DOI: 10.1161/CIRCEP.114.002580

Alternative Research Funding to Improve Clinical Outcomes: The Model of Prediction and Prevention of Sudden Cardiac Death

Running title: Myerburg et al; Alternative Research Funding and SCD

Robert J. Myerburg, MD1 and Steven G. Ullmann, n, PhD2

1

Division of Cardiology, Caard rdio iolo io logy lo g , University U iversity of Miami Miller Un Milller School of Medic Medicine, ciinne, Miami; Mia i 2Center ffor tor M to anage geement and Policy, y,, University of Miami,, Co oraal Gables, FL Health Sector Management Coral

n en nden nd ence ce:: ce Correspondence: Myerburg, Myer erbu burg bu rg, MD Robert J. My Division of Cardiology (D-39) Professor of Medicine and Physiology American Heart Association Chair in Cardiovascular Research University of Miami Miller School of Medicine P. O. Box 016960 Miami, FL 33101 Tel: 305-5855523 Fax: 305-5857085 E-mail:[email protected]

Journal Subject Codes: [100] Health policy and outcome research, [8] Epidemiology, [5] Arrhythmias, clinical electrophysiology, drugs, [121] Primary prevention

1 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

Abstract Although identification and management of cardiovascular risk markers have provided important population risk insights and public health benefits, individual risk prediction remains challenging. Using sudden cardiac death (SCD) risk as a base case, the complex epidemiology of SCD risk and the substantial new funding required to study individual risk are explored. Complex epidemiology derives from the multiple subgroups having different denominators and risk profiles, while funding limitations emerge from saturation of conventional sources of research funding without foreseeable opportunities for increases. A resolution to this problem would have to emerge from new sources of funding targeted to individual risk prediction. In this analysis, we explore the possibility of a research funding strategy that would offer business incentives to the insurance industries, while providing support for unresolved goals. esol es olve ol vedd research ve rese re sear se arch ar ch go goal all The model is developed for the case of SCD risk, but the concept is applicable pplicab ab ble to to other othe ot herr areas he a ea ar e of the medical enterprise. al en ente nte terp rpri pri rise se. se

Key words: sudden cardiac risk s su s: sudd dden e cardiac car a di diac ac ddeath, eaath eath h, ca card rddia iacc ar aarrest, reestt, ri iskk sstratification, trattif ific icatio attion,, cost-effectiveness, cos ostt-ef effectiveness, ef fffeecttiv i en e es ess, s, research rese res esee methodology, research funding

2 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

Introduction Despite major advances in the prediction, detection and management of cardiovascular disorders during the past 40 years, cardiac arrest and sudden cardiac death (SCD) remains a major worldwide public health problem.1 The 40% reduction in age-adjusted cardiovascular mortality between 1990 and 20082 does not appear to have had the same impact on cumulative numbers and proportional contributions of cardiac arrest and SCD. 3 Current estimates of the SCD burden remain in the range of 45 to 50% of all cardiovascular deaths,4 accounting for 424,000 nontraumatic out-of-hospital cardiac arrests (OHCA) responded to by emergency erge g ncyy rescue sys systems ysste tem m annually, with a mortality rate of approximately 90 percent,5 accounting ng for or tthe h eestimated he sti st timat imatted 8 ,00 82, 0000 SC CD’ D s per per year in the U.S.4 Thiss is is ten times the th he annual an nnual mortality rate forr 300,000-382,000 SCD’s breast cancer, cer, and equiva ce equivalent aleent to o the he ann annual nua u lm mortality orta tality y rates rates es ffor or Al Alzheimer’s lzhheim mer’ss disease, diseaase, breast breaastt cancer, c cervical cancer, ncer, r, ccolorectal olorecttall cancer, di ol ddiabetes, iab abetes, HI ab HIV/ HIV/AIDS, V/AI V/ A DS AI S, prostate prosta pr tatte cancer, ta can a cer, as we well lll as ddeath eatth ffrom fro ro ,4, 46 guns, residential entia i l fires, fi car ar accidents, accid idents, and sui id suicide icid de co combined. mbin bi ed d.22,4,6 Moreover, M oreove veer, ind individual divid duaal ri rrisk sk

prediction contin continues tin es as an unresolved nresoll edd challenge hall ll tthat hatt iiss tthe he rate rate-limiting att li limiting miti iti step te for fo progress in i clinical practice. It is unlikely that major impact on individual risk prediction will emerge from anything short of large-scale research efforts to identify small subgroups at very high risk that are currently concealed within larger general population subsets at low cumulative risk. This analysis is designed to summarize the epidemiological and clinical factors that limit prediction and prevention of SCD, develop the rationale for a new approach to research funding, and propose a different resource for such funding. Analysis of the Problem The first principle for defining the SCD challenge derives from its complex epidemiology. The

3 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

total population at risk consists of several distinct subgroups, each having specific epidemiological and clinical characteristics.6,7 The 5 general categories into which SCD risk can be separated have different risk densities, with identifiable overlap among them. Historically, the categories have been classified as: (1) the general population, (2) a subgroup with known risk factors in the absence of identified cardiovascular disease, (3) those with cardiovascular disease and preserved cardiac function, (4) the class of patients with known heart disease and poor heart function, and (5) a category associated with confirmed or suspected genetic abnormalities associated with cause or risk of SCD.7 The category referred to as the general population literally represents resen nts t the the he to ttotal ota tall ppopulation, opuu op ge has has become beccom omee cconceptualized onceptualized as a largee po ppopulation, pulation, excluding ex xcluddin ing the other 4 subset but by usage subsets identified byy their their specific speciffic ris risk sk char characteristics. aracteeri ar ristticcs. A Available vaiilaablle st sstatistics atistiicss sug suggest ggesst thatt th the he ccumulat cumulative um mulat mul Ds emerging emerg rgin i g fr ffrom om the 4 specific speciifi ficc ca ccategories atego goriiess iiss a smal ssmall malll ffraction raction t off tthose hose occurr r number of SCDs occurring in the generall po ppopulation, p latiion, the pu th he latter latter being b ing do be dominated minat i ted by subj subjects bjectts wit bj without itho h ut identified ho identiified id d cardiovascular card rdio rd i vascc io disease or hi high gh h impact i t risk riiskk markers markers. ke T Therefore, Therefore he fo ffor o th the purpose p rpose off research h pl planning l nii and ndd development of preventive strategies, it is logical to consider the general population category as those individuals at unrecognized risk for cardiac arrest, often expressing as the first clinical manifestation of previously undiagnosed or unsuspected heart disease. In the second category, characterized by the presence of risk factors for development of disease, SCD risk is considerably higher than the segment of the population without risk markers, but still low enough to create a challenge for accurately predicting and preventing SCD in the individual subject. The risk factors in this category are used in various scoring systems and other clinical risk profiling strategies. The major risk factors, including diabetes, cigarette smoking, obesity, hypertension, obesity, diet, and sedentary life style, associate with increased

