Estimates of treatment shares without the new drug typically start with estimates of the current treatment shares derived either from market research data such as IMS Health marketing data or from observational databases. There are no standard sources for the estimation of changes in expected treatment shares over the analysis time horizon without the new drug. These estimates may be made by the manufac- turer or the researchers after consultation with physicians who treat the condition of interest based on the current treatment patterns.
One factor that should be considered is that if other new drugs have recently been added to the formulary, their treatment share might be expected to increase over time, taking away treatment share from older, possibly less-effective drugs. If cur- rent treatment patterns have not changed recently, then the current treatment shares might be expected to remain constant.
1 In this chapter, we make the simplifying assumption that the budget-impact analysis is based on the introduction of a new drug to the current mix of drugs for treatment of a condition. Changes in our recommended approaches to estimate the budget impacts of other types of health-care inter- ventions (i.e., vaccines, diagnostics, surgery, and devices) are discussed in Chap. 13.
61
A second factor to consider is if new branded drugs are expected to be launched within the analysis time horizon. If this is the case, an attempt should be made to include estimates of uptake for these drugs in the analysis. Alternatively, the com- puter program created to perform the analysis can be developed to allow the user the option of adding anticipated, new branded drugs to the current treatment shares table and estimating how these shares will be taken from current treatments.
A third factor to consider is if one or more of the current branded drugs are expected to be marketed in generic formulations during the analysis time horizon.
If drugs are expected to go generic, the current treatment mix could include the generic formulation with a 0% market share until the generic form becomes avail- able. Alternatively, the drug could be listed once within the treatment mix but with the user entering detail about the portion of the drug’s share that is due to the branded versus generic formulation. In this case, the mix of branded versus generic formulation of that drug may change over time. Because of the introduc- tion of the generic formulation, it might be expected that the total share of the drug (brand plus generic) will increase with the availability of the generic formulation.
It is important that the assumptions made about changes in treatment share with- out the new drug are clearly stated and are both credible to and changeable by the users of the analysis to reflect their own expectations.
In Box 4.1, we present an example of a study that estimated changes in treatment shares with the introduction of generic drugs.
With entry of either a new branded drug or generic drug over the analysis time horizon, in addition to estimating impact on treatment shares, developers of the analysis need to consider treatment cost and potential impact on clinical outcomes
Box 4.1 Treatment Shares Allowing for Generic Entry
In a budget-impact analysis for understanding the impact of long-acting injectable paliperidone palmitate in the treatment of schizophrenia in Japan (Mahlich et al. 2015), risperidone and quetiapine were available in both branded and generic forms at the onset of the analysis. Olanzapine and aripiprazole were expected to enter the market in generic formulations at different times during the 4-year time horizon of the analysis. The authors estimated the current market shares by combining sales figures, dosage information according to the Japanese label, and estimated future market shares based on market research data. As can be seen from the table below authors expected that a generic formulation of aripiprazole would be intro- duced during the third year of the analysis and a generic formulation of olanzapine during the fourth year. Both drugs continue to have a large mar- ket share in total for all 4 years, although the market share for the branded versions is expected to fall when the generic formulations are introduced.
4 Estimating the Treatment Mix
and population size. For the new branded drug, a price similar to other branded products can be assumed with a premium if the efficacy shown in phase 2 or 3 clini- cal trials is superior to current products. Estimates of clinical efficacy can be taken from the published clinical data, and associated changes in population size can be estimated as discussed in Chap. 3. For a generic drug, published estimates of the branded drug price with discounts based on time since generic entry can be used to estimate the price over time. Efficacy can be assumed to be equivalent to the corre- sponding branded drug.
