City of Paris

“If you build it, will they pay?” So asked George Poste, DVM, PhD, at the 5th Annual Burrill Personalized Medicine Meeting, and the panel of venture capitalists offered good insights. While all agreed about the potential of personalized medicine, they differed in their definition of the field and how they assess the risks and opportunities. Risa Stack of KPCB suggests starting with the key clinical question: What is the physician’s information need for a given therapeutic category? Is that information available through diagnostics? How important is that information to the physician and the patient, and therefore what is its economic value? Brent Ahrens of Canaan Partners agreed that how a diagnostic affects therapy is the key to value creation, but that the acuity of the problem being addressed (e.g., severity, public health impact) is an added consideration. Stack suggested that the cost of therapy is a key part of the equation, in that the more expensive the therapy the more leverage from diagnostic information. Sue Siegel noted that her firm, Mohr Davidow Ventures, takes a broader view of personalized health (vs. personalized medicine) in the investment portfolio. As the panel discussed the importance of IP position, she stressed the importance of execution and considerations of reimbursement and how a diagnostic impacts workflow in the clinical context. Michael Shuster of Fenwick & West also felt that the importance of IP depends on the company.

The discussion about regulatory risk highlighted interesting contrasting views. Stack believes that, across the board, regulatory approval will be tougher, since the FDA appears to be taking a more conservative stand. She also noted that the current rules that apply to single or few genes and specific analytes do not translate easily to multivariate analyses. She echoed Margaret Hamburg’s comments about the need for educating regulators, and mentioned the efforts by the 21st Century Coalition on Medicine to work with policy makers to balance regulation with innovation. Siegel and Shuster, on the other hand, both felt that regulatory risks are there, but not higher than in other areas; Siegel suggested that there are ways to mitigate those risks.

In closing comments, Ahrens suggested that entrepreneurs reach out to the venture community and explore multiple avenues for funding as a key strategy for success. Siegel urged companies to demonstrate their understanding of risks. Stack suggested that sound knowledge of the targeted venture capital firm’s portfolio and strategies is the price of entry. Steve Burrill commented that most entrepreneurs are poorly prepared to discuss valuation in a meaningful way, particularly within the framework of perceived value. He suggested that funding is available—much more so than a year ago—but valuation is at a low today. In fact, in his opening remarks, he suggested that entrepreneurs put the concept of “dilution” out of their minds and focus on securing the needed capital.

George Poste, DVM, PhD, Arizona State University, and Leroy Hood, MD, PhD, Institute for Systems Biology, both gave scintillating talks at the 5th Annual Burrill Personalized Medicine Meeting last week. Poste offered a big-picture view of healthcare and the sociopolitical, regulatory and business environment within which innovation must thrive and advance. He spoke of the need for “harmonizing incentives for diverse constituencies” and the inefficiencies and waste associated with the current system of siloed care and procedure-based reimbursement, and pointed to the estimated annual cost of $177 billion attributed to patient noncompliance. Hood, on the other hand, made an eloquent overview of P4 (predictive, personalized, preventive and participatory) medicine as the inevitable wave of the future. He described the elegant science and the diagnostic and prognostic power of organ-specific blood fingerprinting, and boldly predicted the availability, within the next five to ten years, of $1,000 genome sequencing, routine screening of 2,500 blood proteins as part of wellness assessment and the analysis of 1,000 transcriptomes on a single run for disease stratification.

But if there was a common thread through these two talks, it was the enormity of the informatics challenge at hand. Poste used biomarkers as an example of the current “evidence dilemma.” There are more than 300,000 biomarkers in the literature today, many of which are poorly characterized and based on small-cohort studies. There are no standardized assays and little statistical robustness. Poste envisions a strategic triangle and a Dx-Tx-iHx (diagnostics-therapeutics-informatics) technology convergence, and the rise of open-source networks and consortia as scientists learn to share results and collaborate to address what he calls the omics data tidal wave. On a patient care level, Poste lamented the inefficient use of information today and envisions a new era of connectivity, where informatics is an integral part of clinical decision support.

