The term "value-based price" has often surfaced in the recent discussion about pharmaceutical prices in Japan. This raises the more fundamental question of how the value of pharmaceuticals and medical devices should be assessed.
These days, there is much assessment of the value of brand names, intellectual properties, and other intangible assets. In assessment of specific intellectual properties, their value may be equated to the difference in corporate value that their possession affords. This approach, however, cannot be applied to pharmaceuticals and medical devices. The reason is that, although an assessment of intellectual properties can be made from the corporate standpoint and in conformity with accounting standards, pharmaceuticals and medical devices must be assessed from the standpoint of society as a whole.
As such, how should one set about assessing the value of pharmaceuticals and medical devices? In our view, their value can be assessed in terms of enhanced longevity and quality of life (QOL; see Figure 1). This is to say that it is a function of the length of time for which a new drug, for example, extends a user's longevity and/or improves their QOL. Are there any other indices for assessing the value of pharmaceuticals and medical devices? Modifications to dosages to facilitate the intake of drugs by patients may be assessed in terms of an improved QOL. Regardless of area (pharmaceuticals/medical devices), all medical practices can ultimately be assessed with reference to longevity and QOL. The concept of "quality-adjusted life years" (QALYs) enables assessment of value using a single index for these two components.
The QALYs concept is simplicity itself, and can readily be grasped by taking a look at Figure 2. On this figure, the horizontal axis indicates the life years, while the vertical axis is QOL. QOL is represented as a number between 0 (death) to 1 (perfect health), known as a utility. Let us consider the change in the longevity and utility values for two different patients.
Patient A was initially in perfect health (utility = 1), but was afflicted by a disease in the tenth year, whereupon their utility value fell to 0.8. After spending a ten years in this state, Patient A was further afflicted with another disease in the twentieth year (from the start of assessment), causing the utility value to fall to 0.5. The utility value slipped to 0.2 ten years later, and the patient died ten years after that.
Patient B, by contrast, was not in very good health from the start, and had a utility value of 0.5. However, the utility value remained at this level for the succeeding 40 years, at the end of which Patient B died.
At the outset, the two patients both had 40 years left to live, but clearly differed from each other in respect of the quality experienced during these 40 life years. Because it does not incorporate a mechanism for taking into account any change in QOL, a survival analysis for the long-term prognosis would unfortunately not draw any distinction between these two patients. QALYs was devised to remedy this shortcoming. The thinking behind it is very simple; it rests merely on weighting the life years by the utility (QOL).
In the case of Patient A, for example, QALYs would be calculated as follows.
* Patient A's QALYs = [10 years · 1.0] + [10 years · 0.8] + [10 years · 0.5] + [10 years · 0.2] = 25 QALYs
The same calculation for Patient B would be as follows.
* Patient B's QALYs = [40 years · 0.5] = 20 QALYs
QALYs has enabled a clear distinction to be drawn between two patients who are not distinguished by mere survival analysis.
Although almost all diseases have an influence on both life expectancy and QOL, clinical trials have conventionally considered only one of these factors. Even where they took both into account, they generally evaluated each separately instead of integrally. While this approach may have been sufficient for clinical evaluations, objective assessment of pharmaceutical value requires assessment based on QALYs.
Value-based pricing demands relative assessment of both the value of the pharmaceuticals or medical devices and the related cost. This sort of assessment is the very substance of pharmacoeconomics. In other words, pharmacoeconomics is the science of setting a price commensurate with value (see Figure 3).
While the value of pharmaceuticals/medical devices can be measured by means of QALYs, a certain practical problem is involved. Let us consider the case of assessment of an antihypertensive agent. Assessments of antihypertensives have been based on their effectiveness at lowering blood pressure as determined in clinical trials. By contrast, QALYs-based assessment requires quantification of the extended life expectancy and change in QOL throughout due to the antihypertensive effect. This, in turn, makes it necessary to estimate differences, due to this effect, in respects such as reduction of strokes, the distribution of degree of disability after strokes, and life years according to the degree of disability. The duration of clinical trials required for production approval or application for registration of new drugs is a few months at the most, and only a few years even in mega-trials. This suggests that it would be utterly impossible to obtain all of the data needed for QALYs-based assessment solely from clinical trials.
