Patent Application
20080065411
An embodiment of the invention is herein described, by way of example only, with reference to the accompanying Figures and Tables in which:
Table 1 shows base case and test-adjusted case settings for personalised medicine revenue drivers in a hypothetical example relating to the diagnosis and treatment of genital herpes;
Table 2 shows a series of test uptake values used for developing a series of test uptake scenarios to test the method and system for developing a personalised medicine business plan of the present invention on a hypothetical example relating to the diagnosis and treatment of genital herpes; and
Tables 3a, 3b and 3c comprises profit and loss sheets generated by the method of the present invention using the personalised medicine revenue drivers (shown in
For simplicity, in
The following discussion will provide a broad overview of the architecture and function of the present invention. This will be followed by a brief description of key variables (or personalised medicine revenue drivers) used in the present invention. The discussion will continue with a more detailed discussion of the calculations performed during a first and second operational phase of the invention. The discussion will finish with a description some experimental results generated by the method of the present invention.
The present invention uses references from the literature on medicine, pharmaceutical marketing and diffusion theory applied to the healthcare industry to develop a hybrid predictive-modelling case-based reasoning approach for predicting revenue from a personalized medicine business plan and validating the same. More particularly, and referring to
The present invention has two operational modes. In the first operational mode, the revenue prediction model 10 is provided (by a user 22) with values 12 from an intended personalized medicine business plan. The revenue prediction model 10 is also provided with values of variables (known as personalized medicine revenue drivers 14) that have been previously identified as being key to determining the revenue from the joint marketing of a therapy and theranostic. The values of the personalized medicine revenue drivers 14 may be provided by the user 22 or may be calculated from the precedents in the case-study archive 20. The revenue prediction model 10 integrates the personalized medicine revenue drivers 14 and user-inputted variables 12 in a tailored, non-linear fashion to predict the revenue 16 from the intended personalized medicine business plan. The complex non-linear relationships represented in the revenue prediction model 10 contrast with the simpler linear (cascade) relationship models traditionally used in pharmaceutical modelling and prevents the present invention from over-simplifying the impact of a theranostic on a therapy. The revenue 16 is forwarded to an investment return model (not shown) whose function will be described later.
The CBR comparator 18 collates the user-inputted values 12, personalized medicine revenue drivers 14 and predicted revenue 16 into a hypothetical business scenario. The CBR comparator 18 compares the hypothetical business scenario with the precedents stored in the case-study archive 20 to identify the precedent that most closely matches the hypothetical business scenario. Details are extracted from the closest matching precedent to provide supporting evidence 24 for the hypothetical business scenario and guidance on how the intended personalized medicine business plan might be achieved. More particularly, chosen values for the variables used to generate the business scenario are compared with values from the case-study archive 20 to justify/validate the values chose for said variables. This contrasts with traditional pharmaceutical business models that focus on revenue prediction and do not provide any guidance on strategies for achieving the required financial target.
In the second operational mode, the revenue prediction mode 10 is provided with user inputted values 12 from an intended personalized medicine business plan and the revenue 26 desired therefrom. The revenue prediction model 10 uses these variables to calculate the values of the personalized medicine revenue drivers 28 required to reach the desired revenue 26 and/or the timescale 30 in which the desired revenue 26 can be achieved.
As in the first operational mode, the CBR comparator 18 collates the user-inputted values 12, calculated personalized medicine revenue drivers 28, desired revenue 26 and timescale 30 into a hypothetical business scenario. The CBR comparator 18 identifies a precedent from the case-study archive 20 that most closely matches the hypothetical business scenario and extracts details therefrom to provide evidence 24 in support of the hypothetical business scenario (and guidance on how the intended personalized medicine business plan might be achieved).
The personalised medicine revenue drivers 14, 28 of the present invention are derived from a thorough understanding of both the diagnostics and pharmaceutical markets and contrast with the drivers used in traditional pharmaceutical modelling that do not consider a test's impact on therapy revenue. Furthermore, since the revenue prediction model 10 includes multiple drivers of theranostic impact on therapy, the model embraces the highly variable and unpredictable nature of diagnostic clinical utilization as it impacts a linked therapy. This contrasts with traditional pharmaceutical revenue models, which assume that inefficiencies in the diagnostic market are not a major factor in determining the revenue from a therapy.