4 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

risk in this population subset based on the number of risk factors, and risk factor interventions offer benefit from a population perspective. However, individual effect size and prediction remain a challenge that should be a target of future research. The importance of identification within this subgroup is further highlighted by a recent observation that out-of-hospital cardiac arrest victims who are admitted to a hospital alive are increasingly likely to have only risk factors, as opposed to manifest disease.8 The complexity within both of these first two categories is that the number and power of the specific risk factors available for either groups or individuals establish gradients of risk, based on the sizes of the denominators minators and eventt ra rate rates t (Figure 1). Thee thi third recognized clinical risk hird hi rdd aand nd ffourth ouurt rth categories are those rec ecog cog o nized in clin nical all ppractice, ractice, in which ri profiling has dependent upon imperfect that as become become dep pendeent upo pon im po mpeerfectt ssets etss ooff clinical c in cl inic icaal markers ic markeerss tha h t have ha v eevolved ve volv ved d oover years. Thee finall cate ca category, ate tego gory, genetic genetic t determinants det eteermina er ant ntss off ri risk, iskk, include inc ncllude lude bboth othh rare arr ot arrhythmia rhy hytthmiia sy syndromes ndro do and the recognized cogn g izedd ffamilial ami mili liiall risk riiskk off card cardiac diac arrest as the thhe initial iniitiiall manifestation man nif i esta taattiion off he hheart artt di ddisease seasee in the generall pop population. llation ati ti 1 R Rare inherited inhherit itedd arrhythmias arrh rhh tthmias hmiia constit constitute titt ttee a categor ti category te that thatt stands tandd alone all as a specific component, while familial clustering of SCD risk offers the potential to contribute to risk profiling in the general population and categories of common causes of SCD. Each of the defined categories of risk has unique epidemiologic characteristics and clinical challenges, requiring different approaches to the translation of population risk to individual risk and interventional strategies. The relevant epidemiological targets are subclassified into conventional epidemiology, transient risk prediction, interventional epidemiology, and genetic/personalized risk prediction. Conventional epidemiology is best applied to the general population at risk for SCD as a first cardiac event in the absence of recognized structural heart disease. The concept of the epidemiology of transient risk refers to a strategy for

5 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

predicting triggering events in a timeframe that is useful for preventive actions. This highly desirable strategic target has been elusive, but is beginning to emerge with new screening technologies and biomarker profiling. Interventional epidemiology embodies analyses of the elements of clinical trial design and interpretation to maximize benefit of defined therapies and expand their application in new subsets of patients. Clues to future research opportunities for prediction commonly emerge unexpectedly from secondary analyses of clinical trial data. Genetic epidemiology refers to personalized risk identified at a molecular level. The various categories of population risk require further analysis sis in the context ooff th tthe pathophysiology of cardiac arrest itself. These include substrate-based d riskk de dderived eri rive ivedd fr ffrom rom m underlying str structural ruc uctu tura uraal disease d seease analysis, and expression-based di expressiion o -based risk de dderived rive v d from factors that hhave ve the potential into al to to convert a stable stab blee substrate sub ubbstrate te int nto ann active activ ive arrhythmogenic arrh ar rhyt rh y hm yt mogen nicc substrate. sub u straate ub t .9 S Substrate ubbsttra and kss intersect inte in tersectt with te ith thiin the various various ar eepidemiologic pide pi d miiollog de ogic icc strategies sttra rategi gies discussed gi discuss sssed above. abo b ve. expression risks within Expressionn factors in iinclude clludde neur neurohumoral ohhumorall contr contributions t ib ibutiions ttoo electrophysiological ellectrop opphy h sio io ologi gicall in iinstability, sttab abil i it il ity, y y, hemodynamic mic mi i variants ariants rii ts contributing contrib trib ib tting in to to cardiac rdi dia arrest stt risk, risk iskk markers rkk off electrophysiologic electroph el l tr h si siologic ioll ic instability, and the role of genetic variants in each of these pathophysiological subsets. Rationale for Seeking a New Approach to Funding It is evident from the forgoing that the challenge of SCD prediction and prevention is considerable, but the potential rewards are large. Although the cumulative SCD numbers have not changed substantially in recent years, there appears to have been a shift in the distribution of the population at risk among the various categories over time. The size of the population falling into more easily identifiable high risk subgroups has decreased, putatively as a result of medical and interventional progress. The improved outcomes after myocardial infarction in the era of acute interventions are exemplary.10 This has resulted in a shift in focus to the numerically

