Address correspondence to: William T. McGivney, PhD, Chief ExecutiveOfficer, National Comprehensive Cancer Network, 500 Old York Road,Suite 250, Jenkintown, PA 19046; email@example.com.
S510 THE AMERICAN JOURNAL OF MANAGED CARE DECEMBER 2005
Cancer, in its various types and presentations,is the number 2 killer inthe United States.1 This unfortunatecircumstance and position will only be fortifiedas the median age of our populationincreases. While cancer knows no exclusivetarget, those individuals older than 55 yearsof age constitute the majority of Americanspresented with a diagnosis of cancer everyyear.2 For the United States, for majoremployers, and for managed care companies,access to and availability of cancer careitself and the attendant costs of such carewill present an increasingly substantial managementchallenge. This monograph willreview clinical issues, coverage and reimbursementprocesses, evolving quality-ofcaremodels, and the emergence ofconsumer-driven healthcare as they relateto the cancer care challenge.
Factors Making Cancer Care a Priority
Clearly, a higher prevalence of cancerdiagnoses in the population at large in thecoming years affirms the disease as amajor focus for healthcare managers in boththe public and in the private sectors. Anaging population is not the only reason thatcancer care will impact the healthcare systemin terms of quality, effectiveness, andcost of care.
Research pipelines are replete with innovativedrugs and biologics that hold thepromise of more direct targeting of cellularstructures, nucleosides, and proteins thatcontrol signals for a cancer cell's life cycle,its proliferation, and its death. These drugsand biologics hold promise for patients buthave significant price tags associated withtheir acquisition and use. These agents tendnot to replace existing chemotherapeuticregimens. Rather they are add-ons to suchregimens because of distinct and more selectivemechanisms of action. Additionally,these agents are approved by the US Foodand Drug Administration (FDA) for a singleindication, but, in many cases, have therapeuticpotential for use against an array ofcancer types. Finally, the advent of morespecific and targeted biologic agents providesa more favorable safety/adverse eventprofile than the traditional armamentariumof cytotoxic agents. As such, many cancersare being transformed into chronic diseaseswhere biologic agents are given until diseaseprogression and even through progressionof disease. Taken together, thesefactors presage cancers as diseases that willbe treated more effectively but that willchallenge our ability to pay for such bettertreatment.
Prevalence, Incidence, and Costs
In the United States every year, 1.372million people are diagnosed with cancer.2Of that figure, approximately 76% of newdiagnoses are for people 55 years of age andolder.2 The most prevalent diagnosed cancertypes are prostate cancer, breast cancer,lung cancer, colon and rectal cancers, andmelanoma.1
The total direct costs of cancer care in theUnited States on an annualized basis are $70billion.3 Indirect morbidity costs, or the costof lost productivity due to illness, are $16.9billion, while indirect mortality costs, orthe cost of lost productivity due to prematuredeath, are $103.5 billion.3 Estimates ofnational expenditures for medical treatmentfor common cancers, based on cancer prevalencein 1996 and cancer-specific costs for1995-1998, expressed in 1996 dollars are asfollows: breast cancer—$5.4 billion; colorectalcancers—$5.4 billion; lung cancer—$4.9 billion; prostate cancer—$4.6 billion;melanoma—$0.7 billion.4
The New England Journal of
The rising prices of chemotherapy agentscomprise an increasing percentage of thenation's healthcare expenses. In her commentaryin , Dr Deborah Schrag contemplatesthe impact that the costs of novel treatmentsfor colorectal cancer are having on patientsand their families, the oncology community,and public and private insurers.5 As advocatesfor their patients, physicians often findthemselves in difficult positions as patientslook to them for recommendations regardingwhat treatments to pursue. These recommendationsbecome more complicatedwhen the clinician considers the overallcosts of care and the amount to be coveredfor a particular therapeutic regimen by thepatient's private insurance company. Asmore research progress is made and innovativedrugs and biologics are discovered totreat diseases, all constituents of the cancercare community increasingly will be confrontedwith the challenge of determiningwhich chemotherapeutic regimen is best foreach individual patient, considering the clinicaloutcomes and cost.
Historical Background: Cancer Careand Managed Care
The early 1990s represented a time ofintense contentiousness between the practicingoncology community and managedcare medical directors. Cancer care becamea battleground as managed care companiesattempted to demonstrate the value of the"managed care concept" to major employerswho were both pushing for and paying forsuch services. The focus of managed careactivities in the 1990s was on reining in doubledigit annual increases in healthcareexpenditures. One principal mechanism forachieving this was control of the utilizationof healthcare technology (drugs, devices,procedures) through the establishment ofcoverage policies and often proactive implementationthrough preauthorization requirementswith attendant qualification criteria.With preauthorization, coverage policywas implemented on an individual case levelthrough medical necessity determinations.In cancer, a serious and life-threatening diseasearea, denials of care through theseprocesses became a lightning rod for confrontation,adverse publicity, and majorlawsuits.
