WPC 2BPZCourier 10cpi33x\x6X@8;X@HP LaserJet III (Additional)HPLIIIAD.PRSx  @,\,HX@22BB Z33xCourier 10cpiCourier 10cpi (Bold)Pres Elite 16.67cpi (25inOne!)HP LaserJet III (Additional)HPLIIIAD.PRSx6X@;,\,HX@X01Í ÍX0Í Í20XkF`x6X@;X@<6X9`("Courier 10cpiX?xxx,\x6X@8;X@\?xxx,x `B;XZ QEHHH,#^Hp@U^@X0Í Í  = The Use of CostEffectiveness Analysis by Pharmacy Benefit Management Companies l Henry Grabowski, Duke University P C. Daniel Mullins, University of Maryland  ? ," I. Introduction ă The emergence of pharmacy benefit management companies (PBMs) constitutes a major structural change that has occurred in the distribution of pharmaceuticals during the 1990s. PBMs provide a variety of services designed to influence outpatient prescription drug usage and costs. It has been estimated that over 60 percent of employers utilized some form of PBM services in 1995 with coverage of 137 million beneficiaries (Health Strategies Group, 1995). Pharmacy benefit management firms have evolved over the past decades from very diverse originse.g., claims processing firms, mail order pharmacies, and subsidiaries of insurance firms and managed care organizations. A key factor that has fueled the industry's explosive growth in recent years is, of course, the increasing demand for cost control by health care payors. Also very important in the case of PBMS, however, was the development of online electronic data interchange capabilities at virtually all  ?(# retail pharmacies by the end of the 1980s.YX(#]  ?H& ЍThe Catastrophic Health Care Act of 1988, which mandated outpatient prescription benefits for Medicare beneficiaries, was a major catalyst in the conversion of the retail pharmacies to on-line capabilities. While the Act was repealed prior to implementation of the outpatient prescription drug benefit, one ofh)=0*0*0* its key legacies was the ability of retail pharmacies to perform online claims processing and adjuration. This became a critical component underlying the rapid growth of PBMs during the 1990s.Y(#=0*0*0*ԌAt the most basic level, PBMs perform electronic claims processing, provide a network of retail pharmacies and mail order pharmacy service, and offer pharmacy benefit plan design with patient cost sharing and other incentives. These activities generate cost savings to plan sponsors by reducing prescription claims processing charges, lowering ingredient and dispensing fees, and increasing patient copayments. Beyond these basic services, PBMs also engage in various activities designed to influence drug dispensing and prescribing. These activities include generic substitution, formularies, preferred drug lists, therapeutic interchange programs, treatment guidelines, drug utilization review and prior authorization. One of the main sources of cost savings to plan sponsors arises from the ability of PBMs to secure rebates from manufacturers for including their products on formularies and preferred drug lists. Arrangements between PBMs and manufacturers often take the form of performancerelated contracts which tie the size of rebates to volume targets and other incentivebased milestones. Rebates have been a principal driver of the PBM industry during the 1990s. The PBM industry is a very dynamic one with considerable entry and consolidation over the past five years. It has been estimated that over 40 firms provided PBM services to third parties on a forprofit basis in 1995 (GAO, 1995). Table 1 shows a list of the leading PBM firms in June 1995 with information on their covered#0*((@@ beneficiaries in formulary plans, and their total covered beneficiaries. Because many beneficiaries are covered by two or more firms with different services, one needs to take account of potential double counting in analyzing this industry. Even allowing for this fact, however, it is clear that the top ten PBM firms account for a large share of covered beneficiaries (over 80 percent). While the first part of this decade was characterized by considerable entry into the PBM business, the last few years have been characterized by considerable consolidation including vertical  ? mergers between leading PBM and pharmaceutical firms.:V  ?0 ЍIn the last two years, Merck acquired Medco, SmithKline Beecham acquired Diversified, and Lilly acquired PCS. These mergers have been somewhat controversial, spawning scrutiny by the FTC concerning the possibility of vertical foreclosure of rival pharmaceutical products and related practices. In this regard, Eli Lilly and PCS entered into a consent decree with the FTC to provide an open formulary option and to prevent information acquired by PCS on rival manufacturers competitive practices from being revealed to Lilly. Some of the competitive issues posed by these mergers have been recently considered by the GAO (1995).  While competition in the PBM industry has centered around managing drug benefits on a carvedout basis, many PBM firms anticipate a growing future role in the management of patient care (disease management) for several highcost chronic diseases and conditions (e.g., asthma and diabetes). This would allow PBM firms to profit from the expected movement by managed care organizations toward a more coordinated systembased approach to health care problems. Of course, disease management requires the integration of medical data bases and a much broader perspective than PBMs have employed when managing drug costs on a "carvedout" basis.