release National data base proposed HRS asked to chair the Working Group to develop the registry 11/22/04: Working Group recommendations sent to CMS Purpose of the registry Patients to be enrolled Patient and device data elements to be collected Defining providers as competent and qualified to implant ICDs 9/28/04: CMS published proposed NCD Following SCD-HeFT release National data base proposed HRS asked to chair the Working Group to develop the registry 11/22/04: Working Group recommendations sent to CMS Purpose of the registry Patients to be enrolled Patient and device data elements to be collected Defining providers as competent and qualified to implant ICDs

6. ICD Registry Developing the Baseline Registry CP1262561-4 1/27/05: CMS published final NCD Expanded ICD indications CED process described Data collection using QNet Temporary data collection tool 1/27/05: CMS published final NCD Expanded ICD indications CED process described Data collection using QNet Temporary data collection tool

7. What does CMS hope to gain from the Registry? CMSs goal is to determine whether primary prevention ICDs are appropriate for the Medicare beneficiaries who meet the clinical conditions identified in the agencies NCD of 1/27/05. Coverage with evidence development (CED) Hammill, Phurrough, Brindis. HeartRhythm, 2006

8. CED Coverage with evidence development Develop evidence on what works best in clinical practice... explicit, rapid, evidence based on a process that is predictable with transparency... to improve the knowledge base by which patients and providers can make better treatment decisions. Mark McClellan Administrator, CMS 2/14/05 Conference call

9. ICD Registry Developing the Baseline Registry CP1262561-4 1/27/05: CMS published final NCD Expanded indications CED process described Data collection using QNet Temporary data collection tool 3/05: HRS asked to reconvene the Working Group Define questions that should be answered Define the core characteristics of a national clinical registry 5/19/05: Recommendations sent to CMS 1/27/05: CMS published final NCD Expanded indications CED process described Data collection using QNet Temporary data collection tool 3/05: HRS asked to reconvene the Working Group Define questions that should be answered Define the core characteristics of a national clinical registry 5/19/05: Recommendations sent to CMS

10. ICD Registry Developing the Baseline Registry CP1262561-5 10/27/05: CMS selected the ICD Registry developed by ACC and HRS based on the NCDR

11. ICD Registry Developing the Baseline Registry CP1262561-5 10/27/05: CMS selected the ICD Registry developed by ACC and HRS based on the NCDR 1/1/06: All data submitted to ICD Registry; QNet phased out 4/1/06 Hospitals encouraged to submit data on all patients SCD-HeFT – median age 60 yrs Medicare – median age 74 yrs 10/27/05: CMS selected the ICD Registry developed by ACC and HRS based on the NCDR 1/1/06: All data submitted to ICD Registry; QNet phased out 4/1/06 Hospitals encouraged to submit data on all patients SCD-HeFT – median age 60 yrs Medicare – median age 74 yrs

12. ICD Registry Developing the Baseline Registry CP1262561-5 10/27/05: CMS selected the ICD Registry developed by ACC and HRS based on the NCDR 1/1/06: All data submitted to ICD Registry; QNet phased out 4/1/06 Hospitals encouraged to submit data on all patients SCD-HeFT – median age 60 yrs Medicare – median age 74 yrs 4/07: Quarterly benchmarking reports sent to hospitals Data Quality Reporting process Random auditing 10/27/05: CMS selected the ICD Registry developed by ACC and HRS based on the NCDR 1/1/06: All data submitted to ICD Registry; QNet phased out 4/1/06 Hospitals encouraged to submit data on all patients SCD-HeFT – median age 60 yrs Medicare – median age 74 yrs 4/07: Quarterly benchmarking reports sent to hospitals Data Quality Reporting process Random auditing

13. ICD Registry Data Quality Program National Onsite Audit Program Annual review 10% random sample of eligible sites Comparative analysis of audit findings and sites original data submission (MD training discrepancy, lack of complications, etc) National Onsite Audit Program Annual review 10% random sample of eligible sites Comparative analysis of audit findings and sites original data submission (MD training discrepancy, lack of complications, etc)

14. Why is Benchmarking Important to Hospitals? Quarterly benchmarking compares a hospitals outcomes with hospitals of similar size and a national aggregate Improved Patient Care Meet State Regulations and Payer Requirements Detect Inefficiencies Improved resource utilization Quarterly benchmarking compares a hospitals outcomes with hospitals of similar size and a national aggregate Improved Patient Care Meet State Regulations and Payer Requirements Detect Inefficiencies Improved resource utilization

15. Potential Benefits of the Registry Reveal the degree to which clinicians are managing a disease in accordance with evidence-based medicine Enable clinicians to compare their own outcomes with those of other MDs Provide insights for clinical investigation Highlight a products performance outside of clinical trial constraints Reveal the degree to which clinicians are managing a disease in accordance with evidence-based medicine Enable clinicians to compare their own outcomes with those of other MDs Provide insights for clinical investigation Highlight a products performance outside of clinical trial constraints Hammill, Phurrough, Brindis. HeartRhythm, 2006

