I've discussed some of the scientific and policy challenges that surround personalized medicine, but I've left for last a much harder task: defining the ethical issues. From my perspective, policy and science problems share at least one important common feature - even if no one can agree on the optimal solution, it's possible to propose a solution, consider it's appropriateness in a reasonably objective fashion (hopefully using some pre-determined metric), and make adjustments based on its performance. Ethical issues are a bit trickier. I would guess that most people are willing to agree on what (most) the issues are, but that's about as far as things go. Trying to decide on appropriate solutions is much harder, because it's essentially a balancing act, and success looks less like an objective criteria and more like alienating the fewest people possible.
Becuase these issues are so much more difficult, I'm going to limit myself to three different classes of issues:
Protecting patient privacy is one of the most important things that must be done before ordinary people will be willing to take advantage of individualized medical care, and just about everyone agrees that patient's have a right to keep details about their health private from most people (even if not from, say, their insurance company or in some cases state or local governments). But how far does that right extend? Does it cover a person's genetic makeup? That is something that undeniably influences health, and a fair amount of information about what diseases a person has or is at risk for can be extracted from genotype and gene expression information like what would be collected for personalized medicine services. How do you keep that information private and what uses are OK? For example, if a person has this information collected for use in risk profiling or diagnosis, should that then automatically commit them to allowing their data to be used for diagnosing and profiling others? While this can be done without identifiers, this information is, in effect, personally identifying, so it can never be truly anonymous. Additionally, what about the privacy of other family members? Families share genetic information, and by knowing something about their risk, a person also learns about their relatives' risks.
One of the issues of privacy is also directly related to patient autonomy - the right of a patient to choose what happens to them. The question of what uses of a patient's data are permissible is not exclusively a question of privacy but also one of autonomy. Is it OK to require a person to allow their data to be used for risk profiling or diagnosis as a condition of performing the service for them? The reasons for doing this are compelling (because large sample sizes are necessary for high quality risk prediction), but is it undue pressure on person to require it? That is to say, if agreeing to this is a condition, can that agreement really be considered free and consensual? What if the condition was not to allow the data to be used for diagnostic work but for research? Should a company be allowed to offer to collect this data for free in exchange for allowing its use in pharmaceutical development or trials? It's a long established concept in research that the incentives offered for participating in trials and research studies should not be so high as to . High-throughput data collection will likely be fairly expensive even for those with insurance (and probably nearly unattainable for those without insurance), what level of incentive is overly tempting? Another privacy issue is also an autonomy issue: if a relative learning something about their risk of a disease says something about your risk of a disease, should you have to consent before that relative can have those tests performed? How can we deal with this information being effectively forced on people?
Cost, just like with the policy issues last time, is a significant ethical issue as well. Something like 46 million people are without health insurance today, and many more have insurance plans that cover only the most basic things. How can we provide access to personalized medicine to everyone? Is access for everyone a reasonable goal? Is it an attainable one? These are some of the toughest questions that have to be answered, but luckily they're very similar to questions that have already been asked about traditional health care. There will likely be a number of levels of personalized care, and if universal access is a goal, what level is acceptable? Is a tiered system where those who can't afford to pay on their own receive some level of service while those who can afford it get more comprehensive care ethically sound? It's something that looks much like our current health care system, but many people are unhappy with that model.
Although the ethical issues facing personalized medicine are tricky, they're actually not that much different than the ones that face traditional medicine. This means that the research and dialogue about them can be transferred, so the conversation doesn't have to start from scratch. It's back to science next time, with a more in depth look at pharmacogenomics.
As with any significant undertaking, the challenges facing personalized medicine are not limited to the science behind it. A large number of public policy challenges exist that must be addressed before personalized medicine can become a reality. Each of these challenges must be dealt with not by a single person or group, but by all of the stakeholders that are affected by it. Who are the stakeholders? That seems like an easier question than it actually is, but in general, the stakeholders are physicians, health care organizations like hospitals and health networks, private insurance providers, public insurance providers such as medicare and medicaid, pharmaceutical companies, state governments, the federal government, and, of course, patients. Not all of these are affected by each issue, but solutions will only be possible when the affected stakeholders work together.
As with the scientific issues, in no way is my listing complete, nor is the discussion about the problems. Rather, I want to give a sense for how broad the policy issues are and who they affect. The main issues I want to describe are
Risk prediction is an important part of personalized medicine, but it requires a fairly large sample of individuals on which to base the prediction. Moreover, the individuals used should be fairly similar to the patient (or patients) whose disease risk we want to estimate. One of the most promising ways of doing this is by using data collected by hospitals. This is called a "secondary use" of health data since it is not being used for its original purpose - making a diagnosis about the patient from whom it was collected. There are other potential secondary uses also, including using this data for pharmacogenomic drug development - another important goal of personalized medicine. These secondary uses are not currently possible, and policies must be developed to allow this data sharing in such a way that protects a patient's confidentiality and their right to opt-out of these uses.
