The market analysis estimates how costs and revenues accrue over time in the development and commercialization of a prospective cocaine medication, and presents plausible scenarios of prospective pharmaceutical companies and their drug development decisionmaking process. For pharmaceutical companies, the risk associated with developing and marketing a cocaine medication is considerable. While it is possible to construct plausible scenarios in which half a million patients per day would be using a cocaine medication, a user population that is half or one quarter this size is more plausible. Notwithstanding recent advances in the science base for cocaine addiction pharmacotherapies, it is unrealistic to expect most pharmaceutical firms to consider investing in such products in the near future.

In general, pharmaceutical firms seek competitive returns on their investment. As noted above, one threshold used in the industry is the net present value (NPV) of the product. Another financial threshold used by pharmaceutical firms is a drug's expected peak annual revenue (PAR). A conservative threshold target PAR for a prospective new product in the pharmaceutical industry is about $200 million for a single drug, although the larger companies tend to seek PARs of $250 to 500 million or more. This report uses a market model to analyze the NPV using plausible estimates of relevant parameters, including market size, market penetration, costs of developing a medication, and price. This model is presented and discussed later in this document.

This section provides national estimates of drug abuse in general and for cocaine abuse in particular; national expenditures for alcohol and drug abuse, and for cocaine abuse and addiction; payment sources for cocaine treatment; and a scenario involving potential revenues for a potential pharmacotherapy for cocaine abuse and addiction.

Need for Drug Treatment in the U.S.

Prevailing national estimates for the need for treatment of substance abuse, in general, and cocaine abuse, in particular, are based on results of the National Household Survey on Drug Abuse (NHSDA), sponsored by the Substance Abuse and Mental Health Services Administration (SAMHSA; Office of Applied Studies, 1996) and are summarized in Figure 4 (above). Due to insufficient coverage of selected populations, e.g., criminal justice populations, NHSDA data likely provide low estimates of treatment need (Gerstein and Harwood, 1990). In order to generate more realistic estimates of need for treatment, most analyses augment the estimates from the NHSDA with data from other sources. Such estimates are used as well to gauge how well the national treatment system is able to deliver care to the population in need, and to estimate the need for additional funding for substance abuse treatment (Figure 5 below).

¹Based on 1980 data with a 16.5% adjustment to account for the population increase from 1980 to 1990. This is an underestimate of cocaine abuse, as cocaine use increased dramatically in the late 1980s.

In the U.S., a specialized treatment system delivers the vast majority of substance abuse treatment services, including treatment for cocaine and opiate addiction (Harwood et al., 1994; Rice et al., 1990; Harwood et al., 1984). The great bulk of substance abuse treatment services are delivered by institutional providers; only small shares of services are delivered by physicians in private practice (including general practitioners and psychiatrists), private practice mental health specialists, and general hospitals.

Based on findings of the 1993 National Drug, Alcohol Treatment Unit Survey (or NDATUS, Office of Applied Studies, 1995), the treatment system includes 11,500 institutional providers, including specialized freestanding clinics, mental health clinics, and specialized hospitals and hospitals with specialized units. Among these 11,500 institutional providers, the overall one-day census, i.e., people currently enrolled in substance abuse treatment, was 944,000 patients (Figure 6). NDATUS estimates that 38,000 of these persons are in the criminal justice population. Setting aside this likely unreliable estimate, there are approximately 900,000 patients currently enrolled in substance abuse treatment.

The total number of people that receive substance abuse treatment services at least once during a year is not directly available from NDATUS because, although all institutional providers provided data for current enrollees, some did not provide data for the total number of patients served during the year. Among the institutional providers that did report on this, 2.8 million patients were served at least once during 1993. Among this same set of institutional providers the overall one-day census was 720,000. This yields a turnover in daily census of about 3.9. However, that is an overestimate of turnover, because some patients enroll with more than one institutional provider during any given year. Based on findings from The California Drug and Alcohol Treatment Assessment (CALDATA; Gerstein et al., 1994), Denmead et al. (1995) estimated the true turnover rate to be about 3.6. Applying this turnover rate to the daily census of 900,000 from NDATUS yields 3.2 million people that were treated for substance abuse at least once during 1993.

Estimate of Cocaine Users in the U.S.

The following paragraph provides a summary of some plausible estimates of cocaine use in the United States. The remainder of this section describes the national surveys on drug abuse from which the estimates of cocaine use are derived.

There are in excess of 2 million addicted or "heavy" cocaine users in the U.S. Of these, about 800,000 to 900,000 enter treatment at least once in a given year. On any given day, there about 250,000 cocaine users enrolled in treatment (i.e., at a residential facility or have been served at an ambulatory treatment center within the previous 30 days). Of these 250,000 currently enrolled users, roughly 150,000 are primary cocaine abusers and 100,000 are secondary cocaine abusers. Thus, of all heavy cocaine users, slightly more than 10 percent are enrolled in treatment on any given day.