16. Potential Benefits of the Registry Provide a detailed view of the morbidity, mortality, and resource utilization associated with a particular disease Perform local hospital needs for QA and QI Serve as a hospital and physician response to Pay for Performance initiatives of health plans Provide a detailed view of the morbidity, mortality, and resource utilization associated with a particular disease Perform local hospital needs for QA and QI Serve as a hospital and physician response to Pay for Performance initiatives of health plans Hammill, Phurrough, Brindis. HeartRhythm, 2006

17. ICD Registry A Quality Improvement Tool AHRQ CP1262561-5 Observational registries can quickly accumulate large amounts of data on real- world practice and effectiveness of new treatments and procedures. Physicians and hospitals can use these data to further QI efforts at a local level and physician associations can evaluate data to determine the effectiveness of existing clinical guidelines. Registries for Evaluating Patient Outcomes: A Users Guide. A Users Guide. Agency for Health Care Research and Quality Agency for Health Care Research and Quality

18. National ICD Registry 2006 CP1281777-3 Age (yr)68 Male/female (%)74/26 Race (%) White83 Black/African American12 Hispanic 5 Asian 1 American Indian/Alaska Native 0.4 Native Hawaiian 0.2 Other 4 Age (yr)68 Male/female (%)74/26 Race (%) White83 Black/African American12 Hispanic 5 Asian 1 American Indian/Alaska Native 0.4 Native Hawaiian 0.2 Other 4 Hammill et al: Heart Rhythm, 2007

19. National ICD Registry 2006 CP1281777-2 Hammill et al: Heart Rhythm, 2007 Total implants (no.)206,604 Hypertension (%) 74 Diabetes (%) 36 Chronic lung disease (%) 22 Ischemic heart disease (%) 66 Non-ischemic dilated cardiomyopathy (%) 31 Hx of cardiac arrest (%) 11 Hx of CHF, no. (%) 77 NYHA class II-III (%) 81 QRS duration (ms) 129 Ejection fraction (%) 28 Total implants (no.)206,604 Hypertension (%) 74 Diabetes (%) 36 Chronic lung disease (%) 22 Ischemic heart disease (%) 66 Non-ischemic dilated cardiomyopathy (%) 31 Hx of cardiac arrest (%) 11 Hx of CHF, no. (%) 77 NYHA class II-III (%) 81 QRS duration (ms) 129 Ejection fraction (%) 28

20. National ICD Registry 2006 CP1281777-1 Total implants (no.) 108,341 Single chamber ICD (%) 23 Dual chamber ICD (%) 39 Biventricular ICD (%) 38 Adverse procedure-related event (%) 3.24 Death in laboratory 0.02 Cardiac perforation 0.07 Hematoma 0.99 Lead dislodgement 0.97 Hemo/pneumothorax 0.54 Transient ischemic attack/stroke 0.06 Other 0.61 Total implants (no.) 108,341 Single chamber ICD (%) 23 Dual chamber ICD (%) 39 Biventricular ICD (%) 38 Adverse procedure-related event (%) 3.24 Death in laboratory 0.02 Cardiac perforation 0.07 Hematoma 0.99 Lead dislodgement 0.97 Hemo/pneumothorax 0.54 Transient ischemic attack/stroke 0.06 Other 0.61 Hammill et al: Heart Rhythm, 2007

21. National ICD Registry 2006 CP1281777-5 Hammill et al: Heart Rhythm, 2007 Physicians % Implants performed % Training category EP fellowship + boards 51 77 EP fellowship only 6 6 Surgery residency 11 2 HRS ICD Guidelines 17 9 None of the above 15 6 Training category EP fellowship + boards 51 77 EP fellowship only 6 6 Surgery residency 11 2 HRS ICD Guidelines 17 9 None of the above 15 6 3,899 implanting physicians 206,604 procedures 206,604 procedures

22. Are payors and providers adversaries or partners? adversaries or partners?

23. Hospital and PhysicianProfiling CMS Health Plans State and Federal Lawmakers State Departments Of Health Patients And Families Insurers

24. Administrative data versus Clinical data

25. CP1265452-1 Conclusions – Cardiac surgery report cards using administrative data are problematic compared with those derived from audited and validated clinical data, primarily because of case misclassification and nonstandarized end points.

26. FL MA O O NJ CA MD IL MI WV State Regulators Are Using the Cath NCDR® Regulation Passed Regulation Pending Regulation Passed Regulation Pending PA IA AL

27. Payers and the ICD Registry May 24, 2007: United Health Care, ACC and HRS announced an initiative that will improve patient care by expanding the collection and use of clinical outcomes data for heart rhythm devices by cardiac care facilities, cardiac surgeons, cardiologists, and other healthcare professionals Hospitals seeking designation as a United Health Premium Cardiac Specialty Center will be required to submit data to the National ICD Registry May 24, 2007: United Health Care, ACC and HRS announced an initiative that will improve patient care by expanding the collection and use of clinical outcomes data for heart rhythm devices by cardiac care facilities, cardiac surgeons, cardiologists, and other healthcare professionals Hospitals seeking designation as a United Health Premium Cardiac Specialty Center will be required to submit data to the National ICD Registry