The revenue prediction model 10 is developed using “real-world” or historical, peer-reviewed, published case-study data. Similarly, the CBR comparator 18 validates predictions from the prediction model 10 using precedents (e.g. from a seven year revenue stream for a blockbuster drug and a nine-year revenue stream for a speciality therapy) in the case-study archive 20. These two features further differentiate the present invention from traditional pharmaceutical revenue prediction models that are based on market research data (of hypothetical future doctor reaction to drug value propositions) and tend to be highly subjective (and/or fail to reflect real market conditions). More particularly, by being based on actual “real-world” historical data, the present invention removes the uncertainty and subjectivity associated with traditional pharmaceutical revenue prediction models and provides the pharmaceutical industry with a peer-reviewed, validated, repeatable, benchmarked methodology for the repeated assessment of diagnostic impact on a targeted therapy, which more realistically reflects the market conditions of personalized medicine.
Whilst one of the primary functions of the present invention is to predict the revenue obtainable by allying a therapy with a theranostic, the invention also generates estimates of other financial variables which describe the overall shape of the market impact of the therapy-theranostic alliance and are instrumental in understanding how a predicted revenue may be achieved. The other such financial variables include estimates of:
For example, consideration of different scenarios might determine what factor affecting sales of the therapy the theranostic is chosen to impact (e.g. diagnostic efficiency, adherence).
The modeling methodology of the present invention enables further case studies to be added to the case-study archive 20 as they become available. This in turn will lead to an improvement in the performance of the CBR comparator 18 and the provision of more accurate business intelligence to users. Accordingly, the present invention contrasts markedly with traditional pharmaceutical modelling methodologies, in which one-off models are built and the knowledge acquired therein is not carried over to other models. Furthermore, since the modelling methodology of the present invention is not tied to a specific medical disorder, the present invention can be generalised to a wide variety of therapeutic areas.
The variables used as personalised medicine revenue drivers are described in more detail below. However, it will be appreciated that the present invention is not limited to these specific variables and could instead be implemented with any other suitable variables. In particular, additional personalised medicine revenue drivers may include the erosion of sales of a given therapy and/or theranostic after Patent protection therefor has expired.
2(a) Responders (Percentage of Patients Taking a Test Who are Most Likely to Benefit from the Therapy)
The ability to identify patients most likely to benefit from a drug prior to initiating therapy can have both positive and negative effects on a drug's profitability. In particular, the ability to identify patients more likely to respond to a therapy prior to enrolment can significantly reduce the size, duration and cost of a clinical trial. On the other hand, such pre-screening reduces the potential patient population for a drug, particularly if third-party payers begin to mandate screening in an effort to control high drug costs.
2(b) Screening Effect (Diagnostic Efficiency)
The ability to properly diagnose patients and begin therapy is a critical variable in the market size of a drug. Consequently, a diagnostic, that is targeted toward a normally under-diagnosed disease can have a positive effect on the overall market size of the corresponding therapy. In the present invention, diagnostic efficiency is determined through an estimate of the number of patients diagnosed with a condition using a given test or test modality compared to an estimate of the total number of patients both diagnosed and undiagnosed with that condition (using other inferior testing methodologies or empirical analysis).
2(c) Price Premium (Effect of a Diagnostic on the Pricing of a Therapy)
Despite the reduction in market size produced by targeting therapies, it may be possible to demonstrate significant therapeutic advantage to such targeted groups, wherein this advantage is sufficient to support premium pricing for the relevant therapy.
2(d) Competitive Market Share Advantage (Percentage Share of Total Available Market)
Targeted therapies provide companies with an opportunity to gain an advantage in the competition for market share that is not tied to promotional strategies. In particular, targeted therapies offer physicians a clinical basis from which to determine the best treatment option for their patients. This in turn, increases switching costs for the patient and physician and reduces the likelihood of losing market share to similar drugs. Similarly, in smaller therapeutic areas, or areas that feature less competition, a targeted therapy offers the potential of gaining market share from drugs that arrived earlier to the market and would otherwise hold the greatest share thereof. In the present invention, competitive advantage data is taken from published data for over 129 pharmaceutical cases (Grabowski and Vernon) and modulated according to whether the therapy achieves top decile, second decile or average market share.