While there is no definitive method for estimating the number of cocaine abusers in treatment on any given day, estimates of cocaine users currently enrolled in treatment can best be derived by applying estimates of the proportion of all substance abusers who are cocaine abusers to the NDATUS estimate of 900,000 substance abusers.

Several recent surveys have estimated that up to 17 percent of patients entering specialty substance abuse treatment have primary cocaine abuse (e.g., smoked or other) (Figure 7 below). The State Alcohol and Drug Abuse Profile (SADAP) surveys suggest that the proportion of primary cocaine admits has increased during the 1990s; however, it contains no information on cocaine abuse secondary to other substances. Application of this 17 percent rate yields an estimate of 150,000 daily patients with primary cocaine abuse.

Sources:

TEDS: Office of Applied Studies, SAMHSA, 1997.

SADAP (all): National Association of State Alcohol & Drug Abuse Directors, various.

DSRS Phase I: Brandeis University, 1993a.

DSRS Phase II: Brandeis University, 1993b.

The Treatment Episode Data Set 1992-1995 (TEDS) survey indicates that an additional 19 percent of substance abuse enrollees are secondary cocaine abusers, i.e., are primarily abusing one or more other substances such as alcohol or heroin. This yields an estimated 170,000 secondary cocaine abusers. However, because cocaine abuse may be the third or fourth drug problem for many of these patients, a more conservative estimate of 100,000 is used. Adding this conservative estimate of 100,000 patients to the SADAP estimate for patients with primary cocaine abuse (150,000) yields a combined estimate of 250,000 primary or secondary cocaine abusers.

The SADAP and TEDS surveys must be used advisedly. First, they draw data exclusively from publicly-reimbursed specialty providers. Second, although such providers have 80 percent of the daily patients (among providers reporting funding sources), there is some uncertainty about how complete and representative these surveys are in their coverage of treatment admissions. TEDS reports on 1.3 million admissions and SADAP on 1.9 million admissions, both of which are purported to be a census of admissions to their provider frames. Other studies have estimated that there are over 3 million admissions to specialty treatment providers per year (Harwood et al., 1994; Denmead et al., 1995; Office of Applied Studies, 1997). Because SADAP and TEDS cover publicly-reimbursed patients only, they may over-represent cocaine abuse relative to the entire population, which may have somewhat lower levels of cocaine abuse relative to all types of drug abuse.

The estimate of the number of cocaine addicts varies somewhat. The Drug Services Research Surveys (DSRS) indicate that by 1990 there were in excess of 200,000 cocaine patients per day (Brandeis University, 1993a, 1993b). In linked studies, DSRS examined drug problems of patients according to the one-day census of treatment (Phase I) and annual discharges from treatment (Phase II). Based on Phase I findings, about 30 percent (i.e., 210,000) of the 720,000 patients enrolled with "drug" or "drug and alcohol" treatment providers (excluding "primary alcohol" providers) had a diagnosable cocaine problem. Although little detail was provided for the estimates, it can be deduced that cocaine was the unambiguous primary problem for about 11 percent (75,000) of all enrollees and that cocaine was accompanied by, and may or may not have been secondary to, alcohol for the other 135,000 enrollees.

In Phase II of DSRS, data on type of substance abuse were available for only 81 percent of the 1.05 million discharged patients. Of these patients, 38.5 percent reported having drug or combined drug and alcohol abuse. Application of proportion to the NDATUS estimate of 620,000 current enrollees in drug or combined drug and alcohol treatment (i.e., 900,000 total substance abusers minus 280,000 alcohol-only abusers; Office of Applied Studies, 1995) yields an estimate of about 240,000 cocaine patients per day.

The estimate of a total of 900,000 cocaine users who are treated at least once in a given year is derived by applying the turnover rate of 3.6 (noted above) to the 250,000 persons with primary or secondary cocaine diagnoses. Figure 8 (below) provides a summary of the general characteristics for cocaine treatment in the U.S.

Source: Analysis by The Lewin Group.

National Expenditures on Cocaine Treatment

National expenditures for specialized alcohol and drug treatment were about $6.5 billion in 1993, or $19.71 per-patient-per-day for all treatment (Figure 9 below). Approximately $2.6 billion of this amount was spent on outpatient care (i.e., $1.94 billion and $0.65 billion for outpatient-standard and outpatient-intensive care, respectively), or only about $8.89 per-patient-per-day for outpatient care. In contrast, expenditures for the 103,000 daily enrolled patients receiving 24-hour care (i.e., all modalities except for detox-ambulatory and both outpatient modalities) were $3.87 billion, or about $103 per-patient-per-day. Expenditures for the approximately 80 percent of substance abuse patients who are in standard outpatient care were $7.38 per-patient-per-day.