Issues of the 1990s: Bone Marrow Transplantand Use Beyond the FDA Label
Managed care companies faced 2 majorissues and attempted to address them boththrough the processes just outlined. The firstissue was the advent and widespread utilizationof a very expensive extension of traditionalchemotherapy, that is, high dosechemotherapy with autologous bone marrow/stem cell transplant (HDCT/ABMT/PSCT).The therapeutic strategy was to achieve totalor extremely high percentage cancer cellkill by purposely providing doses ofchemotherapeutic agents in regimens toxiceven to "normal" cells, including progenitorred and white cells in the marrow.6 This technologyhad been more fully developed in thesetting of hematologic malignancies and wasbeing applied broadly in solid tumors, especiallyin breast cancer. The focus of the therapywas advanced disease such that theattendant risks, including death, were deemedreasonable.
The second issue was the increased focusbeing paid by public and private insurers touse-beyond FDA-approved labeling. Whilethe costs pale in comparison (on an inflationadjusted basis) to some of the price tags weare facing for interventions with chemo/bioregimens today, insurers imposed preauthorizationschema on these same drugs and, infact, began denying coverage for some drugsand some regimens on the basis that the proposedindication constituted an "investigational" or "experimental" use of such agentor such regimen.
The issue of HDCT/ABMT/PSCT becamean intense battleground for 2 reasons. First,the charges associated with provision of thefull range of services were often in the rangeof $150 000 to $250 000 in the early 1990s.Even as payors were growing accustomed toconsidering these charges, requests for repeattransplants or tandem transplants wereoffered up. Second, most institutions offeringthis service were unwilling to proceed andinvest the necessary resources without firstachieving agreement by the payor to coverthe continuum of such services. Thus, theprovider side actually established a preauthorizationprocess of their own thatinvolved upfront notification of the payorbefore the services were provided.
Payors set up processes based upon clinicalinclusion criteria to evaluate requests forcoverage of HDCT/ABMT/PSCT. Additionally,outside review processes were put inplace to involve "unbiased third partyexperts" to evaluate and provide opinionsabout the appropriateness of treatmentrecommendations. Most requests were approved.Some denials were upheld. Butmany denials ended up featured on and local news shows and in court(Fox vs. HealthNet, Goodrich vs. Aetna) withresultant judgments against insurers of$90 million and $120 million, respectively.7 In the late 1990s, both employers, aspurchasers of care for their employees,and payors, as financially liable for theircoverage/medical necessity determinations,backed off tough stances and denials oftreatments for cancer. The philosophy ofmanaged care/insurers transitioned to a"hands-off" policy for such cases.
United States Pharmacopeia
Formulary Service Drug Information
The use-beyond FDA-approved labelingissue followed in the draft of the backing offof the HDCT/ABMT issue. Upwards of 40states and the federal government adopted legislationthat was translated into regulationand ceded authority on decisions on appropriatenessof off-label use to 2 established drugcompendia, the and the .
The processes of oversight described inthis section provide a flavor of the mechanismsbrought to bear and the tensions thatarose from same as payors attempted to reinin the rate of rise of healthcare expenditures.The major impact and success thatpayors had was in negotiating down reimbursementrates for the provision of servicesboth on the facility side and on the individualprovider (mainly physician) side. Inessence a significant portion of the "fat" available in the system was cut through suchcontractual negotiations. These tacticsresulted in a brief, transient stabilization ofhealthcare expenditures in the mid to late1990s. However, we have entered the newmillennium and are at once offered large andrich pipelines of new drugs, devices, andprocedures in all areas of medicine and challengedby the potential costs of widespreadadoption and use.
Indeed, the rate of rise of healthcareexpenditures is again accelerating driven bythe introduction and often rapid diffusionof new, innovative healthcare technologies.The challenge this time is that the opportunitiesto find savings by negotiatingreimbursement levels with providers andprovider institutions are much more circumscribed.Thus, the focus necessarily turns tosystems, processes, and mechanisms toassure the most appropriate, effective, andefficient way to use and pay for healthcaretechnologies on behalf of patients. Theattention of all necessarily will be on care forcancers given the multiplicity of factors discussedabove.
Coverage in the 21st Century
The system of development of coveragepolicies by managed care companies/insurershas changed little since the 1990s. Largemanaged care companies have small departmentsusually overseen by a medical directorwith a PhD—scientist or nursing support.It is the responsibility of this group tomonitor the fields of medicine and identifysignificant new technologies that haveimplications for wide application in care andfor the medical loss ratio. Individuals inthese groups develop coverage policies basedupon short in-house analyses of technologicaladvances and/or upon analyses/evaluationsdone by third party specialty groupssuch as ECRI, the Hayes Group, NationalBlue Cross and Blue Shield TEC, etc.
Development of Coverage Policy
The basis for such evaluations and policiescontinues to be scientific data on thesafety and effectiveness of the application ofa healthcare technology for a specific clinicalindication. Evaluations of drugs, devices,and procedures identify relevant health outcomesfor patients and seek data from ahierarchy of sources. For drugs and devices,the approval status and approved indicationsalso serve as a starting point. In-houseversions of these analyses have becomemore abbreviated than some of the more indepthassessments performed in the 1990s.Most large companies develop formal coveragepolicies based on abbreviated technologyassessments. The transparency of theprocess and policy continues to improve asin many cases both public and private payorshave posted final policies and some supportingsynopses on their Web sites.