2(e) Rate of Patient Adherence with Therapy and Effect of Intervention Thereon
A significant factor in patient adherence is the perception of benefit from a therapy. Many patients stop taking their therapy because they do not perceive the lack of a negative effect as a significant positive benefit. Clinical evidence from diabetes, HIV and coagulation management studies suggests that when diagnostic monitoring tools make patients aware of their progress, their adherence improves. Thus, it can be beneficial for a pharmaceutical company to seek a closer alignment with diagnostic monitoring tools, either to help achieve correct therapeutic dosages or to make patients feel that they are in control of their condition or improving, thereby encouraging patient adherence and driving revenue.
2(f) Early Adoption (Speed at which a Therapy Achieves its Maximum Market Share)
A novel therapy can take three to four years (or more) to be adopted. Thus, this process has a direct impact on when a drug will reach its peak-year sales. The availability of a diagnostic that identifies a patient's candidacy for therapy (or demonstrates the value of treatment through post-therapy monitoring) removes some of the uncertainty associated with a novel therapy, thereby influencing adoption rate and revenue of the new therapy. In other words, with a theranostic approach, adopters who would otherwise have waited for evidence that a drug would work for a particular patient will adopt the drug earlier, because the patients for whom it has been proven that the therapy will work have already been identified by the test (thereby removing the need to “wait and see”).
2(g) Early Launch (Effect of Diagnostic on Launch Date of Therapy)
Diagnostics could cut the time from target identification to drug launch from 10-12 years to 3-5 years, thus reducing pre-launch development costs per drug to about $200 million. This is achieved because the availability of a test will reduce the number of patients required to power a clinical trial, thereby accelerating the trial, facilitating regulatory approval and thus reducing the overall time to market.
Referring to
The personalised medicine revenue drivers 14 and other user inputted values 12 are used in the revenue prediction model 10 to calculate the combined sales 32 of the desired therapy and theranostic. The revenue prediction model 10 also calculates the uptake 34 of the theranostic and corrects the combined sales 32 calculation accordingly. The corrected sales figures are used to calculate the revenue from the joint marketing of the therapy and theranostic.
The predicted revenue 16 is then forwarded to an investment return model 36, whereupon it is combined with other financial variables 38 provided by the user to calculate the net present value (NPV) 40, return on investment (ROI) 42 and internal rate of return (IRR) 44 from the intended personalised medicine business plan. The NPV 40, ROI 42 and IRR 44 variables are output to the user in graphical or tabular format of value per year from launch; or as a series of scenarios calculated over a given time period. It will be appreciated that the present invention is not limited to these NPV, ROI and IRR output variables and could instead provide other metrics for assessing investment benefits.
A more detailed analysis of the operations leading to the calculation of the combined sales of the desired therapy and theranostic is shown in
An intermediate value 64 is calculated from the potential market size 62 and the personalised medicine revenue drivers comprising the competitive advantage 66, early adoption 68, effect of the theranostic on the launch date 70 of the therapy and the effect of the theranostic on the pricing 72 of the therapy 72. The intermediate value 64 is in turn used to calculate the sales 32 of the desired therapy and theranostic. The sales may be calculated for a user selected period of time 74 encompassing the entire sales cycle of the therapy or a portion thereof.
A more detailed analysis of the operations leading to the calculation of the theranostic uptake 34 is shown in
As previously mentioned, during the second operational mode of the present invention, the revenue prediction model is provided with user inputted values from an intended personalized medicine business plan and the revenue desired therefrom. The revenue prediction model uses these variables to calculate the values of the personalized medicine revenue drivers (within their case limits) required to reach the desired revenue (e.g. $300 m in year 3 sales). To achieve this, the values of the personalized medicine revenue drivers are seeded with mean values from the case-study archive and then adjusted through a feedback loop in accordance with the:
In addition, as previously discussed, during the second operational phase, the revenue prediction model may indicate that a target revenue is not achievable or not achievable within the required time interval (e.g. will take 5 years rather than the desired 3). Furthermore, the personalized medicine revenue drivers may be fixed to the mean values (from the case history archive) and the revenue prediction model used to show the revenue generated using such mean values. Alternatively, the personalized medicine revenue drivers may be fixed to maximum and minimum values (from the case history archive) and the revenue prediction model used to show the span of achievable revenues therewith.
5.1 General Example
The method of the present invention was used to predict the number of patients treated with a particular therapy (wherein the prediction also considers the effect of the sales of a theranostic for the relevant condition). The prediction from the method of the present invention was compared with a similar prediction made by a traditional pharmaceutical forecast model (that does not take into account the effect of a theranostic).