( 0*((@@ԌThe development of more sophisticated disease management capabilities also has been a major rationale offered for the mergers and alliances that have occurred between pharmaceutical firms and PBMs. Many of these joint efforts are initially focusing on highcost, chronic diseases such as asthma, diabetes and depression. These are diseases in which patient compliance and undertreatment are major issues. In this paper, we are especially interested in analyzing how PBMs utilize costeffectiveness studies in the management of drug benefits. Information for this analysis was obtained from several  ? sources including personal interviews at five leading PBMs.:V  ? ЍThe firms interviewed were Diversified, Medco, PCS, Pharmacy Gold and Value Health. In most companies, we spoke with individuals involved with P&T Committee decisions, outcome analyses, and business strategy. These firms provided formularies, treatment guidelines and other relevant documents to us. This was supplemented by materials from some of the smaller PBMs and from third parties that compile data and information on the PBMs industry. In the next section of the paper, we examine the nature of formularies and other interventions that PBMs currently utilize to influence drug prescription choices. The use of costeffectiveness studies in these activities has been limited. However, as the scope of PBM activities broadens over time to various disease management activities, the importance of costeffectiveness studies to PBMs is expected to increase significantly. Section III of the paper considers these issues as well as the role of government`"x0*((@@ policy makers like the FDA in regulating costeffectiveness claims made by manufacturers to PBM firms and other managed care organizations.  ?x d II. Formularies and Related Drug Product Selection Activities  ?  In our interviews, we obtained estimates from the leading PBMs of the potential savings to payors, relative to an unmanaged plan, from PBM interventions designed to affect drug product selection. This information is presented in Table 2. These activities can produce estimated savings between 14 to 31% in a health plan's total drug expenditures (25 to 40% when one also takes account of basic PBM activities like claims processing, pharmacies and mail order services, and benefit design). The major savings accrue from generic substitution and formularies, including various compliance measures.  ?  A. Generic Substitution As shown in Table 2, generic substitution can save payors up to 10% of their total drug costs. In this regard, a managed plan can achieve generic utilization rates of 35 to 45%, compared to rates of 15 to 20% for unmanaged plans. Generic substitution is typically encouraged through various benefit designs. A common feature is a twotiered copayment or coinsurance approach, with generics paying significantly less than the branded product. Popular multiplesource products also may be subjected to a maximum allowable cost reimbursement policy. In addition, physicians that write DAWs disproportionately may be targeted in retrospective drugH&0*((@@ utilization review letters, including a qualitative analysis of how much patients could save through generic prescribing. PBM policies encouraging generic prescribing have become a major factor contributing to the more rapid erosion of branded drugs sales revenues when patents expire. Some representative results from a forthcoming study by Grabowski and Vernon (1996) are presented in Table 3. As this Table shows, the more recent cohorts of major drugs coming off patent since 1989 have experienced greater sales losses to generics. In this respect, the 1993 cohort lost over 50 percent of their market share (in units) during the first six months of generic competition, compared to 31 percent for the 198990 cohort. Furthermore, the cohort of similar drugs coming off patent in the period 19841988 didn't experience market share losses of 50 percent to generics until more than two years after generic competition was introduced (Grabowski and Vernon, 1994). The large number of commercially important drugs whose patents are scheduled to expire between now and the year 2001 will make generic substitution an important source of ongoing savings in drug benefit expenditures over future periods. One recent analysis indicated that drugs with 13 billion dollars in value are scheduled to come off patent between 1994 and the year 2001 (Treppel, 1994). Another analyst projects that generic dispensing could potentially account for 65 percent of all prescriptions by the end of the decade (Shaw, 1992).%0*((@@Ԍ ?  B. The Formulary Decision Process Because PBMs have a set of clients that range from traditional indemnity plans to PPOs and more restrictive HMO organizational forms, they have developed a spectrum of formulary instruments. These can be classified into open, preferred and closed formularies. Open formularies are merely lists of recommended drugs for which relative cost information is provided. Preferred formularies impose lower copayments for drugs on the formulary. Closed formularies provide reimbursements only for drugs on the formulary. It is up to the client firms to select how aggressive and restrictive they wish to be in their formulary interventions. A large share of the users of PBM services are employers that have evolved recently from feeforservice to PPO or gatekeeper models of managed care. These plans usually elect, at least initially, an open formulary that gives a wide choice in prescription medicines. The main idea here is to educate physicians and patients on the differential costs of alternative medicines. Employers then generally move over time to incorporate the economic incentives associated with a preferred drug formulary. These formularies programs are generally complemented by retrospective drug utilization reviews and therapeutic interchange programs (where pharmacists call physicians in an effort to persuade them to switch prescriptions to formulary listings). Closed formularies, in which only listed drugs are reimbursed, are%0*((@@ unusual in employer plans and are largely restricted to stafftype  ? HMOs.:V  ? ЍIn a survey of 200 medium and large employers with drug carveout plans, by Health Strategies Group (1995), they found that 68% had open formularies, 38% had formulary copay differentials, and 14% had closed formularies. A key decision for PBMs is which drugs should be included in their formularies. We found that PBMs typically employ a twostage or bifurcated decision process in this situation. The initial evaluation of a new drug entity resides with the pharmacy and therapeutics (P&T) committee. Some PBMs utilize outside experts for this purpose, while others rely on inhouse staff. The key decisionmakers on the P&T committee are physicians and clinical pharmacists. The focus of this committee is on a new drug's comparative efficacy and safety and its substitutability with existing drugs. Usually new drugs are evaluated within a short time interval around their approval by the FDA. All available information is utilized in preparing the dossier for the committee's review and action, including the FDA's approved labeling, published studies, and materials supplied by the drug manufacturer (Blissenbach, 1993). The P&T committee decision leads to three possible outcomes. They can decide that a new drug must be included on the formulary (because it offers significant therapeutic advantages over existing  ?