6 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

larger cluster of events that are diluted in much larger population denominators, creating lower population risk, but large cumulative numbers. Identification of small pockets of high-risk density within large population subsets offers the best opportunity for the reduction of the SCD burden, but constitutes the greatest challenge (Figure 1). The research models will differ for the first event subgroup, those with risk factors in the absence of defined disease, and those with defined disease with preserved cardiac function. In part, these are currently addressed as general preventive strategies in an attempt to achieve a population effect that is desirable and necessary, but not sufficient to have a major impact alone. Additional studies aree needed to identif identify fy pathways of investigative efforts that will yield high-resolution risk profiling identify targets ofiliingg tto o id iden entti tify ta tar of opportunity n ty nity y ffor o iindividual or ndiv nd i id dual risk prediction. Suchh studies stu udies will bee complex com mpl p ex and expensive.99,11 They will also alsoo require designs al deesiignns that thatt are are much m ch mu ch lon longer nger than th n cconventional onvvenntioonall sstudies on tud udiies to ud o yyield ieeld outtcome come data. co datta. However, H wever, Ho r w hen de deal aliing with al wiith a condition con ondditi on tion o that on thatt accounts acco couunts for co for one-half one-hh of meaningfull outcome when dealing all cardiovascular ascular ddeaths, eath hs, wit with i h half half lf off those emerging emergi gingg as the gi th he first fi manifestation m ni ma nife festtatiion off pprevious fe previously revi re vii vious unrecognized ed d ddisease disease, is oft often fte dduring ri ring i tthe he productive prod d cti ti e years ears off llife, life if the the jjustification stification tifi ti fi ti off such s ch h an investment appears appropriate. More than 10 years ago, the U.S. Department of Health and Human Services’ Agency for Healthcare Research and Quality (AHRQ) highlighted: “Patients at high risk of sudden cardiac death are hard to identify… accurate targeting is necessary for cost effective treatment.” 12 The realistic dilemma, however, is sources of funding.13 Traditional sources of research funding from the federal government and professional societies are not positioned to make new investments of this magnitude (Figure 2). Private philanthropy and foundations can help, but are usually limited to specific areas of focus and resources as well. The other major component of the research pool, the pharmaceutical, biotechnology and device industries, have contributed

7 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

substantially to the nationwide research commitment, accounting for more than 50% of the $110 billion U.S. biomedical research budget,14 while the NIH budget has recently remained constant in the range of $30 billion. However, in contrast to the past, the vast majority of industry support is currently allocated for product development and testing, and meeting regulatory requirements. Very little goes to the type of preclinical research that would improve pathophysiological pathway identification and population science for improved risk prediction, endeavors where the promises for future efficiencies are hidden. On the surface, hope for substantial increases in the type of support ppo p rt proposed propo p sed ap appe appears peaa pe dismal. But perhaps we have remained too focused on the notion that funding fundi ding di ng ooff bi biom biomedical omed edic ed ica ic research should hould houl uldd be ul b restricted res estric ictted to public and private sources ic souurces with a pprimary so rima mary research mission. If ma society hass so much to gai gain, in, and and cconventional onnvent ntio t onal ssources ouurcess off ffunding unddingg hav un have ve blun blunted, unted, oother thherr el elements lem me c id consid ider ered red in th the fu ffunding nding d g mi mix. should be considered Potential Impact I p ctt off Research Impa Rese Re s arch se h Investment Investtmentt by the t e Insurance th Insurance Industry I du In d stry try y Prior suggestions estions ti for fo expanding e panding ndi din the th research ch h ffunding nding di base b ha hhave a e incl iincluded l ded dedd establishment tabl bliishh ntt off ttaxadvantaged biomedical innovation trusts, preferential funding for new research institutes, creation of a new class of bonds dedicated to support discovery, and deferring patent protection to later in the discovery chain in return for funding.15 Another example, given only limited consideration, are the insurance industryies (health, disability, and life), which might have much to gain from improvements in prediction, prevention, and therapeutic efficiencies. The issues and opportunities can be considered individually for each component. Health Insurance The health insurance industry has multiple levels of exposure directly associated with cardiac arrest and SCD. Costs are generated for both OHCA and in-hospital cardiac arrest (IHCA)

8 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

survivors. First is the cost of immediate short-term care of hospitalized OHCA survivors and subsequent short-term rehabilitation. Among those who survive to hospital discharge, additional costs are generated for longer-term rehabilitation and chronic care facility costs where needed. Finally, insurers bear the cost of long-term strategies for prevention of SCD, primarily ICD’s, in population subsets at high risk for cardiac arrest. The latter include both SCD preventive therapy targeted to survivors of cardiac arrest (secondary prevention) and prevention of SCD within high-risk subgroups, based on risk profiling (primary prevention).16 Hospital costs for OHCA victims who are resuscitated and arrive ve at hospi hospital p tal aliv alive ve ar aare substantial. A recent meta-analysis of 140,000 individuals who experienced ienced ed d oout-of-hospital utt-o -off-ho h sppit ho itaa cardiac arrests, ests, est ts, s found fou ound nd that thaat the survival rate to hospi hospital ita tall admission was w s 23.8% wa 23.8% and the survival surviv rate to hospital discharge discharge wass 77.8%. .8% %. 17 A Applying ppplyin ng th these hese ra rat rates tes to o tthe hhee eestimated sttimaateed numbers numbe beers of of Emergency Em mergg r rdiac aarrest rrees rr es runs annually, est y 5 yyields ields an a eestimated stimated t d 1100,912 00,99122 hhospital ospit ital ad admi admissions missio mi i ns per yyear, Rescue cardiac with 33,072 72 patients patiients discharged disc di sccha h rg ged d alive ali live and li d 667,840 7 84 7, 8 0 dy dyin dying i g in-hospital. i -hhospi in pita pi taal. In another anoth her study, stu t dy dy, the tthhe direct d incident-related llated attedd cost stt off providing pro id idi iding in care to t a patient ti t experiencing e periencing rii cii an OHCA OHCA was as estimated sti timatted d to t be $102,017, 18 including acute care and early rehabilitation. Allocating proportional costs from the meta-analysis data to differences between survivors to discharge and those who died during hospitalization generates an annualized cost of $2.8 billion, based on the estimate of post-OHCA hospitalizations.5 An additional $3.01 billion is allocated to care in long-term facilities and $800 million is expended for initial ICD implants in OHCA survivors, assuming 80% of survivors receive devices. These figures are not inclusive of other post-cardiac arrest long-term therapies such as pharmaceuticals (ranging from $214-$470 per month) and related surgical procedures (ranging from $30,000-$200,000). An estimated additional 200,000 cardiac arrests occur inhospital (IHCA) annually, with a 2009 survival estimate of 22.1%. 19