The need for a formal process to establishcoverage policy proceeds from contractualmandates that specify coverage for healthcaretechnologies that have achieved a levelof scientific proof of safety and effectiveness.In some cases, language regarding degrees ofmedical acceptability still exists. The truefocal point of evaluations and policies is thecontractual term specifying that payors onthe private side will not pay for any healthcareservices or technologies that are "investigational" or "experimental." On the publicside, Medicare is legally bound to pay neitherfor services and technologies that couldbe deemed "investigational" nor for servicesthat would not be considered "medicallynecessary" for the individual patient.
The evaluations upon which coveragepolicies are based are often difficult and controversial.The tension in arriving at suchpolicies is based on identification of appropriateoutcomes to be evaluated and thetype (eg, randomized trial, cohort analysis)and quanta of data necessary to define atechnology as safe and effective and thus,covered for a specified clinical indication.The situation becomes even more difficultwhen one begins to formulate policies relatedto the setting of coverage policies in seriousand life-threatening diseases like thecancers.
Issues in Coverage Policy for Cancer
In cancer care, one is often evaluatingdrugs, devices, and procedures with moderatelevels of supporting data and sometimesdata that is being extrapolated from onetumor type to another. Added to this is thecircumstance that patients with metastaticdisease may have few treatment optionsavailable to them initially or after first- orsecond-line treatment. Often, these patientsare in a real life-or-death battle.
This positions cancers and other life-threateningdiseases in a special and moredifficult category for those individuals settingcoverage policy. A limited number ofmore sophisticated processes have adoptedthe philosophy that the more serious andlife-threatening the illness then the lesserdegree of certitude about effectiveness andthe greater risk of harm that physician,patient, and payor should be willing toaccept.8 Other payors have held firm anddistinguished very little between the thresholdsand levels of evidence necessary toestablish treatments as safe and effective forlow back pain versus for metastatic lungcancer.
This has mattered little in the last 7 or 8years as the bitter battles, adverse publicity,and liability for substantial adverse legaljudgments experienced in the early to mid1990s diminished the appetite of payors toengage in the application of coverage policiesto individual medical necessity determinationsin cancer patients that might resultin the denial for coverage for their care. Butis that about to change?
The very probable influx of expensive biologics,described above, into cancer care isforcing the hand of payors. Discussions withmedical directors and pharmacy directors,even those who experienced the 1990s, indicatemounting pressures from their customers,the employers, "to do something" tostem this new rate of rise in healthcareexpenditures; a rise that is driven by theinfusion, diffusion, and use of healthcaretechnologies. The editorial by Schrag5 citedabove points to issues faced by us all. Thefruits of biomedical research are offering thepotential to combine and develop multiplelines of treatment for solid tumors that havemetastasized. Each line of treatment mayapproximate the cost of the HDCT/ABMT asexperienced a decade before.
Payors on the private side are discussingreimplementation of preauthorization schemaon a more selective basis. Targets would bevery high-cost biologics being used in situations supported neither by an FDA-approvedindication nor by an indication listed inmandated compendia. Cost pressures arebecoming so critical that a high-ranking officialof the Centers for Medicare & MedicaidServices (CMS) offered the possibility thatMedicare might stop paying for uses beyondthe FDA-approved label.9 The suggestionignited an expected firestorm and was discreetlyset aside. However, the issues andpressures facing CMS have created anintense interest on the side of private payorsto watch the "control" mechanisms thatCMS will implement to address their need toremain in the range of budget parameters setby Congress.
CMS is likely to revert to its typicalprocess for handling the issue of use-beyondFDA-approved labeling. This process generallyhas been to cede control over mostissues to Medicare intermediaries and carriers.Indeed, these regional and local arms ofCMS formulate a significant number of policiesincluding coverage policies for usebeyond the label. Over the years, the numberof national coverage policies developedby CMS has been small and focused onmajor issues of controversy or expense. Inoncology, the number of national coveragepolicies pales in number compared to thosedeveloped to address coverage for technologiesin cardiac care.
Recent CMS Draft Decision Memosand Implications
On January 28, 2005, CMS released 2major coverage policies. One focused on theuse of drugs and biologics in colorectal cancersand the other focused generally oncoverage for the use of positron emissiontomography (PET) in cancer care. The coveragedecision memos were noteworthy inthat they reflected issues being driven inpart by the passage of the MedicareModernization Act (MMA). The first issuewas the aforementioned implications of newagents and combinations of same to treatcancer.
The second issue related to a concernabout the appropriate use of imaging withPET in cancer care. This was being drivenby imaging advantages created by a technologythat can describe metabolic activitiesand uptake of other agents and thus can beused in a variety of settings including evaluationof metastatic disease, follow-up oftreatment, etc. PET also was of concern toCMS in that it was anticipated that thereduction in reimbursement for the use ofdrugs and biologics through the new averagesale price (ASP) plus 6% regulation (seebelow) would result in substantial increasesin the utilization of and billing for otherservices. PET scanning was indeed a serviceof concern to CMS.