@ drugs),!X@x:V  ?' ЍThe proposition that unique new drugs offering significant therapeutic advances will be approved for formulary inclusion was'0*(( consistent with the information we reviewed. However, since these medicines are often expensive and lead to rising drug budgets, PBMs may recommend restrictive use provisions or prior authorization. Many biotech entities such as human growth hormone, were included in the list of drugs recommended for prior authorization.! it must not be included on the formulary (because it has@x0*((@@ inferior clinical properties), or that it may be included on the formulary (because it is determined to be broadly interchangeable with existing drugs). Many new drugs are part of an existing chemical class of compounds (e.g., ACEinhibitors, calcium channel blockers, etc.). If the P&T committee determines that a new drug is basically interchangeable with other drugs in its class, then issues of cost become important to the formulary decision. At this point, business decisionmakers at the PBM evaluate the new drug's cost relative to substitute therapies and become involved in the negotiation of rebates with manufacturers. We found that the formularies of the leading PBMs are relatively inclusive. This is illustrated by the data in Table 4. This Table shows the drugs on the formulary for different leading PBMs in several cardiovascular classes. It is noteworthy, for example, that the market for ACEinhibitors includes only eight drug products (seven distinct chemical entities), and that there are between five and eight drugs on the national PBM formularies of  ?  these five top PBM companies.X x:V  ?# ЍIndividual health care plans may opt for more restrictive formularies, but this is mainly done by HMOtype plan clients. The inclusive nature of current PBM formularies reflects a number of factors. First, physicians and patients historically@ 0*((@@ have been used to little or no restrictions on prescription drugs in conventional indemnity plans. They are gradually becoming acclimated to the necessity of restrictions in choice in health care as managed care grows in scope and scale. Second, the pioneering or firsttomarket products in a new therapeutic area have significant firstmover advantages with physicians. Later entrants in a class generally sell at lower prices, unless they offer important incremental therapeutic advantages. PBMs and other managed care entities have been a strong catalyst to the stimulation of price competition among drugs in major crowded therapeutic class (Grabowski, 1995). To date, however, PBMs have tended to add new lowerpriced drugs to their formularies, without deleting a corresponding number of rival brands in its class. Hence, rather than having only two or three drugs in a class like the ACE inhibitors, they have five to eight brands. PBM formularies list drugs by therapeutic class and provide relative cost indices on drugs deemed to be close substitutes. In particular, each product listed is accompanied by one to five (or sometimes six) dollar signs ($ to $$$$$) to denote relative costs. Products that are available generically are given an asterisk or some special marking. Table 5 provides an example of formulary listings for the ACE inhibitors class for the five PBMs in our survey. Most products in this class span a range of $$ to $$$$ with the newer products in the class having lowered costs than the established market leaders (Capoten and Vasotec). As expected, there is a high rank' 0*((@@ correlation in the relative cost indices of different PBM firms. Differences in cost indices reflect different rebate arrangements as well as differences in scaling. The degree to which these relatively inclusive formularies can result in significant cost savings depends on their associated incentive and compliance measures. As discussed above, many plans include lower copayments for formulary drugs. Other forms of compliance actions focus on physicians prescribing behavior through retrospective drug utilization evaluations (DUE). Under this approach, physicians that fall outside prescribing norms are subject to "alert letters" and possible financial penalties. At the present time, there is a strong dynamic toward the development of more restrictive formularies (i.e., fewer drugs in  ?P a class) and stronger compliance measures for clients by PBMs.@P:V  ?p ЍAll of the leading PBM firms in our survey indicated this to be the case, and it was generally borne out by a recent survey undertaken by the Health Strategies Group (1995). They found, for example, a growing use of economic incentives in carved out employerbased formularies. While 38 percent of employers now use differential copay for formulary drugs, this is expected to grow to 51 percent in the near future. Among other things, this reflects the fact that pharmaceutical manufacturers are willing to negotiate larger discounts in exchange  ? for increased market share.@ :V  ?(# ЍBeyond offering employer plans, the option of more restrictive, or even closed formularies, several PBMs have earmarked certain drugs in their formulary for priority status. PBMs then employ a variety of incentive measures to increase usage of these priority drugs, which in turn triggers larger rebates from the manufacturer. These measures can include therapeutic switch campaigns in which clinical pharmacists phone physicians. '0*(( Physician behavior may also be influenced by the use of withhold pools and other financial incentives. Similarly, patients and retail pharmacists are also offered incentives to increase usage of the priority drugs in a given class. Staff HMOs have long recognized this 0*((@@ fact and utilized it to obtain large cost savings in their drug expenditures.  ?   