Source: Office of Applied Studies, SAMSHA, 1995

The estimates of revenue were based on the NDATUS compilation of $3.9 billion from centers that reported revenue, plus an estimated $2.6 billion to account for centers that did not report revenue. Estimates were based on the number of daily patients, by type/modality of care, multiplied by revenue per day (column 4 of the figure above) from reporting "single modality" providers.

Expenditures for specialty treatment for cocaine abuse total approximately $2.1 billion, including about $1.26 billion for the approximately 150,000 primary cocaine addicts and $0.84 billion for the approximately 100,000 secondary cocaine addicts (Figure 10 below). Expenditures averaged $9.00 per day per patient enrolled in standard outpatient treatment and $23.00 per day per patient enrolled in any modality of care (i.e., including outpatient and inpatient care).

Source: Analysis by The Lewin Group. Daily enrollment and annual revenue (imputations for missing data) from 1993 NDATUS (OAS, 1996). Share of clients by modality from TEDS (OAS, 1997). Cocaine share by modality increased by 2/3 to adjust for share of secondary cocaine patients.

Although cocaine addicts account for 17.1 percent of all substance abuse admits, they account for higher percentages in the more intensive and expensive treatment settings such as rehabilitation centers and outpatient-intensive care (OAS 1997). Consistent with development of the estimate of daily cocaine addicts in treatment, the revenue estimate is based on 250,000 daily patients, including 150,000 daily primary cocaine patients and 100,000 daily secondary cocaine patients.

Sources of Payment for Cocaine Treatment

One of the major questions in developing a new cocaine medication is whether there will be sufficient and reliable payment for the medication. Payment for substance abuse treatment, including for cocaine abuse, remains heavily dependent upon government sources. Government provided almost 80 percent of the financing for substance abuse treatment in 1992 (Harwood et al., 1994). This represents only a modest shift to private sector sources since the 1982 NDATUS, which estimated that private sources (insurance plus out-of-pocket) accounted for about 6.5 percent of payments for specialty provider substance abuse treatment (Rice et al., 1990).

The limited sources with reliable data about payment sources in the substance abuse population suggest that only a fraction of patients seeking care for substance abuse treatment have either private insurance or sufficient earning capacity to reasonably afford a medication that would be taken for an extended period of time (e.g., several months or more). According to 1995 TEDS data (Figure 11 below), more than two-thirds of all enrolled cocaine abusers had no health coverage, and another 17.6 percent had Medicaid coverage. Only 9.1 percent of enrolled cocaine abusers had coverage through private insurance (e.g., Blue Cross/Blue Shield, HMO). Wages or salaries were the primary source of income for only 28.1 percent of enrolled cocaine abusers, while 51.9 percent were either on public assistance or had no income. More than three-fourths of the enrolled cocaine abusers were either unemployed or not in the labor force, and nearly 40 percent had not graduated from high school.

Source: TEDS, Office of Applied Studies, 1997.

The Drug Abuse Services Research Survey (DSRS, Brandeis University, 1993) produced similar findings to those of the TEDS data and are arranged by site of care (Figure 12 below).

Source: DSRS, Brandeis University, 1993a, 1993b.

Thus, the source and magnitude of funding for cocaine are key variables to consider in our market analysis. Payment for a new pharmacotherapy may come out of the $2.1 billion now being spent, from additional funds that could be made available, or from funds derived from a combination of existing and new sources.

The potential impact of the cost of a new medication for cocaine abuse may be perceived relative to the current cost of care. The wholesale price of certain commonly used psychotropic medications currently fall in the range of $2 to $4 per daily dose. If the wholesale price of a new cocaine medication is in that range, then its retail cost to payers would represent a large increase relative to current payments for cocaine treatment, particularly the average daily per-patient payment ($9.00) for the majority of cocaine patients in outpatient care (Figure 13 below). As for any health care intervention, the cost of a new cocaine pharmacotherapy should be considered in light of development costs as well as any health and economic benefits that may accrue from its use.

Source: 1996 Drug Topics Red Book. Medical Economics, Montvale, NJ, 1996.

Psychosocial vs. Medical Treatment Model

Substance abuse treatment generally follows a psychosocial, rather than medical, model. The most recent surveys that have examined staffing patterns confirm that the substance abuse treatment system involves little or no physician time in the treatment of patients (Office of Applied Studies, 1993; Brandeis University, 1993). Even when on staff, physicians are often addressing primary health care needs or other mental disorders, rather than providing specialized substance abuse treatment services.

Methadone treatment for heroin addiction would appear to be the most medically oriented model of drug treatment. However, the role of physicians in methadone clinics is generally small and circumscribed to initial diagnostic assessments (i.e., of heroin addiction), management of methadone dosage, and some primary health care services. Most clinic services are oriented to the behavioral and psychosocial needs of the patients, and are delivered by counselors, social workers, and, less often, psychologists (Institute of Medicine, 1990).