Both decision memos were noteworthybecause CMS offered a "novel" mechanism/policy for coverage of new technologies.The policy is being called "Coverage withEvidence Development" (CED). The policysought to address head-on the issue ofpromising new technologies that diffuse rapidlyand achieve widespread use across anumber of indications beyond their intendedinitial application. The policy laid out ameans to cover the initial applications thatwere substantiated by scientific data onsafety and effectiveness. The policy alsooffered the opportunity through coverage tomake available on an earlier but more limitedbasis the technology for other indicationswhere data were incipient, scarce, ornonexistent.
In the colorectal agents' decision memo,the focus was bevacizumab, a biologic withantiangiogenic activity approved by the FDAin early 2004 as first-line treatment for colonand rectal cancers. The CED policy offeredcoverage for FDA-labeled indications andalso for indications like metastatic ovariancancer. Coverage of the latter indication wasavailable only for patients enrolled as subjectson a specific clinical trial available at alimited number of institutions. Debatequickly ensued across various constituenciesof the oncology community as towhether this policy offered an advantage ofearly coverage but with the disadvantageof significant limitations on what normallywould be a fairly steep trajectory of uptake,use, and payment for such a promising, yetexpensive agent, especially in specificcancers where effective treatments werelimited.
The PET scanning decision memo paralleledthe one for colorectal agents. The decision memo specified a listing of cancers forwhich PET scanning would be covered. Thedecision memo also indicated that otheruses would be paid for if the prescribingphysician registered patients with the indicateduses and other information yet to bespecified on a national registry. The registrymodel was viewed as less limiting for technologydiffusion and adoption of new uses.
Journal of the National Cancer Institute
While CED was promulgated as a newmodel in 2005, the concept was proposed bya private payor in a 1992 editorial in the.10This proposal was never implemented on theprivate side but did generate much discussion.It is likely that the follow-up CMS proposalon CED will find limited application aswell. The CED concept has actually beenapplied once before at CMS as a national policyfor lung volume reduction surgery.11Coverage of the procedure for emphysemawas limited to a national clinical trial runby the National Heart, Lung, and BloodInstitute at several sites around the country.The study took several years to conduct andto complete, eventually resulting in an affirmativecoverage policy for the procedure.
Application of Coverage PolicyMechanisms Today
It is likely that the mounting pressures ofa steady stream of newly introduced technologieswill generate sufficient cost pressuresthat coverage policy will be used as away to slow the introduction and, especially,the diffusion of new technologies. Suchapplications of coverage policy will be modestand selective. Modesty of application willproceed from society's present mindset toprovide access to innovative technologies tothose with life-threatening illnesses forwhich effective interventions are not available.Selectivity will proceed from the samethinking and will clearly be illustrated by theuse of preauthorization and medical necessitydeterminations for technologies characterizedby the potential for broad use at ahigh per unit cost.
Biomarkers in Cancer Care
Expansion of preauthorization and medicalnecessity processes may actually bestimulated by the same scientific advancementthat allows development of new biologicagents. The elaboration by biomedicalresearch of the cascades of biochemicalpathways resulting in the identification ofspecific nucleoside, protein, and gene targetsfor new biologic therapies permits moreexplicit and specific identification of patientsubpopulations for whom the agents are likelyto be effective. Overexpression of suchtargets like the her-2-neu receptor enablethe identification of these targets as biomarkersfor more specific classification ofpatients and the treatments (ie, trastuzumab)that they should receive. Thus, applicationof biomarkers as inclusion criteria forpatients in coverage policies and preauthorizationschema will serve as enablers forpayors that wish to use such mechanisms inselect circumstances where they feel a needand an opportunity to both improve andcontrol utilization.
One more area in the realm of coveragepolicy in oncology needs to be mentioned,that is, the use of formularies. Cancer drugsand biologics have not seen the formularyprocess applied to influence their availabilityand usage. Traditionally, these agentshave been administered parenterally. Thus,they have been covered under the medicalbenefit in private plans and under both PartA and Part B for Medicare. The establishmentof a prescription drug benefit underMMA now provides a clear incentive thatwas heretofore not present. That incentive isfor the development of new agents likecapecitabine, imatinib, and erlotinib, in oralformulations. The very limited coverage oforal agents under Medicare was a disincentiveto pharma/biotech companies for thedevelopment of promising new chemical andbiological entities in oral formulations.Passage of the MMA removes this barrier andproffers the ability of chemo/bio therapy tobe delivered orally. Oral therapy will bringnew conveniences and efficiencies in administrationbut also raises issues such as monitoringof patient dosing, compliance,adverse events, etc.
One open question related to the overalluse and utility of formularies is the establishmentof a regulatory process by which genericversions of biologics will be allowed to bemarketed. The FDA has discussed this issuefor years, but it is still far away from therelease of a draft policy for comment. Manufacturersargue vigorously that the biologicand biochemical manufacturing processesnecessary to assure the quality and purity ofbiologic agents are exponentially more complexthan the chemical processes necessaryto manufacture traditional cytotoxic chemicals.Thus, for some time, it is likely that theuse of formularies will be a very limited toolin controlling costs through the preferentialpositioning of generic biologics and controllingavailability and utilization through selectionof biologics with similar mechanisms ofaction. To this latter point, development ofseries of agents with the same or very similaroverall mechanisms of action (eg, antiangiogenesis)may actually result in the use ofmultiple antiangiogenic agents as they willdiffer in their enzymatic, receptor, or othertargets. As such, agents of the same pharmacologicclass (to use a traditional term) mayexhibit an additive therapeutic benefit at theleast, if not a synergistic one. The unravelingof and the specific definition of nature's codeof cell cycle control proffers great benefit forpatients but offers a broad array of possibilitiesfor those attempting to develop the moremundane, yet critical forecasts of technologyutilization and costs.