C. The Use of CostEffectiveness Studies in Formulary Decisions The use of pharmacoeconomic studies in PBM formulary decisions has been limited. PBM decisionmakers that we interviewed provided several reservations concerning the applicability of existing pharmacoeconomic studies for their formulary decisions. First there is the issue of the comparators used in these studies. There are few costeffectiveness (CE) studies that compare closely substitutable products in the same class (e.g., one ACE inhibitor versus another). Rather, the comparison is typically with an oldertype therapy or a placebo. Second, there is the issue of the population studied. Generally CE studies have not specifically focused on the populations with characteristics served by the PBM's clients. Third, there are a number of perceived methodological issues that remain open with regard to CE studies (measurement of costs and benefits, discount rates, etc.). Fourth, there are perceived concerns about the objectivity of firmsponsored CE studies. Finally, some PBMs mentioned the need for bettereducated sales representatives who would be able to address questions concerning CE claims. Given the lack of persuasive evidence from CE studies showing clear differences among competing drugs in the same class, PBMs`" 0*((@@ feel that their current decision approach to the evaluation of a product's cost effectiveness is justified. In particular, in their view, the cost effectiveness of one ACE inhibitor versus another turns primarily on differences in drug costs (including rebates). They believe that drug formularies constructed on this basis can yield significant savings to a plan's drug expenditures without leading to changes in health outcomes, problems of cost shifting, or losses in quality of life for patients. Table 6 presents a list of what PBMs would like to see in CE studies to make the findings more applicable to their formulary decisions. These points address the various perceived limitations discussed above. Beyond wishing to see timely, headtohead comparisons among market leaders in relevant populations, our surveyed firms also would like to see independent sponsorship or nostring funding to insure objectivity of these studies. However, many individuals indicated that this is an ideal but possibly unrealistic expectation, given that the main financial incentives to do the studies now reside with the manufacturers. Publication in highly regarded peer journals is viewed as an acceptable second best solution in most instances. We found it significant that while PBMs were concerned about the objectivity of studies financed by manufacturers, there was little support for formal FDA regulation on the CE claims made by manufacturers to PBMs and other managed care entities. Draft regulations recently have been proposed by the agency (FDA, 1995). PBM firms felt that this would lead to less timely dissemination of' 0*((@@ information and also would constrain potentially valuable sources of information that do not conform to FDA concepts of evidence regarding cost effectiveness (i.e., the need to perform two wellcontrolled studies for CE claims). These issues are discussed further below.  ?  III. Treatment Guidelines, Step Protocols and Disease Management ă  ?(  A. Treatment Guidelines for Unique Therapies The costeffectiveness evaluation of unique new therapies is likely to be a central concern in the future for both PBMs and pharmaceutical firms. Pharmaceutical firms are now devoting a much larger share of their R&D budgets to the search for drugs with unique therapeutic needs. They seek to command premium prices for these drugs on the basis of their advances for patient well being and/or their savings in overall system costs. This underlies the strong interest of major pharmaceutical firms in health economic studies of these types of drugs. At the same time, PBM firms are currently moving to risk sharing or even capitation arrangements with their larger clients. This makes them increasingly concerned with the costs and benefits of new therapies that lead to rapid growth in drug expenditures. For example, several PBM firms mentioned the desirability of seeing costeffectiveness and cost benefit evaluations of new drugs like Immitrex, an important recent advance in the treatment of migraine, that increases the expected cost of treating this condition. As discussed above, new pharmaceuticals that offer significant therapeutic advances over existing therapies are'0*((@@ currently admitted to the formularies of PBMs by the P&T Committee. However, PBMs will make recommendations concerning their optimal utilization in guidelines disseminated to physicians. In some cases, PBMs may recommend to their clients that expensive new drugs be subjected to prior authorization. One case example that we focused on in our discussions with PBM firms was the drug misoprostol (Cytotec). It is indicated for prophylaxis in individuals at high risk for developing complications from NSAIDinduced gastric ulcers. This drug has no close substitutes. It has been the subject of several costeffectiveness analyses. (For a survey, see Barradell et al., 1993). All of the PBMs we surveyed included misoprostol on their formulary and none of them subjected it to prior authorization. However, they recommended its use only for patients at high risk for developing complications from NSAID use (e.g., the elderly, those with concomitant debilitating diseases and those with prior history of gastric ulcers). This is generally consistent with the finding of CE studies on misoprostol (Barradell, 1993). When queried, the PBM firms indicated that their guidelines for misoprostol were based on the medical decisions of their P&T Committees. In this situation, the CE studies were viewed as providing useful background support for decisions of the P&T Committee. In contrast to the PBM treatment of misoprostol, all of the PBM firms we surveyed targeted human growth hormone for prior'0*((@@ authorization. This is a very expensive drug (  $25,000 per year), PBMs wished to insure that its use was confined to approved indications. Several of the other biotech drugs were also earmarked for prior authorization.  ?