The high cost of adverse events has been estimated to be at least US$177 billion per year in the U.S., and prescription drugs overall have efficacy only approximately 50% of the time, representing a potential waste of approximately $350 billion of the worldwide $700 billion or more drug spending (Miller et al., 2011).
The wide range of diseases and conditions targeted by a new generation of diagnostic tests (NDx) offers hope for improved targeting of treatments to the patients who can most benefit. These tests include a wide array of biomarkers and genomic tests which have the ability to detect the risk, presence and progression of diseases, and in the case of “companion diagnostics,” enable optimal pairings of patients and treatment strategies.
In spite of rapid global growth in the past several years, however, there are some important challenges to guiding the development and use of diagnostics toward measurable improvement in health outcomes and efficiency. Although there are a number of critical issues surrounding the development and use of diagnostics, the three most important issues in the U.S. are: (1) discovery and development, (2) integration into clinical practice, and (3) economic evaluation and payment.
1. Discovery & Development
The development of NDx is complicated by several factors. First, the most fundamental challenge in the development of diagnostics is financial risk and return on investment (ROI). The unit costs of diagnostic tests tend to be considerably lower than the costs of treatment within a therapeutic area, yet the total costs of developing and testing a diagnostic test can be high, as regulatory approval and caregiver acceptance hinges on the ability of the diagnostic to disrupt and improve the status quo—both of which demand high accuracy in the form of sensitivity, specificity, and clinical utility. Large pharmaceutical companies with diversified portfolios have the ability to spread risk, but their incentives to develop tests may be limited to the therapeutic areas in which they have treatments on the market or firmly in the pipeline, as in the case of companion diagnostics. Moreover, not all companion diagnostics have a clear business case; in some therapeutic areas, there may be concern on the part of developers that improved diagnostic accuracy could result in decreased sales revenue insofar as tests are able to identify appropriate and inappropriate candidates for treatment.
Another important limitation to the role of large pharmaceutical companies in NDx development is the problem of incentives. Small diagnostic developers have high-powered incentives to discover and innovate, as the leaders and senior scientists typically have a high proportion of their compensation tied to the success or failure of a small portfolio of products. In the case of an imminent sale of the diagnostics company to a larger entity (e.g., investors; pharmaceutical companies), these compensation incentives can be very high. Conversely, once a small company is merged with a larger one (or the larger company diversifies by setting up a small NDx development department internally), those incentives are unavoidably attenuated. This does not imply that it is impossible for large companies to innovate—Roche, Abbott, and Novartis have been very successful innovators in the NDx space—but the pace of innovation will be likely comparatively slower. Indeed, many of the NDx tests in the portfolios of Roche, Abbott, and Novartis were developed by small NDx companies and either licensed or acquired by the larger entities post-discovery. In sum, larger companies face a classic “make-or-buy” tradeoff, driven by three main factors: (1) uncertainly of success and high variance in ROI; (2) incentive attenuation post-integration or acquisition; and (3) the asset-specific investments (i.e., sunk costs) in the case of companion diagnostics, where the uncertainty in NDx development is compounded by uncertainty surrounding whether the companion treatment is able to gain regulatory approval and market access.
2. Integration into Practice
Another important challenge to the development and diffusion of novel NDx is acceptance and adoption in the provider community. But in the case of diagnostics, acceptance is not straightforward. Medicine is as much art as science, and many physicians continue to rely disproportionately on intuition, traditional forms of diagnosis, and the standard practices of the medical community—all of which are reinforced by practice cultures (which vary by setting, geography, etc.) and legal incentives (i.e., “defensive medicine”).
For a diagnostic to have “value”—clinical or economic—the results of the test must in some way lead to a change in provider behavior in the form of an altered or augmented treatment approach, the ordering of additional tests, or referral to additional services. This may be less of a problem for companion diagnostics, where the bundling of tests and treatments has the potential to expedite adoption and change provider behavior. But other diagnostics are only effective if provider behavior changes in response to test results. With companion diagnostics, the focus is seeking appropriate treatment, whereas with other diagnostics, the focus is in the avoidance of inappropriate or low value care. With the possible exception of “defensive medicine,” the existence of an ordered test does not necessarily compel caregivers to deviate from the status quo.