The 1991 NDATUS study of specialty substance abuse providers surveyed 9,000 treatment centers (out of a total of about 11,500) with 811,000 patients enrolled on the survey date. The survey identified only about 2.2 full-time equivalent psychiatrists and other physicians, respectively, per 1,000 enrolled patients. Nationwide, there were about 88,000 full-time equivalent direct care staff, including about 1,800 psychiatrists and other physicians, respectively. These estimates were not adjusted for survey and item non-response (Office of Applied Studies, 1993).

The 1990 DSRS survey focused on 7,200 "drug" abuse centers (excluding "alcohol only" centers), serving 540,000 patients with primary or secondary drug problems. That survey found that there were about 1,000 full-time psychiatrists and other physicians, respectively, on staff at the 7,200 centers, very similar to the estimates of full-time physicians from NDATUS. The DSRS also found about 4,500 psychiatrists and other physicians working part-time or on contract; however, the report does not translate these numbers into full-time equivalents or into the proportion of facilities having any physicians on staff. These estimates were not adjusted for item non-response (Brandeis University, 1993).

Involvement of private practice physicians in treating cocaine addiction is virtually negligible. Of all national expenditures for specialized substance abuse treatment services, including alcoholism, less than 1 percent are for psychiatrist visits. Only about 1 percent of all visits to psychiatrists are by patients with a primary diagnosis of substance abuse, including alcoholism; another 2 percent are by patients with a secondary diagnosis of substance abuse (Harwood et al. 1994). Of all national expenditures for substance abuse services, excluding alcoholism, 2 to 3 percent are for visits to physicians of any type (Harwood 1984; Rice et al. 1990).

Comparison of Cocaine Market with Other Markets

The proportion of the entire population of heavy substance abusers that is considered to be a realistic target market for pharmacotherapy is small relative to the effective target market of pharmacotherapy for other diseases. Although there are other clinical conditions for which the market penetration of pharmacotherapy is proportionately low, the absolute magnitude of the overall market may be so large that even such small penetration of a substantially priced drug can be financially attractive to industry, as in the case of the smoking cessation market.

A combination of factors minimizes the attractiveness of the cocaine abuse market to industry, including: a relatively modest potential market (2.1 million heavy users), low proportion of users currently in treatment (250,000 enrollees on any given day), concerns about compliance in this population, and apparent market expectation of a low price point. This appears to be corroborated by the methadone market, where a relatively effective, low-priced medication in a well-established, long-standing treatment system achieves, at best, a 25 percent penetration of the small population of opiate addicts (i.e., about 125,000 patients enrolled in methadone treatment per day out of about 500,000 opiate addicts). The penetration of LAAM is less than five percent of the methadone market, i.e., less than one percent of the opiate addiction market, and the use of naltrexone for opiate addiction falls below that of LAAM. On the other hand, the market for medications to treat the 2.1 million epilepsy patients generates $400 to $500 million per year (IOM, 1995).

Drawing inferences about the potential market for a new cocaine abuse medication from the market experience of other medications for substance abuse must consider the market-limiting characteristics of the treatment systems for opiate addiction to which methadone and LAAM are subject. The market conditions for a new cocaine abuse medication would differ if it is provided via more traditional means of physician prescribing and distribution through pharmacies, rather than if it is a Schedule II or Schedule III controlled substance warranting the forms of controlled treatment delivery required for methadone and LAAM. Nevertheless, their status as medications for illegal substance abuse, similarities in the user populations, and other characteristics make methadone, LAAM, and naltrexone useful, though imperfect, market comparators for a new cocaine abuse medication. As a group, the markets for these medications more closely resemble the potential market for a cocaine abuse mediation than markets for other medications, and industry looks to these markets accordingly when assessing the market for potential cocaine abuse medications.

In comparison to other populations with clinically manifest cancer, heart disease, or other disorders, most substance abusers practice denial. Most abusers do not acknowledge a need for treatment and some even actively resist treatment. Many abusers that do acknowledge a need for treatment fail to seek or sustain treatment. Most patients that enter treatment do so under duress or coercion from legal authorities, employers, family members or friends (Institute of Medicine, 1990). Thus, social policy may be a primary driver of the proportion of cocaine addicts that can be directed to treatment in a year. However, such policies can be badly undermined when patients that do seek to enter treatment are greeted with waiting lists. Once placed on a waiting list, a patient is far less likely to be contacted and admitted when a treatment slot does open.

As noted in this report, it is estimated that there may be almost 900,000 treatment admissions or episodes per year by perhaps 800,000 cocaine addicts, although this number may be smaller due to high rates of relapse and treatment re-entry. Assuming that 800,000 users enter the system at least once during a year, this would constitute nearly 40 percent of the total number of heavy cocaine users, and might appear to be a more substantial market. However, the prospects of retaining this number of patients in ongoing treatment do not appear favorable, given the limitations of a treatment system and characteristics of a population that yield only 250,000 daily enrollees.