@  B. Step Protocols At the present time, most PBMs have disease treatment "guidelines" that are provided to participating physicians along with the drug formularies. These guidelines often take the form of step protocols in which older, cheaper drugs are the firstline therapies. Except for a few tightly controlled HMO clients, these guidelines are voluntary and educational in nature. While PBMs engage in a variety of incentive measures to influence drug selection within a particular chemical class of drugs, they are relatively passive in their approach to treatment guidelines. The choice of treatment regimen continues to be largely at the  ?p physician's discretion.' pY  ? ЍAs indicated above, PBMs in alliance with several pharmaceutical firm partners are attempting to develop disease management programs for a number of pharmaceuticalintensive disease categories. These activities are considered below.' We investigated how treatment guidelines are developed and found that PBMs tend to follow very closely the recommendations of national commissions, government agencies and leading medical associations. For example, their guidelines often reference the Joint National Committee on Detection Evaluation and Treatment of High Blood Pressure, the National Cholesterol Education Program Expert Panel, the guidelines of the Association for Health Care`"x 0*((@@ Policy Research (AHCPR) and so forth. Leading medical texts and recent journal articles are also frequently cited in the formulation of these guidelines. In our analysis of PBM guidelines, we decided to focus in particular on blood cholesterol reduction as a case study. This is an interesting case because a major new class of drugs have been introduced since the late 1980sthe HMGCoA reductase inhibitors or "statins." These drugs offer some significant therapeutic advantages compared to older therapies, particularly in the reduction of side effects. However, they are also considerably more expensive than some older competing therapies like nicotinic acid which are available generically. Cholesterol reducers have also been the subject of several costeffectiveness studies in recent years. We found the guidelines used by PBMs in the case of the treatment of high blood cholesterol have been strongly influenced by the Expert Panel of the National Cholesterol Education Program (NCEP). The first report of the NCEP was issued in 1988. It provided recommended targeted blood cholesterol levels for preventive treatment tied to a person's risk factors for coronary heart disease (CHD). The Panel also recommended a stepped therapy approach consisting of first dietary therapy, then drug treatment with a bile sequestrant agent or nicotinic acid, and finally with a statin, only if cholesterol goals were not achieved with alternative therapies. The rationale for the use of bile sequestrant agents and nicotinic acid as firstline drug therapies' 0*((@@ related to their demonstrated longterm safety compared to the statins. At the same time, however, the statins were noted as highly effective in lowering LDL cholesterol and had a more favorable profile of side effects than the older therapies. The 1994 NCEP panel made some significant changes in the guidelines. First, the 1994 guidelines recommend a more aggressive drug treatment strategy in the case of secondary prevention. This is consistent with the results of CE studies that are cited in the  ? report.m  Y  ? ЍBetween this initial report in 1988 and the revision of the NCEP guidelines in 1994, a number of costeffectiveness studies were undertaken of cholesterolreducing agents (Martens, 1989; Schulman, 1990; Hay, 1991; Goldman, 1991). They are essentially modeling studies which combine cost information from various sources with epidemiological data from the Framingham Heart Study. While these studies have been criticized for methodological shortcomings, they also have received considerable attention in academic and policy circles. A general finding which emerges from these analyses is that the cost to produce health benefits are lowest for groups with high nearterm risk for CHD. This is consistent with the NCEPtargeted approach for drug treatment and the recommendation for more aggressive treatment of secondary prevention in the 1994 guidelines. In addition, the newer HMGCoA reductase inhibitors had more favorable CE ratios than the bile sequestrant agents in a few studies that made comparative evaluations (see, for example, Coyle and Drummond, 1993).m In addition, the guidelines change the step protocol approach previously advocated. In the new guidelines, the statins along with bile acid sequestrants and nicotinic acid are classified as "major drugs" for treating cholesterol reduction. The report lists relative advantages and disadvantages of these alternative therapies, but does not make any recommendations concerning which should be the drugs of choice. This is an issue on which the panel apparently would like to see more research.p 0*((@@ԌMost PBMs that we surveyed have adopted the new guidelines and now list the statins among the firstline drug therapies in their own guidelines. But not all have done so. A forthcoming study by Oster et al. (1996) should provide further useful information to PBMs and other care providers in this regard. It examines the relative cost effectiveness of a step protocol approach versus the use of an HMGCoA reductase inhibitor as the firstline agent,  ? under the conditions of normal clinical practice at a major HMO. ` Y  ? ЍThis study involves a prospective randomized trial being undertaken at Southern California Kaiser Permanente in which half of the patients were randomized to the prior 1988 NCEP step therapy drug guidelines and the other half received Mevacor (Lovastatin) as their firstline drug therapy. The study seeks to approximate conditions of normal clinical practice at Kaiser with regard to provider and patient compliance and drug reimbursement terms. This type of study should address some of the concerns raised above on the need for CE studies to examine patient compliance costs and overall effectiveness under real world practice settings in order to be persuasive to providers. The main lesson that emerges from our cholesterol case study is that the national consensus guidelines have exerted a strong influence on PBM decision processes, and these consensus guidelines in turn have been at least partially responsive to the costeffectiveness studies performed to date on this issue. There are of course several other major therapeutic areas characterized by a range of newer and older drug treatment options (e.g., hypertension and depression) in which costeffectiveness studies could also play a very useful role. It will be important to ascertain where stepped care based on lowercost medicines as the first option represent a costbeneficial approach and where it leads to a sub  0*((@@ԫoptimal result. This is likely a central issue in the application of pharmacoeconomic and health economic studies for the foreseeable future.  ?  C. Disease Management As discussed above, the leading PBMs see disease management as a major area for future growth. Table 7 shows some of the principal disease management targets of the five leading PBMs. They have initiated various programs targeted at several highcost chronic diseases (diabetes, asthma, depression, etc.). In contrast to their treatment guidelines discussed in the last section, these programs will entail very active intervention strategies. A motivating factor behind the PBM movement toward disease management is the premise that medical costs in many diseases are concentrated in a small portion of the affected population. Furthermore, many of these costs are also associated with outofcontrol cases that have greater emergency room use, hospitalization and other resourceintensive interventions. Hence, if one can identify highrisk patients and intervene in advance to change their treatment and compliance patterns, there are potentially large cost savings as well as improved quality of care for patients. The choice of diseases like diabetes and asthma have been selected as disease management targets by all of the leading PBMs because they fit the model presented above and offer large potential economic opportunities to those firms that can successfully implement disease management programs.% 0*((@@ԌA comprehensive disease management program involves data integration, analysis, implementation and outcomes evaluation. Currently PBMs have very comprehensive prescription drug information and potential access to other medical cost data through their clients. A major activity at the present time is the integration of data with compatible coding systems from physicians, laboratories, home health care and hospitals with prescription drug expenditures. Some firms are analyzing health outcomes with the data sets of HMO clients that are already integrated. A typical project might involve a retrospective analysis of integrated medical data bases for asthma patients. The objective would be to identify factors leading to aboveaverage treatment costs (e.g., inappropriate drug treatment or poor patient drug compliance in certain demographic groups) and to develop management techniques to change these adverse outcomes (e.g., a targeted monitoring and compliance program using pharmacists, nurses, case managers, physician education, etc.). Implementation of a disease management program for asthmatic patients would concentrate on controlling disease, thereby reducing costly emergency department visits and hospitalizations. Appropriate patient and provider education would lead to improved compliance and correct administration of inhaled bronchodilators, as well as proper use of antiinflammatory agents such as corticosteroids. An asthma management program should also involve environmental and lifestyle modification. This would include avoidance of triggers and allergens. Prevention of asthma' 0*((@@ exacerbations is a principle objective of a successful disease management program, as documented in the Expert Panel Report on Guidelines for the Diagnosis and Management of Asthma (HHS, 1991). Outcomes analysis is also an essential component of these programs in order to demonstrate their effectiveness and cost savings. Similarly, a disease management program for hypertension would develop protocols to implement primary prevention, screening, and treatment programs. For initial drug therapy, PBMs will need to determine if newer products with higher direct acquisition costs, such as calcium channel blockers and angiotensin II receptor antagonists, are more cost effective than traditional antihypertensive medicines such as diuretics and blockers. The key to a successful hypertension program, however, also involves individualized patient regimens, reinforcement and continuous care, social support, and the collaboration of various health care professionals, as described by the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure (Gifford, 1993). Thus, PBMs will be challenged with not only selecting the appropriate drug therapy for patients, but also will need to focus on lifestyle modification including tobacco avoidance, weight reduction, moderation of alcohol and dietary sodium intake, and promotion of physical activities. These may be accomplished through patient newsletters, direct patient contact,  ?% or innovative patient education programs.7% 0*((@@ԌAnother aspect of disease management involves integrated management of medical and pharmaceutical care through an electronic online network at the point of medical care. While virtually all pharmacies currently participate in oneline networks, only a small percentage of physician practices do. PCS, which was a prime mover in getting pharmacies online, is also taking the lead in this regard with medical practitioners. Their vision for the future is "electronic coordinated care" with directtophysician electronic communications for guiding drug treatments in particular diseases. The interest of PBMs in disease management implies an increased attention on their part to outcomes analysis and costeffectiveness studies. The leading firms have been building a sizeable inhouse expertise with several Ph.D. and other professionals in this area. They have begun a number of their own costeffectiveness analyses, both independently and in conjunction with academic centers, pharmaceutical firms and HMO partners. Their activities can be expected to grow over time as disease management becomes a more central focus of PBMs and other managed  ?  care organizations. x Y ?x ԍThere are a number of interesting industrial organizationtype research issues concerning the emerging role of PBMs in disease management. These include the comparative advantage of PBMs in disease management, how the market will be structured, and how firms will be compensated for performing these activities. These questions are beyond the scope of the current study, but remain important topics for future work.  0*((@@Ԍ ?   