Hood spoke of the digitalization of biology and biology as an information science. He framed the problem of too much data as one of signal to noise, where noise is winning today. He illustrated how probability models can be used to make dramatic improvements using the recently completed genome sequencing of a family of four. Using Mendelian genetics, the research team was able to reduce errors tenfold and narrow down the disease gene candidates more quickly. Hood is obviously a proponent of sequencing genes of families, where Mendelian genetics can be applied as part of the strategy to streamline analysis and reduce errors. Like Poste, Hood noted that a sound strategy for mining patient data and overcoming the technological and legal-social barriers for doing so is a high priority. And he echoed Poste’s call for standardization, especially in areas such as proteomics.

The illustrious panel of scientists, physicians, regulators and businesspeople at the 5th Annual Burrill Personalized Medicine Meeting last week once again reaffirmed the inevitable trajectory toward personalized medicine and, interestingly, that some of the forces propelling it are also the roadblocks that must be overcome. This paradox came across most clearly for me in the discussions about physician adoption, reimbursement and regulatory perspectives, the latter by FDA Commissioner Margaret Hamburg, MD.

Over and over again, panelists spoke of the flaws of the current reward system for healthcare professionals and the environment within which they operate. The system rewards doctors for the number of patients they see and the procedures they perform, leaving them little time to learn about personalized medicine or even to think about it. Education is key, starting with the incorporation of genomics in medical school curricula. Communicating to physicians in the language they understand (clinical validation) is also critical. And, Geoffrey S. Ginsburg, MD, PhD, from Duke University and Joseph D. McInerney of the National Coalition of Health Professional Education in Genetics pointed out that we should be thinking beyond physicians and include other healthcare professionals (e.g., nurses, genetic counselors) as we consider the role of the provider in the adoption of personalized medicine. Capturing comprehensive family history, which often contains a wealth of data for risk assessment and treatment decisions, has become a luxury in the time-constrained patient care environment and one that is lacking in systematic, standardized data capture capabilities.

On reimbursement, everyone agrees that the current system and CPT code system discourage innovation. And the fact that in vitro diagnostics (IVD) accounts for only 0.01 percent of healthcare expenditures suggests that there is a long way to go and that the path to personalized medicine will be paved by good strategy supported by abundant resources as illustrated by Genomic Health’s trail-blazing reimbursement strategy. Eric L. Book, MD, of Tethys Bioscience commented that “we are distracted by how much healthcare is costing right now” to the extent that sometimes pressure to improve the bottom line detracts from thinking about “doing the right thing for society.” A sobering thought.

The earnest and disarming Dr. Hamburg vowed to engage with industry and academia to advance the scientific rigor of regulatory decisions—a broad advancement of regulatory science. She spoke of a more holistic, multidisciplinary approach (i.e., more integration of diagnostics and therapeutics in review functions and processes), and more flexibility, more transparency and more clarity. Most of the audience concluded, however, that the regulatory path would get tougher, not smoother, going forward. Last but not least, although Hamburg did not provide a release date for the new IVD/MIA regulations, she did disclose in private conversations that they could be expected before the end of the year.

On April 16, The New York Times published an article, “Genes Show Limited Value in Predicting Diseases,” based on commentaries and a review article on genetic risk prediction in the April 23 issue of The New England Journal of Medicine. The New York Times quoted one of the authors, David B. Goldstein, as saying, “With only a few exceptions, what the genomics companies [that offer personal genomic information] are doing right now is recreational genomics.” (I inserted the clarification to emphasize that Goldstein was not talking about genomics companies in general—only those that are offering personal genomic information to physicians and patients.) Curious, I downloaded the original papers.