Discussion about appropriate ways tocontrol utilization of healthcare technologyduring the tension-filled days of the 1990soften turned to admiration of the fact thatMedicare sought to control utilization less bylimiting availability through coverage andmore through incentives determined bylevel of reimbursement. Indeed, in establishingthe hospital-based prospective paymentsystem in the 1980s to reimburse inpatientcare, CMS, then Health Care FinancingAdministration, established clinically logicalgroupings of interventions and let the clinicians(and hospital administrators) decidewhat technologies would be available tomanage patients under the relevant diagnosisrelated grouping (DRG). Indeed, the conceptwas later extended to the hospitaloutpatient side with the establishment ofambulatory groupers as the basis for reimbursementof this care.
MMA and Changes in Reimbursementfor Cancer Care
The DRG system actually was an effectivemechanism of influencing technology utilizationand costs, including in cancer care.However, cancer care had a system thattroubled those at CMS and the federal government'sOffice of Management and Budgetfor years. These government bodies felt thatreimbursement for oncologic and supportivecare drugs administered in the community-basedoncologist's office resulted in a windfallfor that oncologist and his oncologycolleagues. The older reimbursement systemwas based on the term, "average wholesaleprice [AWP]," which was very hard to define(even for the government) given the discounts,rebates, in-kind services, etc, thatcould be wrapped into purchasing agreementsbetween manufacturers or third partydistributors and oncologists' offices. So,whether reimbursement was based on AWP,AWP minus 5%, or AWP minus 12%, CMS feltthat from a financial standpoint, oncologistswere always far ahead of the game at the endof the day.
Repeated efforts by the federal governmentto address this situation were met bystrident opposition by the oncology communityand its representative organizationswho maintained that the overpayment foroncologic agents only made up for grossunderpayment or complete lack of paymentfor the variety of components (infusion,pharmacy storage, nursing support, inventorymanagement) that composed the necessarycare infrastructure and services in anoncologist's office. Underpayment for theseservices was indeed acknowledged as a realityby CMS. But CMS maintained its positionthat the overpayment on the drug side faroutweighed underpayment on the officeservices side. This issue continued muchlike an old family argument until the passageof the MMA.
Indeed the expansion of benefits underthe MMA offered CMS the opportunity, asthe MMA was in development, to address theperceived inequity for the payment for drugson the physician office side in oncology. Thekey to exerting better control was to establisha more clear definition of the purchaseprice for drugs and biologics to be administered in the community oncologist's office.The adopted and operative term, "ASP," isdefined as the manufacturer's sales to allpurchasers (excluding sales exempted fromcomputation) in the United States for suchdrug or biological in the calendar quarterdivided by the total number of such units ofsuch drug or biological sold by the manufacturerin such quarter.12 Sales exemptfrom this computation include those salesexempt from the inclusion in the determinationof best price and sales that are nominalin amount. Pharma/biotech companiesare required by law to report on a quarterlybasis accurate "ASP" for their marketedagents. Reimbursement for the use of oncologicand supportive care agents is thenprovided at ASP plus 6%. Based on fees collectedby third party distributors, someoncology offices believe that the new reimbursementlevel for them is actually ASPplus 4% to 4.5%.
Regardless, the ASP plus 6% level is meantto better define, account for, and stabilize aMedicare drug spend that is about to undergosubstantial escalation in 2006. Threatsof a mass exodus of oncologists from communitypractice through retirement orother mechanisms have not been realized.Indications that many patients with"unprofitable diagnoses" would be referredto and thus overwhelm the hospital settinghave not materialized to any great extent.Indeed some practices, especially largeones or networks of large practices arealready maneuvering through the newreimbursement schema quite well based ongroup purchasing power, more efficientmanagement, the 6-month lag in actualASP reporting, and diversification of servicesprovided.
CMS from its position is pleased with thefact of implementation of this improvedreimbursement schema. Additionally, CMSwill take back in FY 2006 part of theincrease in the reimbursement for office-basedservices related to the administrationof oncologic agents. An existing issue is: WillCMS continue to give back some of the lostreimbursement to community oncologistsin 2006 as it did through the 2005 qualitydemonstration program? At this writing itis likely, as the US Congress just passed aresolution supporting such continuation.Whether or not the 2005 supportive carequality demonstration is continued, augmented,changed, and replaced is an openquestion. Finally, will CMS shock the systemby becoming a direct purchaser of drugs andbiologicals as the Veterans Administrationhas been doing for some years? This conceptis one that is picking up steam and supportin Congress.