IV. Summary and Policy Considerations ă PBMs have evolved rapidly during the 1990s from claims processors and pharmacy networks to the management of drug product selection using formularies and related incentive measures. The main economic drivers of this process have been generic substitution and manufacturer discounts for formulary listings from among closely substitutable drugs. The emphasis has been on lowering drug costs rather than taking a broader perspective on costeffective treatments. PBMs have demonstrated that they can achieve significant cost savings in the drug budgets of employers and managed care entities from formularytype interventions on a carvedout basis. Over the long run, however, management of drug benefits as a separate component, in isolation of other health care activities, has important limitations. By its very nature, it cannot capture the larger cost efficiencies potentially available from an integrated treatment approach to medical problems and diseases. PBMs appear to recognize that the market is likely to evolve into more comprehensive treatment approaches to drug management in the near future. The leading PBM firms have been investing heavily in disease management programs. A key first step in this process is development and analysis of data bases and outcomes information. This process is already underway with PBMs performing a variety of modeling efforts and costeffectiveness analyses of patient data in several target disease areas (asthma, diabetes, depression, etc.). This is likely to expand over time. PBMs can be expected to do' 0*((@@ their own costeffectiveness studies, but also participate in a variety of cooperative ventures with drug manufacturers, HMOs and academic bodies using a variety of methodological approaches. The growth of PBMs can be viewed as part of a broader evolutionary selection process currently taking place in the health care system. Though a trial and error process, the market is currently searching for organizational structures that can provide increased cost efficiency and other performance improvements. In this dynamic evolutionary market, it is relevant to ask what constraints should be placed on information exchanges between the different parties in the market. In a set of draft resolutions, the FDA has proposed that pharmacoeconomic information provided by manufacturers to prescribers and purchasers be treated as part of a drug's labeling. Furthermore, it proposes that any costeffectiveness claims be supported by two adequate and wellcontrolled studies similar to what is currently required in support of a drug's safety and efficacy (FDA, 1995). PBMs have indicated that they want timely CE studies with "real world" medical practice environments and comparators. In the current marketdriven environment, a number of methodological approaches are emerging to deal with these issues including modeling studies as well as experimental and quasi experimental designs. The FDA approach would constrain the availability of`" 0*((@@ information from many studies which are critical to the evaluation  ? of cost effectiveness in real world settings. Y  ? ЍAs many researchers have pointed out, a randomized controlled trial study design provides the highest degree of internal validity, but it also leaves open many questions concerning external validity. This reflects several factorsthe nature of patient and provider populations participating in trials, the presence of protocoldriven costs and outcomes, the use of extraordinary measures to ensure patient compliance, the time horizon used in trials and so forth. Another set of issues relate to the effects on R&D costs and timeliness of FDA requirements. If manufacturers have to do multiple fullscale randomized control trials and receive FDA approval before making costeffectiveness claims to managed care entities, this would not only increase R&D costs but also would delay the availability of useful information to decisionmaking entities. Costeffectiveness decisions are a moving target, and decisionmaking should have the benefit of information in as timely a fashion as possible. In addition, the higher costs would likely  ?0 mean less information for many NCEs with smaller market sales.#0Y  ?  ЍGrabowski and Vernon (1994) have found that the distributions of returns to new drugs is highly skewed. Their work suggests that it would be unprofitable to undertake separate costly CE studies for the majority of new entities.# Based on our interviews with leading PBMs, it is clear that these organizations also have considerable inhouse expertise in pharmacoeconomics and outcomes analysis. They have a healthy skepticism of manufacturergenerated studies and can put their own weights on the claims of manufacturers and the quality of journal publications where this information is published. Indeed, PBMs areH 0*((@@ currently doing various modeling analyses of particular diseases like asthma or diabetes that are exactly the kind of studies for which manufacturers would not be allowed to distribute information. This is not a desirable regulatory outcome. Further research on the costs and benefits of the government regulation of costeffectiveness claims by manufacturers to managed care entities should be a high priority area for future work. It is significant that the proposed FDA regulations are not something being sought by any of the PBMs that we interviewed. It is doubtful that these regulations could pass a benefit cost test from either a private or societal perspective. To the extent that a thirdparty review of costeffectiveness claims is desirable, market alternatives to government regulation should be explored. A possible model worth considering in this regard is the independent auditing of financial data by publicly certified accountants.p0*((@@ %ReferencesЃ X Barradell, Lee B., Ruth Whittington and Paul Benfield, "Misoprostol: Pharmacoeconomics of Its Use as Prophylaxis Against Gastroduodenal Damage Induced by Nonsteroidal AntiInflammatory Drugs," PharmacoEconomics 3(2): 140173, 1993.