Goldstein’s article, “Common Genetic Variation and Human Traits,” asserts that genomewide association studies are not an efficient way to identify gene variants capable of predicting risk in any meaningful way, and that the more gene variants associated with a given disease, the more time- and resource-intensive it would be to identify a manageable and usable set of risk predictor genes. The article also offers the interesting perspective that the “apparently modest effect of common variation on most human diseases and related traits probably reflects the efficiency of natural selection in prohibiting increases in disease-associated variables in the population.” And Goldstein points to the promise of looking at polymorphisms associated with responses to drugs or infectious agents. (Being a marketer, I am always looking for the silver lining.)

In “Genomewide Association Studies—Illuminating Biologic Pathways,” Joel N. Hirschhorn offers a different perspective—that the main goal of genomewide association studies is not to figure out how to predict individual risk of diseases and traits, but to elucidate underlying biologic pathways. As an example, he notes that of the 23 loci found to be associated with lipid levels in genomewide studies, 11 point to genes that have previously been shown to encode key proteins involved in lipid metabolism. Hirschhorn goes on to enumerate a number of genes encoding sites of action of drugs approved by the FDA that have been similarly identified. Citing the distance between the deciphering of the chemical composition of cholesterol and the development of statins, which were separated by “nearly a century and three Nobel Prizes,” he cautions that “each discovery of a biologically relevant locus is a potential first step in a translational journey” and that “some journeys will be shorter than others.”

Peter Kraft and David J. Hunter raise the question, “Genetic Risk Prediction—Are We There Yet?” and concur that there are few monogenic diseases. For example, Crohn’s disease is associated with more than 30 loci and thus potentially more than 30 genes. The graph on page 1702 of the article shows that the higher the number of undiscovered genetic factors for the disease, the lower the likelihood of assessing the risk level for the disease in any meaningful way. Kraft and Hunter outline factors determining the clinical value of a genetic test, and point out the importance of looking not only at test performance (sensitivity, specificity, positive and negative predictive values) but also at costs and benefits of interventions, as well as outcome data, in determining the clinical value of a genetic test. Most important, the authors suggest that while testing for genetic risk may not be here today, it may only be two or three years away, and advocate the urgent need to develop guidelines for physicians, which are now lacking.

John Hardy and Andrew Singleton’s review article, “Genomewide Association Studies and Human Disease,” is an excellent backgrounder for anyone with any interest in this field. It details the stages of a genomewide association study, summarizes its benefits, misconceptions and limitations and offers a helpful glossary. Importantly, it includes an elegant explanation of genetic variability in gene expression, which adds yet another dimension to the understanding of genetic risk of diseases—“moving from dichotomous to graded genetic risk” as the authors subtitled one of the paragraphs. What I found very interesting, also, is the notion that the interaction between genome and environment, which many people like me believed was a given, has not been demonstrated for the most part. And, like the other reviewers, Hardy and Singleton are largely optimistic about the future and the scientific community’s ability to solve the complex puzzle of genes and diseases.

Francis Collins, former director of the National Human Genome Research Institute, provided a cogent, multifaceted overview at the Achieving ROI in Personalized Medicine: Barriers, Incentives, and Pathways to Successful Commercialization conference hosted by the Personalized Medicine Coalition. He spoke about Moore’s Law, as some other speakers at the conference did, comparing the trajectory of scientific advances in genomics favorably to the exponential increase in computing power that was the basis of Gordon Moore’s original thesis. “Beating them hands down,” Eric Lander was quoted as saying. Collins cited numerous major projects (e.g., the Human Microbiome Project) and reminded the audience that rapid advances are a given and must be taken into account in any development models. He spoke enthusiastically of the need for rigorous data about gene-environment interaction and thus a prospective large-scale study (dubbed, aptly I thought, “Framingham on steroids” by someone in the audience). Collins also proposed some urgent action items, such as the need for more adequate oversight of genetic testing and the help needed by healthcare providers. He suggested that healthcare providers need education on personalized medicine, later supported by Janet Woodcock, FDA, who commented, “Clinicians dislike information without instructions.” Collins also noted that today’s primary care physicians are not adequately compensated for face time with patients.