The uncertainty about the number of cocaine users, especially the secondary users, places a wide confidence interval around the potential size of the market for cocaine abuse treatment. For the purposes of gauging the potential market for a new pharmacotherapy for cocaine addiction, it may be optimistic to use an estimate of 250,000 current daily enrollees in treatment. Projections of market penetration must consider that some primary users and many secondary abusers may be treated for other addictions and with behavioral therapy to the exclusion of pharmacotherapy, particularly in the context of the current treatment system. Given that many cocaine addicts abuse multiple substances and have diverse health and behavioral disorders, it may be that one or a few medications for cocaine abuse will be insufficient for treating this population. To the extent that multiple medications are needed, the market potential for any one medication would be reduced.

From the standpoint of the pharmaceutical industry, the anti-depressant drug market may serve as a useful contrast in the CNS market to the cocaine market. Depression is a condition that afflicts more than 8 million Americans (Irvin 1997), and depressive disorders have a lifetime prevalence of up to 15 percent for men and 24 percent for women (Hirschfeld 1997).

The consequences of depression include patient mortality as well as direct and indirect costs to society. Those that suffer from depression are at a greater risk for suicide and other co-morbidities including substance abuse, heart disease and other medical conditions. In 1990 alone, the cost of depression was estimated to be between $26 and $43.7 billion (Henry 1997). On average, employers incur an annual cost of $4,200 per depressed employee, of which 28 percent represents treatment costs (Hirschfeld 1997).

In terms of utilization and sales of anti-depressant drugs, 98 million prescriptions were written for these drugs and $4.3 billion in total sales were recorded in 1996 alone (Scott-Levin 1996, IMS 1996). The selective serotonin re-uptake inhibitors (SSRIs) were the most prescribed anti-depressant (54 percent) and captured 89 percent of the market dollars, while the tri/tetracyclics accounted for 5.6 percent of the market dollars despite representing 32 percent of all anti-depressant prescriptions.

The vast majority of these anti-depressants are dispensed through the retail sector (IMS 1996). Retail sales (i.e., through chain stores, independent pharmacies, food stores) accounted for 86 percent of the market in 1996, while long term care, federal facilities, HMOs, non-federal hospitals, and clinics made up the remaining 14 percent of the market.

Price Sensitivity of a Cocaine Medication

Since no approved pharmacotherapy for cocaine abuse has been tested on the market, it is not possible to know how sensitive the market would be to such a medication. However, indirect available evidence from other substance abuse medications and the current nature of cocaine abuse treatment and its financing would appear to indicate that the market would be very sensitive to the price of a cocaine medication.

In a market where the average daily treatment cost is a modest $9.00 per patient for standard outpatient care (accounting for the great majority of patients) and $23.00 per patient across all modalities of care, a cocaine pharmacotherapy priced at a daily dose of a few dollars would represent a significant proportionate cost increase. This may be particularly so in the judgment of substance abuse treatment providers that are vested in psychosocial approaches to the exclusion of pharmacotherapy. It is important to note that the price sensitivity of the current treatment system may vary considerably from that of more typical pharmacotherapy markets that involve physician prescribing and distribution through pharmacies.

The price of methadone may exert some pull on the price point for a cocaine medication. The price for that relatively effective medication, used to treat another stigmatized substance abuse population and paid for primarily by government sources, is a mere 50 cents per daily dose. Although the price of medications for smoking cessation is considerably higher, such medications are not paid for out of public sources, but rather by a self-pay population.

As noted in the case study of LAAM (see Case Studies section), in certain states, Medicaid has reimbursed clinics for the cost of methadone (about $0.50 wholesale per daily dose) but would not cover the higher, though still modest, cost of LAAM (about $2.00 wholesale per daily dose). In such instances, clinics either had to negotiate with the state to receive greater funding or had to absorb the additional cost associated with LAAM. Market resistance to naltrexone, which at about $4.00 wholesale per daily dose is twice as expensive as LAAM yet moderately priced compared to other drugs, was complicated by the additional cost of needed adjunctive psychosocial services. As such, state programs cannot afford to treat populations of heroin addicts with naltrexone ($3,500 per patient annually), which carries double the cost of treatment with methadone ($1,200 to $1,700).

Most existing treatment for cocaine abuse is paid for out of fixed annual government appropriations. Treatment centers that must function with capped budgets and that are judged on the number of patients treated would be expected to be price sensitive. Traditional measures of treatment productivity have been the number of patients treated or the cost per patient treated, rather than measures of improved health outcomes. In the current environment, apparent productivity can be increased by reducing the cost per person served, as occurred in treatment funding from the 1970s to the 1980s (Institute of Medicine, 1990). The introduction of any additional and/or more expensive services reduces the number of patients that can be served with a given budget. In the current treatment and financing context, it is likely that treatment systems and providers will be very sensitive to the price of any new cocaine pharmacotherapy.