In pharmacoeconomics, we use models to overcome the limitations associated with clinical trials. Models may be defined as artificial structures for the flow of treatment and long-term disease prognosis. We use models to estimate the occurrence of events, life expectancy, and related costs (medical service costs, nursing care costs, etc.) over a time frame exceeding the duration of the clinical trials (see Figure 4).
A decision tree model is often used for analysis of acute diseases (see Figure 5). It employs a very simple structure, and constitutes a treatment and disease flow from the left to the right. It is constructed on the basis of conceivable scenarios and their probabilities of occurrence. As shown in this figure, an evaluation of the effect of initial treatment may be made on the third day after its initiation. If the treatment is found to be effective at that time, it is continued. If not, a switch is made to a different drug, for example. There is also a difference in the flow line depending on the probability of insufficient effect. The decision tree enables estimation of the probable cost and life years as a percentage in the case of administration of a certain type of treatment. These are referred to as the "expected cost" and "expected life years".
Analysis of chronic diseases, on the other hand, often utilizes the Markov model (see Figure 6-1). Herein, the prognosis for long-term diseases is divided into several stages, and a simulation is made of the patient's progress through these stages over a certain period of time. Figure 6-1 shows a Markov model concerning the prognosis for a hypothetical disease (disease A). Assuming that the new drug A would reduce the progression rate (and mortality) 50 percent more effectively than the conventional therapy, Figure 6-2 shows the survival curve (over a period of 10 years) resulting from a simulation based on the Markov model. From this simulation, it can be seen that there was a difference of 0.9 years between the two groups in respect of the expected life years.
For analysis of diseases in which two or more conditions proceed simultaneously (such as diabetes mellitus), analysts may perform a Monte Carlo simulation.
Models therefore make it possible to estimate the long-term QALYs and cost. This raises the question of how to evaluate the cost-effectiveness of pharmaceuticals and medical devices.
For example, consider the case of assessment of new drug B compared with existing drug A for a certain disease (see Figure 7). If drug B has a lower total cost (meaning not only the pharmaceutical cost but also all related costs generated over the long term) and higher QALYs than drug A, it is easy to reach a conclusion - drug B is the obvious preference. Conversely, there would be no reason to select drug B over drug A if the former entailed a higher total cost but delivered a lower QALYs.
The such decisions become more difficult when drug B entails a higher total cost but also affords a higher QALYs than drug A (although the reverse situation, i.e., a lower total cost but also a lower QALYs, is also possible, but we omit it because delivery of an equivalent or higher QALYs by the new drug would be the more realistic premise). In other words, drug B would be more effective (QALYs) than drug A, but also carry a proportionately higher expense.
In such a case, analysts would evaluate the cost-effectiveness of drug B with reference to the additional cost required for a 1-QALY extension from drug A. This is termed the incremental cost-effectiveness ratio (ICER). Figure 8 presents an example. In this case, the new drug B would deliver a 2-QALYs extension from the existing drug A, but also entail a cost increase of two million yen. The ICER would be calculated as follows.
* 2 million yen/2 QALYs = 1 million yen/1-QALY extension
The question is whether to consider this 1-million-yen ICER high or low. In several countries, pharmacoeconomics is already being applied for official policy-making in the medical field. For example, the ceiling value for an ICER to be deemed cost-effective is set at about 30,000 pounds in the United Kingdom and 50,000 dollars in the United States of America. This would be equivalent to five to six million yen in Japan. Although there has not been much discussion yet on the subject in Japan, some researchers have reported that the ceiling value in Japan is basically in line with these other countries.