Genetic information has the potential to reveal a large amount about a person's risk of developing a disease in the future. Insurance companies are likely to want to use this information to set policy rates and limits or even potentially to deny coverage altogether. Even if insurance companies have no immediate plans to do this, fear of it may prevent people from having this information collected. One policy solution has already been proposed, a legislative initiative called the "Genetic Information Nondiscrimination Act". The US Senate passed this act in both 2003 and 2005, but it was never brought out of subcommittee in the House of Representatives. The bill has been re-proposed in both the House and Senate this year, but its future is still unclear. Legislation of this type is extremely important, and must be undertaken at a national level. Hopefully mounting public pressure will force the House to consider and pass a bill to prevent this discrimination.
As of now, even if the scientific issues I've discussed before were solved, many physicians would not understand how to use this information. Doctors who don't specialize in genetic conditions are exposed to genetics mainly during medical school, and often only at a very basic level. Although many understand the impact of genes on simple Mendelian diseases, they are unfamiliar with the role genetics plays in complex disease and the way that genetic information can assist diagnosing and treating patients. Increased education, both during medical school and through continuing education programs, is the answer to this problem, but fitting that into an already overcrowded curriculum is no small task. As insurance companies, hospital systems, and patients realize the benefits of personalized medicine, though, doctors will catch up out of necessity. Being proactive with policy solutions should make that process both quicker and easier.
Physicians aren't alone in not understanding genetics. Patients have an even more limited knowledge of genetics, with much of their information coming from popular media. And let's be honest, if all a person knows about genetics they learned from watching Heroes or CSI, then making informed decisions about how to use genetics in medical care is going to be difficult. They will have unrealistic expectations about what can be done to correct genetic problems ("Isn't there some sort of gene therapy for this?") and exaggerated fears about what might be done with this information ("I don't want my DNA in some database where the government can get it!"). Bringing expectations in line with reality and talking about the reasonable risks of gathering this information then must be accomplished through public education policies. The exact forms of these campaigns could vary, but drawing on the successes of cancer awareness and safer sex awareness programs should inform the process.
Bringing personalized medicine to the public is a large computational challenge. Databases must be created to store the information, computing power to actually perform risk prediction calculations must be purchased, deployed, and maintained, and access to high-throughput technologies (genome-wide genotyping, transcriptomic profiling, proteomic profiling, and metabolomic profiling) must be increased. What is the best way to accomplish this? Are private doctor's offices ever going to have the resources to build and maintain a cluster of servers? And will their patient population be large enough to use for risk prediction? High-throughput screening for now is limited to research hospitals, but just as most doctor's offices use external labs to perform clinical tests, external companies will likely fill this role for them.
Of course, none of these problems will be solved cheaply. Educating physicians and the public require some cost, but are not terribly onerous, and the government and professional groups are likely to step into these tasks, because they have some expertise in providing this education. The infrastructure, on the other hand, is far more expensive. Some of the cost will likely be covered by private industry as companies spring up to do high-thoughput screening or provide application hosting for doctor's offices, but that cost will then be passed to doctors, insurance companies, and inevitably to patients. Policy solutions must be created to ensure that health care consumers (i.e., patients) don't get stuck with a disproportionate amount of the bill, and that these infrastructure concerns aren't used to hike insurance rates unjustifiably.
So there is not shortage of policy issues that have to be addressed to make personalized medicine a reality, but some solutions have already been proposed. In some cases, the problems are similar to other problems that have been extensively studied. By taking advantage of these, we can make sure that these challenges don't prove to be long-term roadblocks. Next up I'll talk about some of the ethical issues in personalized medicine. I have my very first dissertation committee meeting on Thursday, so I may be a day late, but I'll do my best to get the post finished on time.
The promise of personalized medicine is one that is fundamentally rooted in science. It's based, at least partly, on the belief that drives all science: knowing more (relevant) information about a process can lead to a deeper understanding of how that process works. Much science, however, (and particularly molecular biology) has followed a fundamentally reductionist paradigm. Each part of a system is studied in isolation, and the information it provides is considered additive to the information provided by a separate piece of the system.
But R.B. Laughlin and David Pines write
So the triumph of the reductionism of the Greeks is a pyrrhic victory: We have succeeded in reducing all of ordinary physical behavior to a simple, correct Theory of Everything only to discover that it has revealed exactly nothing about many things of great importance.1Laughlin & Pines are talking about the Theory of Everything in physics, but the principle holds. Human biology is inordinately complex, with variables working not in isolation but in concert. To individualize medical care requires a deeper understanding of that biology, and that is no small order. I'm going to cover three major problems in this post, which is by no means an exhaustive list of the scientific issues facing personalized medicine. Rather, its a subset of issues that are very interesting to me and present significant hurdles to the field. These are
The first problem is how single nucleotide polymorphisms (SNPs) affect a person's risk of disease or how they influence a complex phenotype. Some of these polymorphisms have an obvious impact. They change either an amino acid or an exon splice site, leading to a different protein product. Others might affect a promoter binding sites or have no discernible change whatsoever. Genetic association studies, the typical tool for identifying SNPs with an impact of a disease or quantitative phenotype, have run against a roadblock in recent years; the findings of many promising studies have not been replicated when the studies were repeated in independent samples. The reasons for this lack of reproducibility are numerous, ranging from spurious initial findings to poor choices for replication samples. Although a number of potential solutions, from cross-validation to requiring all association studies to include an independent validation sample, have been proposed, the field still lacks a consistent, realistic, and tractable method for assessing which associations are "real" (i.e., have real predictive power or mechanistic insight and are worth devoting further resources to) and which are either artifacts of the data or are exclusive to the population being studied.