(# X Blissenbach, Henry F. "Pharmaceutical Services in Managed Care," Ch. 14, in Peter Kongfvedt, ed., The Managed Health Care Handbook, Aspen, Gaithersburg, Maryland, 2nd Edition, 1993, 142160.(# X Coyle, Douglas and Michael Drummond, "Does Expenditure on Pharmaceuticals Give Good Value for Money: Current Evidence and Policy Implications," Health Policy, 26: 5575, 1993.(# X Gifford, Ray W., et al., The Fifth Report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure, Archives of Internal Medicine, 153:154181, Jan, 1993.(# X Goldman, Lee, Milton C. Weinstein, Paula A. Goodman and Lawrence W. Williams, "CostEffectiveness of HMGCoA Reductase Inhibitors for Primary and Secondary Prevention of Coronary Heart Disease," JAMA, 265(9): 11451151, March 6, 1991.(#   X Grabowski, Henry G. and John M. Vernon, "Brand Loyalty Entry and Price Competition in Pharmaceuticals After the 1984 Act," Journal of Law and Economics, 35: 331350, Oct. 1992.(# X Grabowski, Henry G. and John M. Vernon, "Returns to R&D on New Drug Introductions in the 1980s," Journal of Health Economics, 13: 383406, 1994.(# X Grabowski, Henry G. and John M. Vernon, "Longer Patents for Increased Generic Competition: The WaxmanHatch Act After One  Decade," forthcoming in PharmacoEconomics, 1996.(# X Hay, Joel W., Ellison H. Wittels and Antonio M. Grotto, "An Economic Evaluation of Lovastatin for Cholesterol Lowering and Coronary Artery Disease Reduction," American Journal of Cardiology, 67(9): 789796, April 15, 1991.(# X Health Strategies Group, Pharmacy Benefit Management, Palo Alto, California, November 1995.(# X Martens, Leon L., Frans F. H. Rutten, D. William Erkelens and Carl A. P. L. Ascoop, "CostEffectiveness of CholesterolLowering Therapy in the Netherlands: Simvostatin versus Cholestramine," American Journal of Medicine, 87(4A): 545585, Oct. 16, 1989.(#'0*((@@ԌX National Cholesterol Education Program, Report of the Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults, Archives of Internal Medicine, 148: 3669, Jan. 1988.(# X National Cholesterol Education Program, Second Report of the Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults, Circulation 89(3): 13331445, March 1994.(# X Oster, Gerry, Gerald M. Borok, Joseph Menzin, Joseph F. Heyse, Robert S. Epstein, Virginia Quinn, Victor Benson, R. James Diedl and Arnold M. Epstein, "CholesterolReduction Intervention Study (CRIS): Randomized Trial to Assess Effectiveness and Costs in Clinical Practice," forthcoming in the Archives of Internal Medicine, January 1996.(# X Schulman, Kevin A., Bruce Kinosian, Terry A. Jacobson, Henry Glick, Mary Kaye Willian, Harris Koffer and John M. Eisenberg, "Reducing High Blood Cholesterol with Drugs," JAMA, 264(23): 30253033, Dec. 19, 1990.(# X Shaw, Hemant K. "Catalysts for Change: The Impact of Generic Drugs on the Pharmaceutical Industry Spectrum Report 388 Decision Resources, Inc., Burlington, Mass., December 28, 1992.(# X Treppell, Jerry, "Generic Industry Drug Interview," Kidder Peabody, New York, April 4, 1994.(# X U.S. Department of Health and Human Services, National Asthma Education Program Expert Panel Report, "Guidelines for the Diagnosis and Management of Asthma," Bethesda, MD: NIH Publication No. 913042.(# X U.S. Food and Drug Administration, "Principles for the Review of Pharmacoeconomic Promotion," Draft manuscript, March 20, 1995.(# X U.S. General Accounting Office, "Early Results on Ventures with Drug Manufacturers," GAO/HEHS9645, November 1995.(#0*((@@ T ddx !ddx  T  ? & Table 1 )t Leading PBM Companies )t  ? %XJune 1995   +Number of Beneficiaries    Company)Formulary +Plans ((Millions)? ;All Plans :p(Millions) b  PCS Medco Pharmacy Gold Diversified Value Health Caremark Advance Paradigm Express Scripts Diagnostec WellpointX,32 -  ,32 -  ,15 -  +012.9 -  ,l7.5 -  ,l5.2 -  ,l5.0 -  ,l4.8 -  ,l4.5 -  ,l2.0X>P55 ? >P42 ? >P15 ? =12.9 ? >8.0 ? =28.0 ? >6.6 ? >6.9 ? =11.0 ? >5.0b R  Less Duplication*(29.0)=(66)R R X All Other PBM Companies,l2.0>P13R    Industry Total~+093.9~>137   Source: Health Strategies Group0*((@@ O !ddx  Addx O  X ? 'H Table 2 *  Estimated Annual Cost Savings * @ PBM Drug Product Selection Activities * XP  Generic Substitution 610%P P X Formularies (including compliance measures) 515%P P  Concurrent Drug Utilization ReviewH 24%P P  Prior Authorization 12%P  H  Total Range 1431%   Source: Industry publications and interviews with leading PBMs. 0*((@@ J Addx addx  J    *  ?p 'H Table 3 *  Patented Drugs Experiencing Initial *  Generic Competition 19891993 *     Cohort8'Generic Market Share (Units) *Six Months After Entry P   198990 4n31%P P 8 199192 4n44%P    1993X 4n51%   Source: Henry Grabowski and John Vernon (1996) 0*((@@ Y addx  ddx! Y "   "*  ?p 'H Table 4 *  Number of Products on Formulary * l Selective Cardiovascular Classes *      p     Diuretics * )@ACE %Inhibitors 7  5Beta 3,Blockers AhCalcium @,Chemical AhBrokers  P    PBM 1P d 18P *8P 615P C17 P P   PBM 2 d 15 *7 614 C17 P P P  PBM 3 d 12 *5 78 C11 P       PBM 4pd 15p*7p610pC14    Source: Information supplied by PBM companies.!0*((@@ m ddx! ddx" m $   $*  ?p 'H Table 5 *  Formulary Cost IndicesACE Inhibitors * "P " * PBM 1#>PBM 2-RPBM 38fPBM 4CbPBM 5"P P " Monopril0 $0$$$0.$$09$$0D ԩ"P P " Lotensin $$$$$.ԩ9$$D $$"P P 0" Altace  $$ $$$ .$$ 9$$ D ԩ"P P " Accupril  $$ $ԩ .$$ 9$$ D$$$"P P  " Zestrilp  $$$p $$$p .$$p 9ԩp D ԩ"P P  " Prinivil  $$$ $$$ .ԩ 9$$ D $$"P P p " Vasotecf $$$$$$$$.ԩ9$$$D$$$"P       " Capotenf $$$$$$$$-$$$$8$$$$C$$$$    Source: Information contained in leading PBM formularies. "0*((@@ Y ddx" ddx#x Y * D  ?p H' Table 6 ă  ?   What PBMs Would Like to See in CE Studies ă    ? Headtohead comparisons among market leaders  ? Studies performed on relevant populations  ?8 Timely availability of information  ? Independent sponsorship or nostrings funding  ?X Publication in leading peerreviewed journals  ? XMore sophisticated drug representatives who can discuss nuances of a studyƔ"  Source: Interviews with leading PBMs.8#0*((@@ Y ddx#x ddx$$ Y " "*  ?p '  Table 7 *  Common Disease Management Targets * " "   v Asthma"> ^Diabetes)6Cardio (Vascular7 2Depression@Gastro >Fintestinal" P " DPS X"X,X7XC"P P " MedcoH  XH "XH ,XH 7XH CbX"P P " PCS  X "X ,X 7X CbX"P  H " Pharmacy Gold    X "> "X , ,X 7 C" H      " Value Health   X"> "X, ,X7 7XC CbXH    Source: Interviews with leading PBMs.