28. Developing the Longitudinal Registry to Answer the CED Questions CP1262561-6

29. It is particularly important that these factors (ICD firing data and survival) be determined in the actual population receiving ICDs, who are older and present more comorbidities than represented in the trial populations. Fortunately, these key factors will be tracked in the ICD Registry Lynne Warner Stevenson, MD Circulation. 2006;114:101

30. CED Questions EF 31-15% EF 31-15% Non-ischemic cardiomyopathy <9 months Non-ischemic cardiomyopathy <9 months Class IV CRT-D patients Class IV CRT-D patients EF 31-15% EF 31-15% Non-ischemic cardiomyopathy <9 months Non-ischemic cardiomyopathy <9 months Class IV CRT-D patients Class IV CRT-D patients

31. Longitudinal Registry Study Design CP1262561-13 Primary endpoint Primary endpoint First delivery of an appropriate ICD therapy (shock, ATP) Secondary endpoint Secondary endpoint Survival probability at 3 and 5 years Death from CV cause Total # and rate of device therapies Ratio of inappropriate to total device therapies Primary endpoint Primary endpoint First delivery of an appropriate ICD therapy (shock, ATP) Secondary endpoint Secondary endpoint Survival probability at 3 and 5 years Death from CV cause Total # and rate of device therapies Ratio of inappropriate to total device therapies

32. Longitudinal Registry Study Design CP1262561-14 350 randomly selected implanting MDs 350 randomly selected implanting MDs 3,500 patients followed 3 years for events and 5 years for survival 3,500 patients followed 3 years for events and 5 years for survival Based on 10% rate of appropriate therapy at 3 years (15% at 3 years in SCD-HeFT) Device therapy follow-up Device therapy follow-up Every 3 months for a minimum of 3 years Adjudication process Data combined with NDI and Medicare claims data 350 randomly selected implanting MDs 350 randomly selected implanting MDs 3,500 patients followed 3 years for events and 5 years for survival 3,500 patients followed 3 years for events and 5 years for survival Based on 10% rate of appropriate therapy at 3 years (15% at 3 years in SCD-HeFT) Device therapy follow-up Device therapy follow-up Every 3 months for a minimum of 3 years Adjudication process Data combined with NDI and Medicare claims data

33. ICD Registry Version 2.0 – Updating the Registry CP1262561-17 Redefine the registry purpose, target audience Enhance the data collection forms Add leads -- implant, revision, replacement Make the registry a performance reporting tool (Guidelines) Post market surveillance (FDA Sentinel Network) Coordinate with longitudinal data – Medicare Claims Data, National Death Index Redefine the registry purpose, target audience Enhance the data collection forms Add leads -- implant, revision, replacement Make the registry a performance reporting tool (Guidelines) Post market surveillance (FDA Sentinel Network) Coordinate with longitudinal data – Medicare Claims Data, National Death Index

34. ICD Registry Research and Publications CP1262561-20 Research requests are reviewed and prioritized by R&P Subcommittee 34+ requests to date ICD-Registry provides financial support for data analysis Yale CORE assists with data analysis 6 abstracts presented at AHA, 11/07 5 manuscripts Research requests are reviewed and prioritized by R&P Subcommittee 34+ requests to date ICD-Registry provides financial support for data analysis Yale CORE assists with data analysis 6 abstracts presented at AHA, 11/07 5 manuscripts

35. ICD Registry Research and Publications CP1262561-21 Sample research proposals How do the baseline characteristics of patients receiving ICD therapy in the general population (real world) compare with the characteristics of patients enrolled in randomized clinical trials of ICD therapy? Are patient outcomes such as morbidity and mortality affected by patient baseline clinical characteristics such as ejection fraction, QRS duration, NYHA class, gender, age, and race? Sample research proposals How do the baseline characteristics of patients receiving ICD therapy in the general population (real world) compare with the characteristics of patients enrolled in randomized clinical trials of ICD therapy? Are patient outcomes such as morbidity and mortality affected by patient baseline clinical characteristics such as ejection fraction, QRS duration, NYHA class, gender, age, and race?

36. ICD Registry Research and Publications CP1262561-22 Sample research proposals What are the characteristics of the physicians implanting ICDs regarding training, experience, and volume; and how does this relate to implantation outcomes? Does age, race, and sex distribution of patients undergoing ICD implantation differ among different regions of the country and different size of hospitals? Sample research proposals What are the characteristics of the physicians implanting ICDs regarding training, experience, and volume; and how does this relate to implantation outcomes? Does age, race, and sex distribution of patients undergoing ICD implantation differ among different regions of the country and different size of hospitals?

37. Why a Registry? Science tells us what we can do; Guidelines what we should do; and Registries what we are actually doing Science tells us what we can do; Guidelines what we should do; and Registries what we are actually doing. Lukas Kappenberger MD HRS ICD Policy Conference Washington DC, 9/16/05 Science tells us what we can do; Guidelines what we should do; and Registries what we are actually doing Science tells us what we can do; Guidelines what we should do; and Registries what we are actually doing. Lukas Kappenberger MD HRS ICD Policy Conference Washington DC, 9/16/05