The next problem is predicting a person's risk of developing a disease. Biology and disease etiology are very complex, and attempting to create a mathematical model of who will develop a disease is impractical. At the same time, being able to estimate a person's risk of a disease is an important goal; it allows people who will benefit from early interventions to be identified, and it can both improve people's health and lower the long-term cost of a person's care. Right now, the only way to predict a person's risk of a disease is to have a medium to large study population that measures the outcome of interest and develop a prediction scheme from that. It's laborious, and if there isn't a study that looks at your outcome, there's no way to create one. Worse, it's difficult to know how well a risk prediction scheme will work when applied to a new patient, especially if that patient is very different from the people used to construct the scheme. These systems are built are imperative to personalized medicine (and, not coincidentally, the subject of my dissertation research, but that's a topic for another time - we're laying out problems, not solutions today), and ways to make them more accurate and more broadly applicable must be found.
Last is a group of problems that fall into their own own category: pharmacogenomics. Creating drugs that have fewer adverse side effects and are more effective has long been a goal of researchers and pharmaceutical companies. The genomics revolution means that more information than ever before can be integrated into this process. But to design more effective drugs requires two very difficult things: a deeper understanding of the mechanisms that underlie different disease subtypes and a more thorough understanding of how drugs act. As an exercise, flip through a drug reference book like the Physician's Desk Reference. The precise mechanism of many (if not most) recent blockbuster drugs is unknown. Doctors understand that some drugs work better for some people than for others, but there is no systematic way to tell who will respond. Targeting drugs to specific disease subtypes should help alleviate that problem. But as an extra layer of consternation, there is significant variability not only in how people respond to a drug, but in what dose they need and what side effects they suffer. This is because a number of factors influence drug metabolism. Polymorphisms in the genes CYP2C19 and CY2D6 are known to affect the rates of metabolisms of many drugs. By modifying the effective concentration of medications, these polymorphisms can either decrease the drug's effectiveness or increase the risk of toxic side effects. Despite the fact that the CYP subunits responsible for processing many drugs are known and the effect of polymorphisms in these subunits is fairly well studied, it is still not typical clinical practice to base drug dosage on this information. Hopefully increased accuracy will change this.
This list of problems just begins to scratch the surface, but it should provide a good sense of the scale of the scientific issues that face personalized medicine. None of the problems are easy, and none of them will be solved overnight. It's not even clear right now what the best solution to some of these problems looks like. In fact, the answer to some of these problems depends on the answer to some of the policy and ethical issues facing personalized medicine. On Monday I'll begin to describe what those policy issues are.
After a long time of thinking about it (and a long time spent procrastinating), I've decided to resurrect the blog. So I've slapped on a new coat of paint, added a few new gew-gaws, and I'm off to the races.
The blog is now actually an experiment of sorts. I'm waist-deep in writing my thesis, which is a risk prediction system that is able to sit at the heart of a personalized medicine system. It's fascinating work and I'm learning incredible amounts both about the mechanism of making a prediction and about what extra steps are necessary to make an algorithm clinically relevant and doctor friendly.
But the prediction side isn't the only part of personalized medicine that's fascinating. It has the potential to drastically change the way medical care is delivered, but there are a number of hurdles to clear first. Some of those are technological, and some of those are at the heart of my thesis project. But other hurdles are policy issues, economic issues, ethical issues, and privacy issues. I have no way of addressing them (other than VERY briefly) in my thesis, both because they're too far afield, and because I have little expertise in those areas.
I like to think broadly about things, however, and I think that this topic is too important to be considered in isolation. This is something that needs to be discussed publicly, both to allay people's fears and also to take advantage of their wisdom. So that's what I'll be trying to do. I'm going to begin by defining some terms and then we'll see where that takes me.
As I said, this is an experiment, so this will be a little different than a typical blog. The posts will tend to be longer and somewhat less frequent. For now the plan is to publish on Mondays and Thursday, but that may change in the future. For me this blog is going to serve two purposes: 1) to keeping up writing while I'm actually doing the work of the thesis (I've just spent a month writing the first two chapters of my thesis, and I know how much constant practice improves writing, so I want to give myself that practice), and 2) to serve as a testing grounds for ideas to put into a publication on the policy of personalized medicine. I'd love to hear input and comments, so please send me an email (my address is in the sidebar).
Reagan Kelly is a PhD student at University of Michigan studying bioinformatics. His thesis is focused on risk prediction algorithms for personalized medicine systems, and he is also interested in the policy and societal implications of individualized healthcare.You can read his CV for more information about him. If you would like to contact him, please send an email to reagank -at- reagank.com