Basic Relationships of Price, Market Size, and Revenues

As described above, on any given day there are 250,000 cocaine patients enrolled in treatment, out of an estimated 2 million cocaine addicts in the U.S. If a new medication were used by every current patient in treatment, the medication would have to sell at just over $2.00 per daily dose wholesale ($2.20 to 4.00 per day retail) in order to generate $200 million in annual revenue, as shown in Figure 14 below. If the peak market penetration is only 50 percent of current patients, the wholesale price will have to be over $4.00 per daily dose. On the other hand, if the number of cocaine patients in treatment can be doubled, the necessary wholesale price would only be about $1.10 per daily dose. In light of the importance of achieving significant market penetration, it is important to note that for the drugs LAAM and naltrexone, market penetration in the first several years has not reached 5 percent (see Case Study section).

Decision makers in industry must consider:

• how many patients would use a cocaine medication (as modulated/governed by their providers)
• how sensitive demand will be to price
• how much payment will be forthcoming for the added cost of medications.

If PAR is substantially below $200 million, or some other acceptable industry threshold, there may still be commercial interest if there is reason to expect a sufficient NPV. This will depend primarily on costs of development and production relative to the expected lifecycle of revenue. In order to elicit significant pharmaceutical company development interest at revenue levels below $200 million, the expected costs of development probably would have to be lower than for a "typical" medication. According to 1994 estimates by the former congressional Office of Technology Assessment (OTA), it cost an average of about $150 million in cash outlays, and a capitalized cost in excess of $350 million over 12 to 15 years to develop a successful new medication, including the costs associated with development of drugs that never reached the market. The cash outlays for any given drug that did reach the market, not including the burden of cash outlays for drugs that did not ultimately reach the market, were generally less than $50 million per successful drug in the OTA report.

Translating Market Barriers and Policy Options into Financial Parameters

As described in this report, pharmaceutical companies' decisions to pursue new products is based largely on financial considerations inherent in risk-reward tradeoffs. Despite the great diversity of market barriers, most can be interpreted as having a direct effect on one or more financial parameters that are factored into these decisions. Similarly, most policy options that exist or that could be implemented can be interpreted as having a direct effect on these financial parameters.

The figures below portray likely relationships between the particular market barriers (Figure 15) and policy options (Figure 16) identified in the 1995 IOM report and six basic parameters relevant to decisions to pursue a new therapy: R&D costs, (time to) product launch, marketing and distribution costs, market size/penetration, price, and (duration of) market life. These relationships were used in part to develop the scenarios shown later in this report. In particular, these relationships were used to translate market barriers, and various means that might lower these barriers, into the financial parameters used to generate the quantitative indicators of PAR and NPV. (Note: the use in Figures 15 and 16 of the market barriers and policy options cited in the IOM report does not imply that this study confirms all of those market barriers and policy options.)

Market barriers derived from IOM (1995).

*No longer applicable.

*Policy options derived from IOM (1995).

Modeling Financial Return

The level of interest on the part of industry in developing a new drug abuse medication depends primarily on financial criteria. Factors that may not appear to have direct financial import, such as social stigma associated with a product, inclination of caregivers to consider pharmacological as opposed to behavioral interventions, or corporate commitment to further the greater societal good, do have financial implications that are considered by companies. The financial impacts of such factors can be estimated and incorporated as such into decisions about pursuing projects.

Companies often have summary financial targets or hurdles that drive investment decisions. Two that are often used in industry and which are used in this report are net present value (NPV) and peak annual revenue (PAR). These and other related terms are defined in Figure 3 (earlier in the Methods section).

In principle, companies pursue projects that have positive NPVs. In addition, larger companies are less likely to be interested in a new drug for which projected PAR is less than $200 to 300 million. Of course, alternative projects that offer higher NPVs and/or PARs tend to be more attractive to companies.

Gauging whether a medication might generate a given level of annual revenue to a company is relatively straightforward, as it depends primarily on the number of patients taking a given amount of medication per day and the wholesale price (i.e., paid to the company) per day of the medication. Figure 17 (below) presents this fundamental relationship.

Source: Analysis by The Lewin Group

As shown in Figure 17, a new cocaine medication could achieve $200 million in revenue with 150,000 daily patients (i.e., who are on a prescription on any given day) and a wholesale medication price of $4.00 per day, or with 250,000 daily patients and a wholesale price of about $2.00 per day. In order to achieve $200 million at a wholesale price of $0.50 per day, there would have to be more than 1 million daily patients. The graph can be interpolated or extrapolated to estimate the number of patients that would be required at other prices in order to achieve an annual revenue of $200 million.

Market Penetration, Price, and Annual Revenue

In order to demonstrate basic relationships between market penetration, price, and annual revenue, the following analysis poses three levels of market penetration and three price levels for a new cocaine medication. The market penetration levels are "doubling," "saturation," and "half entry," as follows.