The incremental costs are

9 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

difficult to define since a large proportion of IHCA’s are deaths associated with end-stage diseases or transient triggers for cardiac arrest. The other major cost center for insurers is that related to implantable device therapy for patients determined to be at high risk for a primary cardiac arrest, with or without heart failure, based on the major ICD and CRT clinical trials.(16) Data from the National Cardiovascular Disease ICD Registry (NCDR-ICD) indicate that approximately 110,000 devices were implanted in 2011, including both initial implants and replacements, 20 at a conservative estimated cost of $3.3 billion for devices alone; this is not inclusive of long-term follow-up w-upp and monitoring ng ccosts for device recipients. An addition $2.5 billion is estimated for lifetimee expenditures exppen endi dittu di ture ress re rrelated late la tedd to te ICD therapy p pr py pre prescribed escr es crib ibed ib e ffor or primaryy prevention indications, indi dica di c tions, based on li ca llifetime f time cost profiles fe attributed to to secondary secondary pprevention reeven ntionn IICD CD D rrecipients. eccip pienntss.18 Cu Cumulative umu mula lativve acu acute utee an and nd lon long-term ng--teerm m costs co s for e sur est urvi ur vivvors aand vi ndd pprimary rimary i y pprevention revent ntio nt ionn IC io ICD D reci ipi pien ents en ts, ts s therefore, theereffore, fa th ffar arr eexceed xceed d $$10 1 bi 10 bbillion i cardiac arrest survivors recipients, per year. Thee fforegoing oii direct dir t costs stt for fo post-cardiac post stt cardiac rdi dia management ntt are a major ajj part, partt bbutt not ott tthe he total, of the aggregate cost of cardiac arrests in the United States, placed by one estimated at $33 billion annually for OHCA.21 To put this in perspective, the entire cost of implementation of the Affordable Care Act, is estimated at $1.36 trillion from 2011 through 2023, or an average of $105 billion annually. So research that would potentially bring about a reduction of the costs of OHCA by 20%-50% could offset a significant portion of the costs – conceivably in the range of 10-15% - of the Affordable Care Act costs. Private insurance providers serving in commercial markets would be major beneficiaries of a reduction in health care costs allocated to short- and long-term OHCA and IHCA survivors, and better efficiencies in primary prevention strategies, helping the industry to remain viable in a

10 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

market place that has become increasingly stressed over the last few years. Further there is significant expansion of private insurance provision into the Medicare Advantage private market place and the State Medicaid programs as these government programs increasingly move toward contracting with private health insurance companies. Private health insurance companies currently cover just under 70% of the U.S. population, or over 220 million individual lives. 22,23 All have an interest in cost savings. The return on investment from a reduction in the incidence of cardiac arrest and its consequences would be significant. When analyzing cost figures, especially for expensive long-term m interventions, itt iiss important to recognize that the health care insurers inherit the limitations ability ons ooff ou ourr ab bil ilit ityy as it investigators physicians subgroups rs an andd ph hys y icia ianns to identify individualss or ia or small subgro r upps at high risk, and to prescribe expensive therapies funding xxpensive peensive therap apiess efficiently. effici cieentlly. ci y These Thesee llimitations imiitaatio onss result ressullt from m marginal m mar argina ar nal al research rese search ch fu patterns forr pastt pivotal pivvotall clinical pi cli linic i all trials trialls that th were wer e e uusually er suall lly sufficient ll suuff ffic iccie ientt to to adequately ad dequate tely te ly def define fine acce acceptable levels of efficacy, fficacy, fficacy y, bu bbutt no not ot optimal. opptimal. i l Most M st were underpowered Mo unde d rppowered d for sstratification t atif tr iffic i atio ti n off iindividual ndiv nd ivid iv idual risk id 16,24,25 ,24 24,25 25 efficiently.16 The llatter Th att tt iis generally generall ll viewed iie ed d as a metric triic off th the proportion ti off a ttreated r tedd

population that will achieve the targeted benefit – either absolute risk reduction (usually a small number) or the number needed to treat to prevent one event (often a relatively large number). Accordingly, one could make an argument that health care insurers, including both private carriers and CMS, should have an incentive to contribute to such research efforts. The rationale, while inclusive of good will and general welfare, can be based on business modeling alone. To the extent that investment in a higher level of individual risk prediction would lead to greater efficiencies in the expenditures of health care funds, the insurance system would achieve improvements in its business model. An example is reduced ICD expenditures based upon improved individual risk prediction. In this example, our limitation in the ability to risk profile

11 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

individual ICD candidates results in clinical Guidelines that lead to only 20-30% of recipients ever receiving appropriate delivered therapy. If we could improve delivered therapy from implanted ICDs to 50-60%, based on more precise identification of higher-risk candidates, health care savings would emerge from an appropriately reduced denominator of recipients. In addition, the research expense for any one project is time-limited, while savings accumulate year-to-year. Nonetheless, the cost for scientifically valid basic and clinical research addressing low incidence, high absolute number events is a major reason why health care insurance companies are investing in “big data,” The intent is to determine a pattern tern of use and/orr to rreduce health care risk and health care cost overtime, but there are concerns about boutt iits t llimitations ts imitattio imit i ns for fo accurately ide identifying ent ntif ifyi if ying yi n eff ng efficacy. ffic ff i acy. Valid clinical effic ic efficacy cacy measures are are more mor o e likely to emerge from “big science.” accommodation conventional ce.”” In large ppart, ce arrt, “b “big ddata” ata” iiss an an acc commo modaation tion tto o th the status sttattuss quo quo of of co onven ntioo research funding. u ng unding ng. With expenditures t annual th annuall health heaaltth care exp pendditures in i the range g off $2 $$2.7 .77 ttrillion r ll ri llio ionn in io in the h U.S., U.S S., and and a federal research approximately earch ch h ffunding nding di bbudget dget d t off appro iimatel tell $30 $30 billion, bbillion illi il li a 1% set-aside set ett aside id for fo research ch h targeted to generate better efficiencies would provide an additional funding pool equivalent to the current NIH budget. The magnitude of this set-aside example is far in excess of what is needed or justified for the model case of cardiac arrest and SCD, which was chosen as a demonstrative of what could be done from the perspective of a personal interest in SCD. It is self-evident that other disease states, not limited to the cardiovascular enterprise, with similar prediction and prevention dilemmas should share in such a funding pool. Health insurance industry contributions for this type of research are uncommon, although there are examples of support for research targeted to compliance and secondary outcomes.26 The strategies proposed for the Patient-Centered Outcomes Research Institute for implementation of