Reimbursement methodology and schemafor cancer care on the Medicare side seempretty well set for the near term. Clearly, privatepayors and their employer clients arewaiting, watching, and evaluating the effectsand effectiveness of the CMS programs.Private payors, on their own, are increasinglylooking for ways to address the same issueas Medicare has through ASP plus 6%.Possibilities for implementation includeintroduction of a similar reimbursementmodel, new attempts at managementthrough the use of specialty pharmacy intermediaries,transition of management of parenteralagents to the pharmacy side from themedical side, case rates, etc. Whatever themechanism, stemming the rate of rise ofhealthcare expenditures attributable to theexpanding use of expensive drugs and biologicalswill be a reimbursement challengefor some time to come.
Quality-of-Care Evaluation:The Last Great Hope?
A variety of strategies and schema havebeen developed and implemented applyingcoverage and reimbursement policies as ameans to achieve some control over the utilizationof healthcare technologies and somerelief from a steady rate of rise in healthcareexpenditures. These efforts have met withmodest success. Now the technologicalimperative is driving more and more drugs,devices, and procedures to the healthcaresystem as the fruits of biomedical researchare being realized. Many policy makers areshifting their view to ongoing and publicevaluation of the quality of care delivered asthe primary means to achieve the biggestbang for one's healthcare buck or as describedin more erudite circles maximizingvalue from the healthcare dollar spend.Thus, as Relman13 proposed in 1988, maybewe are indeed about to enter medicine's "Eraof Accountability."
The quality-of-care paradigm involvesbasic components as follows. Evidence-basedguidelines should communicatespecifically and clearly the appropriate algorithmicprogression of care for patients withcertain clinical characteristics. From thepool of guideline recommendations, qualityindicators are identified that serve as measuresthat can be used to evaluate the caredelivered by providers as clinicians or institutions.These measures can be benchmarkedagainst the median or against bestpractices and then fed back to providers ina continuous quality improvement loop toestablish a basis for action and for improvementby that provider.
The measures and quantitative data canbe taken to the public arena. Public reportingof performance data will be used by thepayors and purchasers, including patients,to better select the setting and the providerlikely to provide optimal care. Attainment ofthe ability to benchmark and rank order orassign a hierarchical categorization toproviders can then be used as a means todistinguish and reward those "high performers" and to encourage their peers to aspireto such recognition. Prospective or retrospectivefinancial rewards for these highperformers have been called "pay for performance" (P4P). Interest in this area is sokeen that it is difficult to open one's mail atwork without receiving an invitation to yetanother P4P conference.
The basis for quality evaluation is theestablishment of the parameters for appropriatecare in clinical guidelines. A majorimpetus for the establishment of clinicalguidelines arose from the studies on thegreat variation in clinical practicedescribed by John Wennberg in his publications14,15in the late 1980s. Wennbergshowed, for example, substantial variationin the per capita rates of laminectomybetween 2 communities, Boston and NewHaven, that were only 100 miles apart andfor which there was no scientific or clinicalbasis for such variation other than physiciandiscretion and preference.14
These studies along with others thatidentified unnecessary use of commonprocedures such as cholecystectomy andhysterectomy drove the call for the developmentof guidelines to literally better "guide" the care being offered to patients. A plethoraof guidelines has been developed since.Most have not been used and have been discardedor lost in nice leather bound editionsthat sit on bookshelves. The major reasonsfor the lack of use are mainly 3-fold. First,most guidelines are delivered "dead onarrival." Laborious processes can take yearsto deliver a single guideline that is out ofdate as soon as it is finished. Secondly, mostguidelines are not specific. A lack of specificityabout what constitutes appropriatecare limits the utility of such documentsto clinicians. Third, and most important,guidelines have not been regularly updated.In a rapidly advancing field like cancer care,lags in updating diminish the utility of theguidelines to the end user. To develop, maintain,and update a comprehensive set ofguidelines in one area of medicine is adaunting challenge.
In 1995, the National ComprehensiveCancer Network (NCCN) initiated developmentof a comprehensive set of guidelinesfor oncology. The NCCN guideline processaimed to overcome the shortcomings ofother guideline efforts and within the spaceof 3 years developed a comprehensivelibrary of 110 guidelines that covers the carefor about 97% of all patients, for their supportivecare, and for prevention, screening,and early detection. The NCCN ClinicalPractice Guidelines in Oncology™are nowwidely recognized and applied in both thecommunity practice and academic practicesettings as the standard for clinical policy inoncology. The NCCN Guidelines are increasinglybeing used by payors to set coveragepolicies. In August of 2005, more than484 000 guideline files were downloadedfrom www.nccn.org. A major reason for thesuccess of the NCCN guideline effort is a truesense of urgency in the translation ofresearch results and FDA approvals intocontinual updates that are available in amatter of a few weeks.
The NCCN Guidelines respond directly toa recommendation made by the NationalCancer Policy Board (NCPB) of the Instituteof Medicine that called for the developmentof an authoritative set of guidelines uponwhich clinicians could rely.16 Additionally,the NCPB called for the establishment ofdatabases that would evaluate the qualityof cancer care on an ongoing basis. Theserecommendations proceeded from a fairlydrastic conclusion of the NCPB that "thereis no national cancer care program or systemof care in the United States?The adhoc and fragmented cancer care systemdoes not ensure access to care, lacks coordination,and is inefficient in its use ofresources." The NCPB, composed of leadingclinicians and patient advocates, hadissued a call to action for the oncologycommunity.