• Doubling the current number of cocaine abusers in treatment to 500,000 patients per day might be achieved if a national program for long-term maintenance treatment was developed. Such a level of market penetration would be equivalent to that of methadone with opiate addicts, i.e., penetration of approximately 25 percent of heavy users.
• Saturation of the current 250,000 patients per day (i.e., the current number of patients plus recently discharged patients) would be a high rate of market penetration. This would involve acceptance of patients for medium-term treatment (3 to 6 months), though not for indeterminate duration maintenance.
• Half entry of 50 percent of the current daily census of cocaine users in treatment would yield 125,000 patients. This would involve medication accepted for use while patients were under care of specialty or health providers (for an average of about 3 months), but not after termination of formal treatment.

These levels of market penetration and service delivery are patterned on currently existing service configurations in the national treatment system. As such, these levels are based on the estimated number of patients currently enrolled in treatment (i.e., 250,000), rather than on the estimated number of cocaine users that have been treated once or more in a given year (i.e., 900,000).

The market penetration levels are linked to payment levels in order to derive variables of financial return to pharmaceutical companies. For purposes of illustration, three pricing levels are as follows.

• $2.50 wholesale per daily dose
• $1.25 wholesale per daily dose
• $0.50 wholesale per daily dose

The wholesale price of $2.50 per daily dose is lower than that of naltrexone, a medication for both opiate addiction and alcoholism, which is about $4.55 per daily dose. The wholesale price of $0.50 is used as a lower price limit; it is roughly what methadone programs pay for their medication. (It is unlikely that a new pharmacotherapy for cocaine abuse would be priced so low.) The price of $1.25 is halfway between the other two, and is near the low end of the price range for LAAM. As a point of reference, providers pay about $1.30 - $2.00 per daily dose of LAAM (i.e., 180 - 280 mg/week at $0.05/mg.).

Estimated revenue to a company is based on the wholesale prices, i.e., the price per dose paid by providers (e.g., pharmacies or drug treatment clinics, depending on how a drug is dispensed) to the pharmaceutical company. However, the retail prices (i.e., that cover the cost of medication plus related costs of storage, handling, and dispensing) that are charged to patients and/or that are passed through to payers (insurance companies, Medicaid programs, other payers) are relevant as well, as any payer reluctance to pay these retail amounts may affect use of a drug. Retail markups vary widely, but are typically on the order of 10 - 100 percent.

In the current context of treatment, a new cocaine medication would represent additional payments at retail prices paid by third-party payers (government or private sector), patients, or other sources. The estimates for retail prices for medication should be considered relative to their anticipated net impact on current spending levels for treatment noted above, i.e., an average of $23.00 per day across all 250,000 enrolled cocaine users and an average of $9.00 for the majority of cocaine users enrolled in outpatient programs. At a wholesale price of $4.00 per day, a retail markup for a typical Medicaid program of 15 percent would yield a retail price of $4.60. Assuming that the cost of the medication would add to, and not substitute for, the current cost of care of $9.00 per day, the cost of the new medication would represent to the payer a 51 percent increase in the daily cost of treatment of cocaine abusers. In comparison, at $0.50 per day, the cost of methadone is only a small portion of the $5 to $20 daily cost of treating opiate addicts.

Peak Annual Revenue

Figure 18 (below) shows expected peak annual revenue to a pharmaceutical company under the nine combinations of market penetration and wholesale price. Under the most optimistic combination of these, the PAR to the pharmaceutical company would be $455 million, well above most companies' acceptable PAR thresholds. However, achieving this PAR would require the following.

• There would be 500,000 cocaine users (double the current 250,000) enrolled in treatment daily, representing about 20 to 25 percent of the estimated total number of heavy cocaine users.
• Payers would be willing to pay roughly $2.75 to $5.00 retail for medication per day per enrolled user.
• Approximately $500 to 600 million of funding for treatment with the new drug (i.e., the retail payments for $455 million wholesale of the drug). This would have to be realized in the form of new funding, reallocation of funds from the $2.1 billion currently spent for treating the 250,000 patients currently enrolled in treatment, or a combination of these.

Achieving these assumptions would likely require creation and funding of a new national substance abuse treatment program for cocaine users. Precedent for such new funding exists in the creation of the national methadone maintenance treatment system in the early 1970s, when nearly 100,000 methadone treatment slots were funded de novo. This was equivalent to 20 to 25 percent of the estimated number of heroin addicts at that time (also the current penetration of methadone treatment). This major funding initiative was supported primarily by the federal government and entailed, as planned, a phase-down of federal dollars over the subsequent five years.

A PAR of $227.5 million could be achieved through: 1) the doubling of current market penetration of cocaine treatment (likely requiring a national cocaine maintenance system) with the middle-tier price or 2) current treatment levels with the higher price. Another way to realize this estimate would be if two-thirds of current daily patients and about one-third of the recent departures from treatment took the medication. Under this scenario, patients would take the medication for an average of three months during treatment, and half might continue for another three months. (These average times for taking medication are illustrative only. Such determinations would necessarily reflect such factors as whether a drug is intended for short-term use, e.g., for detoxification, or indefinite maintenance, as well as patient drop-out rates, which can be considerable.) As noted above, achieving this estimate would require a major infusion or diversion of resources, equal to about 15 percent of current funding for treatment of cocaine abusers.