12 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

comparative effectiveness research, funded by an ACA-mandated fee on covered lives, is another form of industry involvement. Although patient-centered research goals address efficiencies, they do so from a perspective that derives in large part from existing data and guidelines, with their inherent limitations. This proposal focuses more exclusively on generation of new knowledge that would cycle back through patient-centered analysis. Substantial support is envisioned for new, in-depth studies focused on high-resolution individual risk stratification for the purpose of achieving therapeutic and preventive efficiencies that are currently beyond our capabilities. Disability Insurance To the extent are successful, e t that ent th hat a currently cur u rent nttly l available responses to OHCA OH succe c sssfu f l, a population of survivors are disabilities. are ccreated, ar reated, a ssignificant ig gnifficcantt pproportion ropo portio ion ooff whom m have h ve llong-term ha onng-term m neurological neu eurolo ogi gicaal dis sabbill Less than 10% functional status. 1 % of survivors survivors are able to t return n to fully fullly l normal norrmall fu func cti tionall statu tuss. It has tu has bbeen een estimated that survivors t approximately app pproxiimaately pp l 226% 6% off survi ivors hhave ave severe neurological neurol olloggicaal iimpairment mpa p irment nt ((CPC C C CP scores 3-4), ) 27 most stt off whom hhom die die within ithi it ithin hi 6-12 6 12 months months. th Among Am those th with iith th CPC CPC score 1-2, 1 2 long llo o term survival is considerably better, but return to normal function is still limited, primarily for those remaining in CPC category 2. Disability costs accumulate primarily for those in CPC 2-4. At a cost of approximately $130,000 for 30 days of rehabilitation and 365 days of long-term facility care,18 this extrapolates to $1.23 billion nationwide per year, assuming a 1 year mean survival of those with severe disabilities. Costs for those with moderate disabilities are not included, nor are those on long-term respirator or feeding support. For in-hospital cardiac arrest, a survival estimate of 22.1% in 2009 was associated with 28.1% of the survivors discharged with significant neurological deficits (CPC >1), or 6.3% of the total cardiac arrests. Extrapolation from the participating hospitals to an estimated U.S. total of

13 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

200,000 in-hospital cardiac arrests per year adds 12,533 subjects to the significant disability population pool annually,19 at an estimated annual cost of $1.63 billion. Research contributions to improve response systems and post-resuscitation care would offer the opportunity to reduce the pay-out burden on disability insurance providers. Although the total $2.86 billion is smaller than the cost accumulated in the health care insurance model, disability insurance is a less complex model than health and life insurance, requiring fewer variables in the analysis of the relationship between premiums, research contributions, and financial benefits. There is an added component to the disability model that affects ts employees, emp ploye y es,, as well wel e l as the costs of disability that are passed on to the employer. It is estimated d tha that at an employer emp mplo l yeer incurs lo in in a cost 1.75 times imes ime es the the earnings e rnin ea inggs of the affected worker, in r rrelated e ated to the co el ccost stt of of replacing the indiv individual. There are an an eestimated stimated 39, 39,000 ,0000 to 48 48,750 8,750 0 ind individuals ndivid duals who whoo die diee in in the th workplace woorkpplace from from o cardiac cardi diacc di ppsychic syc ych yc chic im iimpact pactt on surv survivors rviv rv iv in ivors n tthe he workp workplace k la lacce aand ndd tthe he nnet ett imp impact pac actt on survi surviving survivin ivin i arrest.5,28 The ps workers performance e erformance iiss si sign significant. g if ifiicant. Life Insurance There would appear to be a business advantage for delaying deaths for life insurance policy holders, allowing companies to earn off of retained funds for longer periods of time. Although the same advantage might not hold for their expanding annuity businesses, it is notable that the life insurance industry did create a Life Insurance Medical Research Fund in 1945. 29 The fund provided support primarily for the generic type of science envisioned here, plus training grants, and structured the program as a foundation, led independently by prominent clinical scientists. The research fund appears to have ceased functioning in the early 1990’s. So the concept is not new, but is worthy of re-exploration in light of current funding dilemmas. It might be argued that successful risk profiling for first-event SCD’s would be a

14 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

disincentive to insurers to issue policies to individuals populating the newly-identified high-risk subgroups. Whereas a person cannot be denied health care insurance under the ACA, nor can they be charged a higher premium for higher risk, there is no such requirement placed on life insurance companies. This disincentive argument is countered, however, by the parallel goal of early interventions for preventing premature deaths once such subgroups are identified, providing a net benefit to the insurer and the general population. Reducing first-event SCDs from 50% to 20% by effective routine risk profiling earlier in life could provide a broad-based societal benefit, encompassing the realms of health care cost efficiency, y, disabilityy payments p ym pa men entt to cardiac arrest survivors with neurological deficits, and premature life insura insurance payouts, ranc ncee pa payo yout uts, s while s, wh also preventing 150,000 nting ntin ng the the annual a nuuall unanticipated prematuree lloss an oss of some pr pproportion oppor o tion of up to 150,00 people, many any still in their an theiir productive prod oductiive years. od yea earss. Revisiting the Ph Pharmaceutical, Biotechnology P haarmaceutic ti al, l Bi iot oteechnol ollogy ogy and og d Medical Med ediical Device Dev eviice IIndustries ndusstr triies In additionn to the insurance medical insuran an nce industries, industrie d i s, the h pharmaceutical, pharmaceutiicall, bbiotechnology i tech io ch hnolo logy lo g andd med gy diccal a ddevice devi evi industries may returning the scope off ddirect ma also ls gain aii ffrom r ret ett rning nii tto contributions contrib nttrib ib ti tions for fo research ch h bbeyond be ond d th product development. Extending the example of ICDs cited earlier, implants appear to be decreasing based upon effective interventions for acute coronary syndromes and other therapies. If they were to decrease further for the common Guidelines-based indications as a result of better risk profiling, the device business would be further reduced – an apparent disincentive. However, this loss might be offset by new indications emerging in parallel by developing insights into pockets of high risk in the general population that currently contribute such large numbers to first-event SCDs. The dilemma of a low incidence concealed in a large denominator translates to large numbers that may identified within subgroups of high risk density (Figure 1).9 The net effect over time can be positive for both the corporations and the public.