The wide acceptance, application, andavailability of NCCN Guidelines have facilitatedtheir use in a variety of quality assuranceand quality improvement programs.More specifically, guideline recommendationshave been the basis for the identificationof circumscribed sets of qualityindicators by a number of different organizationsincluding the National QualityForum, the Commission on Cancer, privatehospital systems, and a joint effort of theNCCN and the American Society forClinical Oncology.
To what purpose? As in diabetes withhemoglobin A1C and acute myocardialinfarction with b-blocker administration,indicators in oncology will gradually find theirway into quality improvement programs. Theidentification of a set of 3 to 5 indicatorswould allow for more focused and efficientdata collection on underlying practice.This is a key point as the collection of dataacross practices in different settings in theUnited States would be a major challenge.
One organization has been able to collectdata over the years through a process usedto accredit cancer centers across the UnitedStates. The Commission on Cancer, a membershiporganization, led by the AmericanCollege of Surgeons, the American CancerSociety, and other major associations, hasdeveloped the National Cancer Databasethat houses data collected principallythrough tumor registries in local hospitals.This model is useful to evaluate in terms ofthe possible development of a nationalreporting system to measure quality.
In oncology, the focus, at this writing, isclearly on CMS. CMS leadership has exhibiteda true openness and willingness to workwith the medical community to achieve areasonable basis for moving forward on thequality and value proposition for healthcare.The quality demonstration project for fiscalyear 2005 involved the use of codes andreimbursement for community practices toevaluate and report on the status of treatedpatients with respect to levels of fatigue,pain, and nausea and vomiting. Physicianshad 4 codes for each symptom and used 1 ofthe 4 codes to denote the degree of samefrom no symptom to severe. While somehave criticized the demonstration project asnot collecting any usable data, others understandthat conceptually CMS is sending amessage to the oncology community thatstarting now reimbursement is likely to betied to some form of data reporting.
The existence of a specific MedicareQuality Demonstration Project in oncologyfor 2006 has not been determined at thiswriting. However, discussions have indicatedthat CMS may establish a demonstration programthat seeks to deliver the message thatguidelines, like the NCCN Guidelines, developedthrough a formal systematic process byan authoritative source should be the basis ofcare delivered. In following these discussions,we see a system evolving that usesguidelines as a foundation and that encouragesthe identification of quality indicatorsfrom such guideline recommendations.
These quality indicators must be scientificallyvalid. Collection of data to establishevaluative processes requires that such dataelements pass the feasibility test in terms ofdata collection. The indicators must identifya practice that has demonstrated variabilityin order that improvement might be achievable.Importantly, quality indicators must beassessed to sufficiently impact patient outcomesto warrant the effort and resourcesthat will be invested in data collection,analysis, and reporting. Also, quality indicatorsmust have value to the end user. Finally,any such process will have to overcome amajor deficit in claims data available tomanaged care companies, that is, the lack ofstage of the cancer.
These activities in both the public and privatesectors eventually will lead to a cancerdelivery system where public reporting of performancedata will be common and expected.Also, P4P will gradually become an appliedmechanism to drive the system to the attainmentof optimal return in terms of patientbenefit for every healthcare dollar expended.
One of the major objectives of qualityevaluation efforts is to provide clinical informationand provider performance data topatients and families to involve the patientside more prominently in the decisionsabout the specific clinical interventions,about the setting of care, and about theselection of the provider of care. Both providersand managed care companies aredeveloping Web sites with informationalcomponents to address these 3 areas andmore. Synopses about disease and appropriateclinical interventions with referenceand links to explanatory and supportingarticles are provided. These informationaland educational efforts seek to mold theclinical decision-making process into apatient/physician partnership.
Patients are being provided with user-friendlyversions of information previouslyavailable to clinicians. Such informationincludes the guidelines and scientific evaluationsof interventions available to clinicians.Additionally, building upon thepublication of specific performance measuresfor cardiac surgeons that continues tothis day in New York State, organizations arebeginning to publicly publish such data oroffer it on a consultancy basis. CMS haslaunched a Web site, www.hospitalcompare.hhs.gov, that seeks to provide patientswith information on the quality of cardiaccare delivered at hospitals across thecountry. Specific quality process measuresfor acute myocardial infarction like "aspirinat discharge" or for pneumonia like "oxygenationassessment" are described and performancewill be detailed. On April 1, 2005,Dr Mark McClellan, administrator for CMS,launched the Web site and commented:
"We are trying to equip consumers tothink more concretely about quality, torealize that quality can differ significantlyacross facilities, that valid and widelyused measures of quality can be a tool tohelp understand quality differences, andeven more importantly, that individualscan impact quality through their questionsand through their choices."17
A variety of organizations on the privateside are poised to contribute parallel comparativeperformance measures and data topatients and providers alike. Nationalmanaged care companies are planning aphase 2 beyond their phase 1 work thatprovides more general clinical information.That phase 2 will emulate what the state ofNew York and what CMS are doing. Companieslike Ingenix, MedStat, etc, are buildingvery large databases to benchmarkperformance and efficiency of healthcareproviders for payors and employers.