As mentioned in Chap. 3, care needs to be taken when using market research data or data from observational database studies, as they may not include the needed level of detail on the characteristics of the patient population. If the size of the indi- cated and reimbursement-eligible population has been estimated, then current treat- ment share estimates are needed for the indicated and reimbursement-eligible population rather than in the total condition population. For example, if the indi- cated and reimbursement-eligible population is limited to those for whom first-line treatment has failed, then data showing the treatment shares for all patients with the condition irrespective of treatment history would not be appropriate to use. However, the market research data might only report the treatment shares for all patients with the condition of interest. In this case, assumptions are needed to convert the market research data into estimates of the treatment shares for the indicated and reimbursement- eligible population, possibly using expert opinion. An alternative approach would be to expand the population included in the budget-impact analysis to all patients with the condition of interest (i.e., not just those who are indicated and eligible for reimbursement) and use the market research data on current treatment
Market share in the treatment of schizophrenia in Japan (Mahlich et al. 2015, Table 4)
Treatment
Market share Year 1 (%)
Year 2 (%)
Year 3 (%)
Year 4 (%) Paliperidone palmitate (LAI) (Xeplion) 0.03 0.19 0.35 0.74 Risperidone (LAI) (Risperdal Consta) 0.64 0.45 0.35 0.27 Risperidone (oral) (Risperdal) 12.44 10.92 9.51 8.70 Risperidone (oral) generic entry 10.68 11.74 12.73 13.04
Olanzapine (oral) (Zyprexa) 14.82 15.30 15.53 12.32
Olanzapine (oral) generic entry 0.00 0.00 0.00 3.46
Aripiprazole (Abilify) 13.76 15.35 14.84 11.43
Aripiprazole generic entry 0.00 0.00 1.55 6.07
Quetiapine (oral) (Seroquel) 13.77 10.88 8.49 7.72
Quetiapine (oral) generic entry 4.93 8.01 10.62 11.58
Blonanserin (Lonasen) 3.80 3.97 4.14 4.30
Conventionals 25.13 23.19 21.89 20.37
LAI long-acting injectable
63
mix for all patients with the condition of interest. With this approach, the predicted treatment shares for the new drug and the new treatment mix would be estimated, taking into account that the new drug is indicated and reimbursed only for a subset of the total drug-treated population.
In Box 4.2, we present a hypothetical example of estimates of treatment shares for the different drug-treated populations with attention-deficit/hyperactivity disor- der in children and adolescents.
Box 4.2 Hypothetical Treatment Shares Over Time for Attention-Deficit/
Hyperactivity Disorder in Total Drug-Treated Population and in Those Who Are Intolerant to Methylphenidate or for Whom First-Line Treatment with Methylphenidate Has Failed
The first-line treatment for children and adolescents with attention- deficit/
hyperactivity disorder (ADHD) is generally methylphenidate (MPH), a stimu- lant that may be effective in controlling the symptoms. Both immediate-release and extended-release formulations are available. The immediate-release for- mulations are generic, while extended-release formulations are just beginning to reach the end of their patent life. However, it is not uncommon for patients to have issues with these drugs. Some parents are reluctant to have their chil- dren taking a stimulant. In other cases, children or adolescents may not toler- ate them, or the stimulant may not be effective. In addition, stimulants should not be used when there is a high risk of abuse in the family. Atomoxetine is a drug indicated for ADHD that is not a stimulant and may be used in children and adolescents who should not take stimulants or for whom treatment with stimulants has failed. Atomoxetine is also reaching the end of its patent life.
Since MPH extended release is the most common first-line treatment and is effective in a high percentage of those treated, the current treatment shares in all drug-treated children and adolescents will be different from the treat- ment shares in those for whom first-line treatment with MPH has failed or in those who should not take stimulants. In these population subgroups, the non- stimulants would have a greater treatment share. In addition, since extended- release formulations of MPH and atomoxetine are losing patent protection within the next 5 years, even without a new drug, the treatment shares both in the total population and in those who should not take stimulants or those for whom first-line treatment has failed are likely to change.
The table below provides hypothetical estimates of the current treatment shares for the total drug-treated ADHD population (which could be obtained from market research data) and assumed values for the current treatment shares for the two population subgroups: first-line treatment without stimu- lants and subsequent lines of treatment when first-line treatment with MPH has failed. The table also provides estimates of the treatment shares for the three populations after 5 years. Current treatment shares among those for
4 Estimating the Treatment Mix
whom first-line therapy has failed could be estimated using observational data or expert opinion from practicing physicians. Current treatment shares for those who should not take stimulants could be estimated using expert opinion or assumed to equal the treatment shares for any non-stimulant drug indicated for ADHD. In our hypothetical example, atomoxetine is the only non-stimu- lant drug. There are no good sources other than expert opinion for the change in treatment shares in all three population groups over time, but we have assumed in our hypothetical example that entry of a generic drug will take treatment share away only from the corresponding branded drug.
We have assumed in this hypothetical example that, despite failure of the first extended-release MPH product used, the majority of drug-treated adoles- cents who continue to a second-line treatment will switch to an alternative extended-release MPH product.
Hypothetical estimates of the current treatment shares for the total drug-treated ADHD population
ADHD drug
Treatment share All drug-treated ADHD population
First-line treatment, no stimulants
Failed first-line treatment with MPH Current
(%)
Year 5 (%)
Current (%)
Year 5 (%)
Current (%)
Year 5 (%) MPH, extended release,
branded
75 65 0 0 70 30
MPH, extended release, generic
0 15 0 0 0 50
MPH, immediate release, generic
20 15 0 0 15 5
Atomoxetine, branded 5 2 100 35 15 7
Atomoxetine, generic 0 3 0 65 0 8
ADHD attention-deficit/hyperactivity disorder, MPH methylphenidate