15 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

Interface between Business Interests, Cost Effectiveness, and Clinical Benefits The interface requires a model that would be acceptable to the various segments of the insurance industries, investigators and universities, government, and the public. The previously suggested creation of research trusts14 is a good starting point, but does not mention sources of funding, independent of tax benefits. The term “set-asides,” mentioned previously in this statement cannot take the form of a federal tax on industry for support of expanding research; that does not appear feasible in this era. In contrast, if the industries see business merit to the concept, a direct infusion into carefully structured, tax-advantaged independent entities,, dedicated to prediction preedi dict c and prevention research, would likely be the most palatable funding and nd ma mana management nage geme mentt m m. mechanism. reserve credibility crediibility y andd be be compliant co omp plian nt with with h thee law, law aw,, such suuch entities enntiitiees would wouuld have ha too bbe To ppreserve t tur truct ured ed ffor or iindependence, nddep epende d ncee, overs sig ight ht andd transparency, ht, tran nsp spar areency ar cy y, as ddiscussed i cuss is ssed ss ed elsewhere. els l ewhe h re.30 For carefully structured oversight, example, in n ad addition dditiion to a charter ch harter de ddefining fiiniingg iintent ntent and d orga organization, g niization, th the h tr trusts rusts might mig i ht ig ht consider con onsi siderr si creating institutes, and stit titt tes ti te having ha h ing in an Executive E ec ti e C Committee mitt itt for f policy polic oli lic oversight o ersight siight ht andd adherence, adh adherence dhe ndd one or more Research Evaluation Committees (study sections), with final decisions recommended by an Awards Committee. In contrast to a recent study supporting the long-held view that public research funding contributes to the general economy beyond the confines of academic research institutions,31 the arguments provided here are largely developed from the perspective of a combination of private and public business considerations and the potential for public health benefits. Given that the costs of cardiac arrest to the insurance industry are so large, and significantly outstrip the death rate and associated costs of so many other health and insured events, it is curious that research is not supported significantly in the health insurance sector. In other sectors of the insurance

16 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

industry, consortiums of individual insurance companies do support research to save life and limb, and resultant costs. For example, the Property Insurance Research Group is part of the Fire Protection Research Foundation whose mission is to recommend research projects, and is funded by property insurance company contributors.32 Regarding research into the prevention of auto accidents, the Insurance Institute for Highway Safety and the Highway Loss Data Institute focus on human loss and economic loss, and undertake research funded by over 100 auto insurance companies and insurance foundations.33 Feasibility would require reorientation of some existing business ess ppractices. ractices. Thee ttendency e en to focus growth and performance in terms of quarter-to-quarter or year-to-year r-to-y yea earr wo woul would ulld ha hav have ve to ve adjust to mult m multiple tip iple le yyears e rs oorr even a decade or more too complete ea complete appropriate appr p opri riaate studies, impleme ri implement e results and rrealize ealize health h aand nd bbusiness usin ness ga gain gains. ns. Inn addition, addiitiion, an in industry-wide ndus ustry-w us wide dee pool pooling ling strategy strategy, sttratteg gy 2 similar to the h model he mod odel el developed develop loped d bby y th the he li life ins insurance nssur uran a ce iindustry nddus ustr try in the tr thee 1940’s, 194 940’ 0 s, 29 would w oulld be be required. requu

It is unrealistic istic to expect expe p ct the the h contributions contrib ibuttions off a few ib few corp corporations porati tions ttoo fund f ndd the fu thhe development devellopme oppme mennt of o a product that its nature all, att bby it natt re would o ld trickle triickl kle iinto ntt the th public p bl bli blic ic domain d aii andd benefit be fi fitt all ll even e en if iitt were er taxadvantaged. 14 Contributions proportionate to corporate size or market share might be principles for compromise. But, as stated above, there is a strong precedent for insurance company consortiums to support research in other sectors of the economy. Finally, setting the business element aside, and thinking of the parallel public health and medical practice gains, the initiatives proposed offer the hope of doing more good for our patients and their families, and addressing, at least in part, the recent emphasis on the needs and directions of relevant clinical research going forward.34,35 Conclusion The aim of this discussion is to initiate a conversation on a proposal for new methods of funding

17 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

critically important research that may impact at both population and individual patient levels. Feasibility, considering the complexities of the health care system and related industries, requires in-depth collaborative analyses by the various interested stakeholders who have access to relevant data and system designs. The creation of an ad hoc national commission to evaluate and model such systems, or an exploratory study by the Institute of Medicine, would be appropriate next steps. The alternative, maintaining the status quo, would leave us locked into the current funding circumstances, facing the consequences of opportunities lost.

Acknowledgments: The authors appreciate the review of this manuscript comments criiptt an andd co comm mmen mm ents en ts bby ts Donna E. Shalala, PhD, Peter J. Schwartz, MD, Hein J. J. Wellens, MD, Fred Lindemans, PhD, D, Fr red L i de in dema mans ma ns,, P ns and Michael supported el R. Rosen, Ros osen en, MD. en MD. Dr. Myerburg is suppor MD rte ted in part by thee American American Heart Association University Miami n Chair Ch in Cardiovascular Carddiovascular Research at thee U niversity ooff Mi iami Miller School of Medicine. Conflict off Interest Disclosure: None Inte teere rest st Dis i closure: l N No ne References: s s: 1. Myerburg RJ, Castellanos A. Cardiac Arrest and Sudden Cardiac Death - Chapter 41. In: Bonow RO, Mann DL, Zipes DP, Libby P, editors, Braunwald’s Heart Disease: A Textbook of Cardiovascular Medicine, 9th edition. Oxford, UK. Elsevier, 2010 2. National Institutes of Health, National Heart, Lung, and Blood Institute: Morbidity & Mortality: 2012 Chart Book on Cardiovascular, Lung, and Blood Diseases. Accessed September 6, 2014 at http://www.nhlbi.nih.gov/research/reports/2012-mortality-chart-book.htm 3. Roger VL, Go AS, Lloyd-Jones DM, Benjamin EJ, Berry JD, Borden WB, Bravata DM, Dai S, Ford ES, Fox CS, Fullerton HJ, Gillespie C, Hailpern SM, Heit JA, Howard VJ, Kissela BM, Kittner SJ, Lackland DT, Lichtman JH, Lisabeth LD, Makuc DM, Marcus GM, Marelli A, Matchar DB, Moy CS, Mozaffarian D, Mussolino ME, Nichol G, Paynter NP, Soliman EZ, Sorlie PD, Sotoodehnia N, Turan TN, Virani SS, Wong ND, Woo D, Turner MB; American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics--2012 update: a report from the American Heart Association. Circulation. 2012;125:e2-e220. 4. Fishman GI, Chugh SS, Dimarco JP, Albert CM, Anderson ME, Bonow RO, Buxton AE, Chen PS, Estes M, Jouven X, Kwong R, Lathrop DA, Mascette AM, Nerbonne JM, O'Rourke B,