Major portals for health information likeWebMD seem to be readying themselves toprovide similar information; witness the purchaseby WebMD of HealthShare Technologythat heretofore had provided hospitals, managedcare companies, and employers "withInternet tools that lead to more informedchoice of hospitals for inpatient care, consideringboth cost and quality."18 Consideringthe WebMD target audience, thegeneral public, one could surmise that thesetools will soon be available to patient consumersof care. Notwithstanding some academicconcerns about validating the modelsand tools used to develop clinical and/orfinancial hierarchies and tiers of providers,these models are being developed and readiedfor delivery to users of a variety of decisionmakers and users of the healthcaresystem; full steam ahead!
As mentioned above, cost, financial, orefficiency of care data will be released aswell. Attempts are being made by payors andemployers to inform patients about the costsof their care and to incentivize them to useproviders who are more efficient in theirdelivery. The models to incentivize arepresently in flux. The move toward tierednetworks where patients would assumeincreasing co-pays based upon some formularizedexpression of the costliness and/orthe quality of care of the provider thatpatients might choose compared to otheravailable providers has not been greetedenthusiastically by patients, not to mentionproviders. Additionally, most employers haveassumed a wait and watch position andhave not moved to make such tiered networkprograms either an option or mandatefor their employees.
A final and quite important considerationin this discussion is the use of "co-pays" tomake patients and families more aware ofthe total cost of their care and the degree ofsubsidization of such care be that by majoremployer, employer coalition, or state orfederal government. The term "co-pay" isused here more generically and broadly thanusual. Co-pay is meant to include the aggregateof the costs borne by patients in theform of coinsurance (eg, 20% of acceptedcharges), deductibles (eg, payment of first$500 for medical care), or actual traditionalco-pays (eg, $10 for a generic prescription).Payors are seeking to calibrate the appropriatethreshold of copayments that will makethe patient explicitly consider the personalworth of the potential clinical and quality-oflifebenefits to be derived from the patient'sexpenditure of her/his healthcare dollar.For example, if a second-line oncologychemo/bio regimen will cost a patient a copayof $3000 and the median increase insurvival seen with the regimen is 2.5 monthsover traditional much less costly regimens,will the patient deem the expenditure of$3000 likely to provide a sufficient enoughreturn on investment in terms of improvedpatient outcome? The same decision-makingprocess might be invoked in evaluating aco-pay for the use of agents to ameliorate apatient's state of fatigue.
The degree to which the co-pay conceptis implemented and at what level of co-paywill be examined on a societal basis. Formost patients, decisions about annualexpenditures of $3250, $5000, or $8000over 2 to 3 years are major decisions. Theintegration of such financial considerationsand incentives/disincentives into benefitslanguage will likely attract much attention,including legislative attention not too fardown the road. As discussed all through thisreview, in serious and life-threatening disease,language, processes, and policies thatimpact the availability of care and access tocare will be scrutinized intensely.
The healthcare system is once again beingchallenged by financial pressures and tensionsreminiscent of the late 1980s and earlyto mid 1990s. An aging population and amajor influx of new healthcare technologieswill serve only to accentuate the challenge.The federal government and major employersmust address the accelerating rate of rise intheir healthcare dollar spends. Policy makersare again reviewing methods and models todo so. Lessons learned from the battles of the1990s between providers and payors are evidentin some models and seem to be disregardedin others. Proposals to institute themore draconian coverage policy-based methodsof the 1990s are scarce. Selectivity andsensitivity in implementation of processesthat might intrude into the physician/patientrelationship are being exhibited. Recognitionof the coverage decision as a risk benefitanalysis that more closely approximates theserious and life-threatening nature of cancertypes is critical. Reimbursement methodologiesare being fine-tuned, but residuum of"fat" to cut is meager.
Ongoing evaluation of the quality of careprovided by clinicians is now viewed as thefocal point for maximizing the value derivedfrom the expenditure of healthcare dollars.Continuous quality improvement, measurement,and reporting programs will emphasizethe necessary foundation of practice asevidence-based practice guidelines aredeveloped, distributed widely, and updatedby authoritative groups in a timely fashion.Physicians and other clinical professionalswill be incentivized through P4P to followpractice guidelines and to be judged onguideline recommendations translated intoquality indicators. Patients, as those withthe greatest investment (both health andfinancial) in the outcomes of care, will beencouraged to participate more fully indecisions regarding the money they spendand in the care that such monies buy. Toaccomplish this, clinical information andcomparative performance benchmarkingdata will become more available to patientsand to corporate purchasers of healthcare.
The bet is that after all the studies and allthe applications of models (eg, capitation,disease management programs) over thepast 20 years, a return to and a focus on thescience of care through the application andimplementation of guidelines with performancemeasurement will lead to the highestquality care, the best patient outcomes, andthe most efficient expenditure of healthcaredollars. It is a trillion-dollar bet!
Address correspondence to: William T. McGivney, PhD, Chief Executive Officer, National Comprehensive Cancer Network, 500 Old York Road, Suite 250, Jenkintown, PA 19046; firstname.lastname@example.org.