On the subject of reimbursement, James Utley, Coventry Health Care, took the audience through the payer’s decision-making process: the determination of medical necessity (essentially a benefits vs. risk assessment) and demonstrated value, which relies on clinical evidence and, in his words, is “unbiased to cost.” The ROI calculation and coverage decision that follows then becomes an actuarial issue, which can affect changes in premiums as well as benefits. Utley referred or deferred to this actuarial decision several times in his discussion. While I thought his presentation shed light on how clinical evidence is reviewed and the level of information required, I was disappointed that it offered no insight on how coverage is determined. Later in the discussion, Utley did make the bold and sobering statement that somehow the nation needs to come to grips with the fact that the pot for healthcare is not unlimited. Indeed, several speakers suggested that the incremental funding needed to advance some of the exciting developments would have to come from reducing wasteful spending.

Which brings me to the subject of healthcare spending and future directions in healthcare delivery. The Hon. W.J. Billy Tauzin of Pharmaceutical Research and Manufacturing of America (PhRMA) gave an impassioned talk about his personal experience with cancer and the very personal treatment decisions he and his doctor had to make. Most important, he made a strong and convincing case for the transition from prevention to intervention, from “sick care” to healthcare, especially in chronic diseases (e.g., obesity, diabetes) that are reaching epidemic proportions.

Change is coming.

Last week, a star-studded panel gathered at the Achieving ROI in Personalized Medicine: Barriers, Incentives, and Pathways to Successful Commercialization conference hosted by the Personalized Medicine Coalition to review the results of the newly completed Deloitte study on ROI in personalized medicine and to offer their perspectives. It was the week after the inauguration. The excitement about personalized medicine was amplified by the optimism around the new administration, and justifiably so. After all, our new president happens to have been the sponsor of the Genomics and Personalized Medicine Act of 2007. At the same time, the economy, rising healthcare costs and the state of the pharmaceutical industry weighed heavily on everyone’s mind. And what resulted was a day of dialogue that was both expansive and sanguine, and a balance of vision, analysis and introspection that I found refreshing.

The fact that the consumer came out as an undisputed winner in the ROI study, led by Paul Keckley, Deloitte Center for Health Solutions, reaffirmed the role of personalized medicine in the future of healthcare. But a host of scientific, business, regulatory, reimbursement and societal barriers challenge the realization of its potential. It is quite clear that better integration between therapeutics (Rx) and diagnostics (Dx) from the start would be desirable in streamlining the development process, but a winning business model for Rx/Dx collaboration remains to be crafted. It was telling when Aidan C. Power of Pfizer pointedly stated that the drug developer would want to keep its options open to work with different diagnostic partners to develop the second-generation biomarker test after launch. Not surprisingly, Randy Scott of Genomic Health commented that the current cost-based reimbursement system for diagnostics is outdated, and he foresees a rebalancing of healthcare spending between therapeutics and diagnostics. This view was supported by a comment later in the day by Janet Woodcock of the FDA, in the context of discussing the need for more data to support science-based medicine, that current reimbursement for diagnostic tests generally is not sufficient to justify adequate study.

Woodcock’s comprehensive review touched not only on the regulatory gaps such as the In Vitro Diagnostic Multivariate Index Assays (IVDMIA) guidance that remains to be taken beyond the draft, and the yet-to-be-published draft of Rx/Dx co-development guidance, but also complex issues around valuation of specific therapies, government incentives for the development of personalized medicine, and the role of academia and federal research funding. Both Woodcock and Francis Collins, former director of the National Human Genome Research Institute, advocated a strong partnership between academia and the private sector. But Woodcock wondered if the reward system within academia, and the scientific community’s innate respect for pure science, would penalize those who engage in translational research.

Next week: The discussion around science, reimbursement and the future of healthcare.