At any rate, while an increase in medical costs is often treated as if it were an unequivocally negative aspect during discussions in Japan, none of the countries officially practicing pharmacoeconomic methods have this kind of outlook. In contrast, they are prepared to accept pharmaceuticals that cost more, provided that they deliver value commensurate with the increase. This approach forms the very basis of pharmacoeconomics.
Model analysis is an extremely powerful tool for estimating cost-effectiveness over time spans that exceed the duration of clinical trials, but it also presents problems. The most significant is the handling of the uncertainty surrounding the parameters applied. In dice, for example, there is no uncertainty about the probability of throwing a "1" (1/6). This is not the case with the efficacy rate for antihypertensive pharmaceuticals. Let us assume that one clinical trial yields a figure of 85 percent as the efficacy rate, but also that the 95-percent confidence interval is 75 ~ 95 percent. In such a case, the true efficacy rate would fall within this interval with 95-percent reliability, but there is no telling exactly where. This is to say that some uncertainty is associated with the 85-percent efficacy rate.
In order to confirm the influence of such uncertainty, sensitivity analyses are conducted in pharmacoeconomic analyses using models. A sensitivity analysis is a way of examining any change in results due to changes in the parameter values applied within a certain scope. In the case of the aforementioned 85-percent efficacy rate, for example, it would proceed by varying this rate in the 95-percent confidence interval (75 ~ 95 percent) and checking the robustness of the conclusion that the pharmaceutical is cost-effective. Such variation in the parameter values naturally leads to a change in the numeric output, but the conclusion that the pharmaceutical is cost-effective will be supported if the ICER, nevertheless, falls within the ceiling value (e.g., 6 million yen). A sensitivity analysis using only one parameter is termed a one-way sensitivity analysis (see Figure 9), while one that simultaneously varies two parameters is a two-way sensitivity analysis (see Figure 10).
- More sophisticated types of sensitivity analysis treat parameter uncertainty as a probability distribution and check the results in terms of probability. These are called "probabilistic sensitivity analyses". The major type is the Monte Carlo simulation (Figure 11). Probabilistic sensitivity analysis enables determination of the probability of the ICER falling under the ceiling value. A graph with the ICER ceiling value on the horizontal axis and the probability of the ICER falling under this value on the vertical axis produces a curve for cost-effectiveness acceptability.
This curve often appears in analyses these days. From Figure 12, for example, it can be seen that the probability of the ICER for the drug under consideration not exceeding six million yen is about 60 percent.
In the preceding sections, we showed that the true value of pharmaceuticals and medical devices can be assessed in terms of longevity and utility, and that QALYs enables assessment by integrating these two factors. We also indicated that long-term estimates can be made by using models and that, in relative evaluations including cost comparisons, ICER enables evaluation of cost-effectiveness. Finally, we described how probabilistic sensitivity analysis enabled assessment of the uncertainty of results in quantified terms.
Assessment of the true value of pharmaceuticals and medical devices through such scientific methods holds enormous benefit in various ways. Obviously, it can be applied in discussion of pharmaceutical prices. Moreover, the techniques are applicable not only to medical drugs and devices, but also to all medical practices.
Additionally, pharmacoeconomics provides precious input in studies of product portfolios by pharmaceutical firms and manufacturers of medical devices. Because QALYs may be equated to pharmaceutical value per se, the aggregate QALYs of a firm's products may be equated to its corporate value.
While a key element of pharmacoeconomics, models can also serve as valuable tools themselves in strategies for development and marketing. For example, various simulations can already be performed for uncertainty related to utility and safety in advance of clinical trials. It is also possible to identify parameters with a significant influence on prognosis and cost-effectiveness, and this can occasion the planning of new studies if evidence for these parameters is not yet available. Furthermore, models structure the flow of treatment and long-term prognosis for diseases, and the thinking they represent can also be applied in discussions of treatment guidelines.
While the proper output of pharmacoeconomics is information needed for making decisions on selection of pharmaceuticals and medical devices as viewed from the standpoint of cost-effectiveness, the techniques are applicable in a diversity of domains.