3. Economic Value & Payment
For NDx, regulatory approval is an important hurdle, as the pace of discovery has allowed little time to optimize the approval process. As providers begin to embrace the new generation of diagnostics, acceptance among payers will be increasingly restricted by questions of value. Economic evaluation of diagnostics is in some case more complicated than the analogous evaluation of treatments, mainly because the economic properties of diagnostics differ depending on whether the core application of the NDx is (1) screening, (2) clinical diagnosis, (3) treatment prognosis, or (4) disease monitoring.
Screening: The most basic value story of an NDx is the ability to detect the presence or likelihood of disease earlier than alternative methods of diagnosis. However, the value of an NDx in disease screening is complex. First, on the cost side of the ledger, population-based screening can increase the total costs of diagnosis rapidly—especially in case of diseases and conditions with relatively high prevalence– and indeed this has been a major concern on the part of payers. But higher total diagnostic costs can be offset by lower total treatment costs. The main connection between screening and treatment costs is whether the costs of earlier treatment are lower than the costs of later treatment. Normally this is the case, as treatment costs typically intensify with disease progression. Thus, the value story hinges on the difference between population-based diagnostic costs—which can be large—and the magnitude of difference between earlier versus later treatment.
Clinical Diagnosis: This is of course the most fundamental role of NDx—to improve clinicians’ ability to diagnose diseases and conditions. This has been the goal of virtually all traditional diagnostic approaches, the basis of which is the face-to-face clinic visit. But NDx offers a much wider range of applications. The ability of NDx to improve the accuracy of diagnosis has the potential to lead to less over-treatment and less under-treatment, and more reliance on appropriate, guideline-based care. Greater accuracy and precision also optimizes the timing of treatment duration, which also has implications for resource use and outcomes.
Treatment Prognosis. Which patients are best suited for a treatment? Improved diagnostic accuracy and precision supports the optimal pairing of patient and treatment strategy, as in the case of companion diagnostics, personalized medicine, and stratified medicine. Pharmaceutical companies are largely abandoning the “blockbuster” mindset, the basis of which was to develop new compounds with the largest possible market share. The main problem with the blockbuster approach is that drugs work differently on different types of patients, resulting in failed clinical trials (the costs of which can be enormous) and, worse, post-market safety and efficacy problems (e.g., in the case of Vioxx). NDx has the potential to (1) reduce treatment adverse events by supporting the optimal pairing of patients and treatments; (2) inform the design of randomized clinical trials by fine-tuning patient grouping and inclusion criteria (which in turn may increase the chances that a treatment will gain regulatory approval); and (3) reduce the uncertainty surrounding the potential value of treatments in the pipeline.
Disease Monitoring. Monitoring refers to the monitoring of treatment effects during treatment and the surveillance of patients during or following treatment. As with screening, monitoring has the potential to save resources if the costs of early treatment are less than the costs of later treatment. The diagnostic and prognostic capability of advanced diagnostics is also of positive economic value if patients with adverse treatment effects or recurrent disease can be stratified into optimal second-line therapeutic regimens.
In sum, although the economic aspects of NDx vary considerably across these four applications, the potential of NDx to “pay for itself” is substantial. The final frontier for NDx is acceptance—and eventual reimbursement—from payers. It is critical that NDx developers continue to fine-tune the economic arguments for their products to overcome the common misconception that added diagnostic tests are simply added costs. In addition to the direct cost savings potential described above, most importantly NDx has the potential to improve outcomes, and there is of course a variety of indirect cost savings associated with improved outcomes. Thus, even a narrow “direct cost” approach is likely to understate the potential value of NDx.
-John E. Schneider, PhD & Cara M. Scheibling
Avalon Health Economics