A PAR of $100 million per year could be achieved under more moderate, though still ambitious, scenarios. Market penetration of 50 percent would require a wholesale price of $2.50 per day to generate this level of revenue. If a new medication reached the current number of daily enrollees, the price can be about half as high, i.e., $1.25 wholesale. Doubling the current number of treated patients would yield $91 million per year at a retail price of only $0.50 per day.

The time to achieve PAR reflects the level of investment in a new drug, and affects the number of years during which peak annual revenues can be sustained, i.e., before a product loses its market exclusivity. This is addressed in the sample scenarios later in this report.

Internal Rate of Return and Reducing the Time to Market

Another important investment criterion used by some companies is the internal rate of return (IRR). The IRR is the rate of interest at which the present value of all net cash flows into and out of a project over a specified time interval equals zero. A higher IRR makes an investment opportunity more favorable.

According to its report on the pharmaceutical industry, the former congressional Office of Technology Assessment (OTA) estimated that the pharmaceutical industry achieves IRRs in the range of about 14 percent (OTA 1993). To maintain such a rate, companies may seek individual investment opportunities with higher rates, e.g., above 20 percent, to account for the risk associated with drug development efforts, the majority of which do not ultimately yield marketable medications.

Consideration of acceptable IRR for prospective cocaine medications readily reveals two key findings. First, if the cost of developing a successful cocaine medication is comparable to OTA estimates of the recent cost of developing other medications, pharmaceutical companies will be hesitant to take on such efforts early in the R&D process. OTA estimated that the fully capitalized cost of a medication at launch was as much as $359 million in 1990 dollars, equivalent to cash outlays of $135 million over 13 years. When these factors are applied to the market scenarios described above, not even the optimistic doubling scenario at the high wholesale price of $2.50 per patient per day meets the 20 percent IRR threshold, let alone the more conservative saturation and half-entry scenarios.

Second, pharmaceutical companies may be more interested in a late-entry scenario in which a company takes on further development after initial R&D - and its inherent risk - have been conducted (or sponsored) by another entity. Late entry might occur if, as has happened with other medications, the government were to complete the early years of product development and then turn over the rights of the product to a company.

As indicated in Figure 19 (below), a late-entry scenario that requires only 8 years to complete development of a cocaine medication, at an estimated total cash outlay of $70 million, may be more favorable. This scenario, entailing a shorter development time at lower cost, may be attained when the government, another company, or some combination of these has invested in early R&D. Thus, a large pharmaceutical company may be more willing to acquire or otherwise collaborate with a smaller firm with a substance abuse compound that has already passed milestones in the R&D pipeline, requiring less investment and fewer years to market. This scenario may also occur when a major advance in the science base has occurred that accelerates the drug discovery and development process. New research indicating that medications in development for other central nervous system disorders (mental illness, analgesia, anesthesia, etc.) have applications for substance abuse as well could shorten the development time table and lower costs.

Note that even the least ambitious penetration market scenario achieves the 20 percent IRR threshold value if the cost of the development phase is sufficiently reduced, Figure 20 (below).

Affordable Levels of Development Cost

The final approach taken to assessing the financial attractiveness to commercial firms is to estimate the affordable level of R&D cost, given the expected peak annual revenue to the company. The base conditions for the model have been used to develop the estimates in Figure 21 (below). In particular, this set of calculations assumes that a company has a potential medication that has completed preclinical development, and only has an expected 8 years until product approval and launch.

The graph represents, for a given PAR, the maximum level of expected development cost that would allow a firm to anticipate a 20 percent IRR. For example, if the acceptable threshold for peak annual revenue of a cocaine medication is as low as $50 million, the maximum anticipated cash outlay would have an expected value of $18 million over 8 years. The expected value takes into account the risk of failure of the development effort; thus, the actual outlays could be as low as $5 - 6 million for a single potential compound.

In this model, only at a PAR near $400 million would a pharmaceutical firm expect a "typical" development effort with cash outlays (adjusted for risk of failure) to have good prospects of yielding a strong payoff. Below $400 million per year, the company would have to either anticipate lower than average development costs, or a higher than average probability of a successful development effort.

Earlier in this report it was suggested that a major pharmaceutical company would want to anticipate a potential peak market of $200 million in order to seriously consider a development project. At this level of expected revenue the company would still have to have very positive prospects for development of a medication in order to proceed. Either there would have to be reason to believe that costs could be kept lower than for an "average" medication development effort, or they would need to anticipate a higher probability of success. These calculations are predicated on the assumption that the annual revenue target will be achieved through successful entry into the market at an adequate price.