18 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

Page RL, Roden DM, Rosenbaum DS, Sotoodehnia N, Trayanova NA, Zheng ZJ. Sudden cardiac death prediction and prevention: report from a National Heart, Lung, and Blood Institute and Heart Rhythm Society Workshop. Circulation 2010;122:2335-2348. 5. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha MJ, Dai S, Ford ES, Fox CS, Franco S, Fullerton HJ, Gillespie C, Hailpern SM, Heit JA, Howard VJ, Huffman MD, Judd SE, Kissela BM, Kittner SJ, Lackland DT, Lichtman JH, Lisabeth LD, Mackey RH, Magid DJ, Marcus GM, Marelli A, Matchar DB, McGuire DK, Mohler ER 3rd, Moy CS, Mussolino ME, Neumar RW, Nichol G, Pandey DK, Paynter NP, Reeves MJ, Sorlie PD, Stein J, Towfighi A, Turan TN, Virani SS, Wong ND, Woo D, Turner MB; American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics--2014 update: a report from the American Heart Association. Circulation 2014;129:e28-e292. 6. Myerburg RJ, Castellanos A. Emerging paradigms of the epidemiology and demographics of sudden cardiac arrest. Heart Rhythm 2006;3:235-239. 7. Wellens HJ, Schwartz PJ, Lindemans FW, Buxton AE, Goldberger JJ, Ho Hohnloser Huikuri H hnlloser SH SH, H u HV, Kääb S, La Rovere MT, Malik M, Myerburg RJ, Simoons ML, Swedberg wedbber ergg K, K Tijssen Tijjss Ti ssen e J, en J Voors AA, Wi Wilde death: current challenges Wild ldee AA ld AA. Ri Risk s stratification for sudden en ccardiac ardiac death h: cu urr rrent status and challe e for the future. 2014;35:1642-1651. ure ure re. Eur Heart Hearrt JJ.. 20 He 2014 14;3 14 ; 5: 5:16 1642 16 42-1 2-1 -16551. 8. Wong MK M MK, K, Mo M Morrison rrisson o L LJ, J, Q Qiu iu F,, A Austin usttin P PC, C, Cheskes Cheske Ch kes S S,, Dorian Dorriaan P,, Scales Sc es DC, C,, Tu Tu JV, JV, Verbeek Verr PR, Wijeysundera sundera sund der eraa HC, HC Ko K DT. DT. Trends Trend ndss iinn shortnd sho hort ho rt- andd llong-term ong ng-term t m survival suurv rviv i all amo among ong out out-of-hosp out-of-hospital t-off ho h sp cardiac arrest e ppatients est atie at ient ie ntss alive nt aliv al ivee at hhospital iv ospi os pita pi tall arrival. ta arrri riva val. va l. Circulation C rccul Ci ulat atio at ionn 2014;130:1883-1890. io 201 014; 4;13 4; 130: 30: 0:18 1883 18 83-1 83 -189 -1 890. 89 0 0. 9. Myerburg r RJ, Junttila rg Junntt ttil ilaa MJ. il MJ. Sudden Sudd Su dden dd en cardiac car ardi diac di ac ddeath eath ea th ccaused auseed by ccoronary au o on or onar aryy he ar hear heart artt di ar dise disease. seas se a e. Circu Circulation u 2012;125:1043-1052. 11043 0433 10 04 1052 52 10. Puymirat E, Simon T, Steg PG, Schiele F, Guéret P, Blanchard D, Khalife K, Goldstein P, Cattan S, Vaur L, Cambou JP, Ferrières J, Danchin N; USIK USIC 2000 Investigators; FAST MI Investigators. Association of changes in clinical characteristics and management with improvement in survival among patients with ST-elevation myocardial infarction. JAMA. 2012;308:998-1006. 11. Goldberger JJ, Basu A, Boineau R, Buxton AE, Cain ME, Canty JM Jr, Chen PS, Chugh SS, Costantini O, Exner DV, Kadish AH, Lee B, Lloyd-Jones D, Moss AJ, Myerburg RJ, Olgin JE, Passman R, Stevenson WG, Tomaselli GF, Zareba W, Zipes DP, Zoloth L. Risk stratification for sudden cardiac death: A plan for the future. Circulation. 2014;129:516-526. 12. Treatment to Prevent Sudden Cardiac Death: Clinical Highlights. May 2003. Agency for Healthcare Research and Quality, Rockville, MD. Accessed September 6, 2014 at http://www.ahrq.gov/research/findings/factsheets/coronary/sudden/index.html 13. Lauer MS, D'Agostino RB Sr. The randomized registry trial--the next disruptive technology in clinical research? N Engl J Med. 2013;369:1579-1581.