Referring to
The business intelligence provided by the CBR comparator of the present invention suggests that in order to achieve the penetration targets shown in
More particularly, the business intelligence provided by the CBR comparator of the present invention suggests that in order to achieve the penetration targets shown in
The method of the present invention was also used to optimise resource requirements as opposed to test uptake scenarios. The results from this exercise suggested that the number of patients treated between 2010 and 2012 could be increased by 11%. However, the increase in sales in years 1 and 2 predicted by the method of the present invention remains lower than that predicted by a traditional pharmaceutical forecast model.
The business intelligence provided by the CBR comparator of the present invention suggests that in order to achieve the desired penetration targets, a minimum of 40% of target providers must behave as innovators and early adopters of the theranostic by the end of year 1, using this on 95% of their patients. More particularly, the business intelligence provided by the CBR comparator of the present invention suggests that in order to deliver this best in class market penetration:
The above modelling exercise has confirmed the benefit to a therapy of focusing resources on sales of the theranostic as early as possible. In particular, the modelling exercise showed that approximately every 1$ spent on theranostic sales and marketing to innovators and early adopter providers, translates into $5 in therapy sales. This compares well with successful direct to customer advertising campaigns, wherein $3 revenue is generated for every $1 spent on TV advertising.
The method of the present invention was also used to predict the benefit to a laboratory partner of supporting a therapy. The method of the present invention estimated that the testing market value will increase significantly, offering between £34-£50 m contribution to overheads (assuming 65% profit margins) to laboratories servicing this market.
5.2 Genital Herpes Example
Genital herpes is estimated to affect, on average, 25% of the US adult population (i.e. approx. 50 million people). The vast majority (approximately 90%) of these people do not know they are infected. Thus, there are approximately 45 million US adults who have genital herpes, but are unaware of it. Of these people, approximately 25% will be truly asymptomatic. Thus, approximately 34 million US adults will display some form of symptoms at some time during a year. Assuming people will only present to their physician during an outbreak of the disease, and assuming 5 outbreaks of 7 days duration per year on average; approximately 3.4 million people per year will present to their physician for testing (however, in reality, the figure is considerably higher).
The standard diagnostic workup for genital herpes involves clinical examination, taking a patient history and viral culture of lesions (if present). The standard diagnostic approach has a relatively poor sensitivity of around 50%. However, a new means of diagnosing genital herpes caused by HSV-2 (around 90% of cases) has been recently developed using type specific serology (TSS), wherein TSS has a much higher sensitivity (typically >90%).
At present, there are no drugs available to cure herpes. However, for the sake of example, assume a company X produces a novel drug, Simplavir, that eradicates the virus from a patient. In this hypothetical example, a one year course of Simplavir (necessary to give >90% clinical efficacy) produces $450 per patient for the company X. The present invention is used to advise company X on how shifting the diagnostic paradigm away from exam and culture to TSS could impact their sales of Simplavir.
The TSS test chosen is called Oran2; it is a lateral flow test designed to be used at the point of care (thereby eliminating the risk of patients not returning for the results of their test). Average sensitivity across all patients is 90%.
Table 1 shows base case and test-adjusted case settings for the personalised medicine revenue drivers (wherein the asterisk superscripts above values indicate that the value shown is a mean values from all the cases). Values for diagnostic efficiency in a base case and test-adjusted case are supported by 4 and 0 cases, respectively (the value for the test-adjusted diagnostic efficiency is flagged as beyond the range of values that the stored case-studies supports). A series of test uptake scenarios are then developed based on the values shown in Table 2.
The method of the present invention is then used to calculate profit and loss sheets for each of the three uptake scenarios (shown in Table 2), wherein the resulting profit and loss sheets are shown in Table 3.
Use of the most likely uptake scenario (novel marker, established platform: start value=0.1, slope=0.024, ceiling=0.6) suggests use of the test could generate an additional $168 M revenue over 5 years. Analysis of variation of NPV with slope (assuming a constant start value and ceiling of 0.1 and 0.6, respectively) indicates an uptake slope of 0.01 will be required in order for a positive NPV to be achieved.
Software, web, and computer readable data storage implementations of the various embodiments and method steps described herein can be accomplished with methods known in the art including programming techniques with rule based logic and other logic.
Alternatives and modifications may be made to the above without departing from the scope of the invention.