19 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

14. Moses H 3rd, Dorsey ER. Biomedical research in an age of austerity. JAMA. 2012;308:23412342. 15. Moses H 3rd, Martin JB. Biomedical research and health advances. N Eng J Med. 2011;364:567-571. 16. Myerburg RJ, Reddy V, Castellanos A. Indications for implantable cardioverter-defibrillators based on evidence and judgment. J Am Coll Cardiol. 2009;54:747-763. 17. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out-of-hospital cardiac arrest: a systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes. 2010;3:63–81. 18. Merchant RM, Becker LB, Abella BS, Asch DA, Groeneveld PW. Cost-effectiveness of therapeutic hypothermia after cardiac arrest. Circ Cardiovasc Qual Outcomes. 2009;2:421-428. utcomes. 2009;2:4 421 2119. Girotra S1, Nallamothu BK, Spertus JA, Li Y, Krumholz HM, Chan Heart an PS; S American Ameriican He Hea Association Get with the Guidelines–Resuscitation Investigators. Trends Survival nds in nS urvi ur viva vi vall after va afte af terr Inte I Hospital Cardiac Arrest. 2012;367:1912-1920. a di ardi diac ac A rres rr e t. N Engl J Med. d 2012;367:19 19112-1920. 19 20. Kremers MS, Hammill SC, Berul Koutras Curtis Blum rs M rs S, Hamm mill SC C, Be eruul CI CI, Ko K outrrass C, C urrtis tis JS, JS, Wang JS W ng Y, Wa Y, Beachy Beachy Be hy J, B lum m Meisnere L, M,, R Reynolds Al-Khatib SM, L Co Conyers de ddell M eynoold lds MR, MR Heidenreich Heid idenrreiichh PA PA, A l--Khattib bS M, Pina Pin na IL, IL Blake Bllakee K, Norine Walsh M,, Wi Wilkoff Masoudi FA, National Registry a M alsh Wilkof ff BL BL, Shalaby S allab Sh abyy A, M asou o di F A, Rumsfeld Rum umsf sfelld J. sf J. The Th Na Nati tioonall IICD ti CD R eg g Report: version 2011. Heart Rhythm. r n 2.1 rsion 2.1 including incl in clud cl udin ud ingg leads in lead le adss and ad and pediatrics pedi pe diat di atri at riccs ffor ri or yyears ears ea rs 22010 0100 an 01 andd 20 2011 11. He 11 Hear artt Rh ar Rhyt ythm yt h . 2013;10:e59-65. 5 65 59-65 65. 21. Kida K K, Ichinose F. P Preventing NO Ichi hi F Pre r enting ti ischemic i hemii bbrain raii injury iinj njj r after ft sudden s ddden de cardiac rdi dia arrest stt using sing sii N inhalation. Crit Care. 2014;18:212. 22. New Census Estimates Show 3 Million more Americans had Health Insurance Coverage in 2012. Accessed September 6, 2014 at http://aspe.hhs.gov/health/reports/2013/CPSIssueBrief/ib_cps.cfm 23. Henry J. Kaiser Family Foundation: Medicare Advantage Fact Sheet, May 2014. Accessed September 6, 2014 at http://kaiserfamilyfoundation.files.wordpress.com/2014/05/2052-18medicare-advantage.pdf 24. Myerburg RJ, Mitrani RD, Interian A, Castellanos A. Interpretation of outcomes of antiarrhythmic clinical trials: Design features and population impact. Circulation. 1998;97:15141521. 25. Myerburg RJ. Implantable cardioverter-defibrillators after myocardial infarction. New Engl J Med. 2008;359:2245-2253. 26. Choudhry NK, Avorn J, Glynn RJ, Antman EM, Schneeweiss S, Toscano M, Reisman L, Fernandes J, Spettell C, Lee JL, Levin R, Brennan T, Shrank WH; Post-Myocardial Infarction

20 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

Free Rx Event and Economic Evaluation (MI FREEE) Trial. Full coverage for preventive medications after myocardial infarction. N Engl J Med. 2011;365:2088-2097. 27. Graves JR, Herlitz J, Bång A, Axelsson A, Ekström L, Holmberg M, Lindqvist J, Sunnerhagen K, Holmberg S. Survivors of out of hospital cardiac arrest: their prognosis, longevity and functional status. Resuscitation. 1997;35:117-121. 28. Ambrose R. “The Cost of Cardiac Arrest in the Workplace,” Lecture 1, 2004. Accessed September 6, 2015 at: ohsonline.com/Articles/2004/09/The-cost-of-cardiac-arrest-in-theworkplace.aspx 29. Atwater RM. Life Insurance Medical Research Fund-Third Annual Report, 1947. Amer J Pub Health. 1948;38:1597-1598. 30. Myerburg RJ, Feigal DW Jr, Lindsay BD. Life-threatening malfunction of im implantable mpl plan an cardiac devices. N Engl J Med. 2006;354:2309-2311. 31. Weinberg g BA, Owen-Smith J, Rosen RF, Schwarz L, Allen BM, Weisss RE, RE Lane Lane J. J Research Re funding. Science c en cien ence ce ffunding unndi d ngg and and short-term economicc ac aactivity. tivity. Scienc Science n e 2014;344:41-43. 2014;344:41-43. 32. Property ty y Insurance Ins n urance R Research eseaarcch Gr Group." roup."" Na N National ationaal Fire Firre Pr P Protection otec ot ection on Association. Assoocia iattion. N.p., ia N.p.,, n.d n.d. d. W Web. 18 July 2014. 14. 14 http://www.nfpa.org/~/media/Files/Research/Research%20Foundation/PIRGStatementOfPurpos w.n pa. w.nfp a or org/~/medi media/Fi /Fi Files/Re les/R sear search h/R /Research% earch% h%20 20F 20 Founda d tion/P on/P /PIR RGS GSttateme mentOf tOf OfP P e.pdf . 33. About the t Institutes." Institut utes ut es." es ." IIHS-HLDI: IIIHS-H IIHS HS-H -HLD HLD LDI: I: Crash Cra rash sh Testing Tes esti ting & Highway ting Hig ighw hway hw ay y Safety. Saf afet ety. et y. N.p., N.p .p.,. n.d. n.d. .dd Web. 188 July 2014. http://www.iihs.org/iihs/about-us. // iiih ih //iih iihh / b t ii 34. Antman EM, Harrington RA. Transforming clinical trials in cardiovascular disease: mission critical for health and economic well-being. JAMA. 2012;308:1743-1744. 35. Lauer MS. Rigorous science as the road to better public health. Popul Health Metr 2013;11:10.

21 Downloaded from http://circep.ahajournals.org/ at New York University/ Medical Center--New York on February 10, 2015

DOI: 10.1161/CIRCEP.114.002580

Figure Legends:

Figure 1: Risk density subgrouping within the general population. The mean SCD risk in the general population over 4 decades is demonstrated. The mean risk is