Personal genetics is moving faster than ever, we began with a few variants in an affordable genetic test, now we have hundreds of thousands, soon it will be sequencing of all our genes (exome sequencing) and finally, the whole genome by 2011, 2012, 2013? Who’s prepared to predict?
What’s the use of it though – the technology is moving much more rapidly than the knowledge of the effects of genetic variation and we still only really have a few “clinically useful” applications? A problem that has not yet been fixed is the oversight of commercial services, the fact that there are not yet any new specific regulations to deal with personal genetics is probably just as well because by the time any law is passed it would be likely to be out of date. The lack of really effective clinical utility and the existence of commercial interests increases the confusion though. It’s hard to sell something that is interesting, “fun”(?), quite expensive, but not actually that useful to the majority right now. Hard to sell means sometimes over the top marketing. It’s also hard for the media to know what to do with it so we end up with confusion borne of ignorance. Many are getting worried that it’s being oversold, others worry about over-regulation. I think we have to be patient, it will become irresistibly useful sometime. The clinical applications will be clear as will the non-clinical uses. Maybe it’s like the early days of the motor car – useless (I can walk faster than that thing and when I need real transport I’ll use my horse – for transport read medicine and for horse read family history) and potentially dangerous but also interesting and “fun” (and expensive).
For now we have what we have, it’s not going away and we need to make the best of it. Since there has just been some twitter discussion about utility and since we wrote a review recently that’s about to be submitted I thought I’d post some maybe relevant excerpts here.
What to Regulate?
The types of services offered include health risk assessments, nutrigenetics, ancestry, parentage, pharmacogenetics and neonatal genotyping – the last category is particularly sensitive especially as services based on infant growth, IQ, dietary requirements etc are starting to appear. These offerings will develop whether or not the evidence exists to support the services (see for example My Child Talent Profile) and the presentations of these highly questionable services will become ever more sophisticated and convincing.
Regulations or strong guidelines are required for at least two purposes, both with the same aim of reducing the possibility of harm while trying not to be overly cautious and exclude possible benefits for the consumer. They are needed to protect the delivery and growth of serious beneficial personal genetics services, allow the existence of harmless even if not necessary beneficial services and to reduce the influence of questionable, potentially harmful services. Regulations serve to protect both the consumer from harm and the reputable service providers from disreputable competition, and they also need to protect from harm by not prohibiting potentially useful services.
There are several areas that are potentially subject to regulation:
1. Analytical validity – this covers the laboratory process and is a measure of the accuracy of the genotyping. It is relatively straightforward and each country has its own laboratory accreditation procedures which cover accuracy and reproducibility although they vary from one country to another.
2. Clinical validity – the genetic results are interpreted based on current knowledge. Clinical validity concerns the accuracy of the interpretation, i.e. the accuracy with which a test predicts a clinical outcome. For example a certain set of SNPs might be predicted to influence cholesterol levels under given dietary circumstances. Clinical validity is less clear cut but generally straightforward for many genetic risks and gene-environment interactions however there is not always wide agreement on when genetic association may be considered valid.
3. Clinical utility – the measure of the likelihood that the recommended therapy or intervention will lead to a beneficial outcome. Clinical utility tends to be assessed after the test is offered and the decision is usually made by the payer such as the government or private insurance. For personal genetic testing which is not yet reimbursed by either payer, the decision maker will be the end user (practitioner and/or patient/customer).
Utility is far from clear
Clinical utility is the most controversial aspect; it is often difficult to define and has to take into consideration many factors including positive or negative psychological or motivational effects on the end user. It is generally proposed that clinical utility can only be thoroughly established through randomised clinical trials (RCT) but these are challenging for the personal genetics environment which is rapidly evolving, often includes diet, lifestyle & behavioural changes and has small cumulative effects over decades (see Gulcher & Stefansson and Ransohoff & Khoury for an example of the current debate) and benefit may extend beyond purely clinical utility (Foster, et al). For example an Alzheimer's risk assessment would not seem to have much clinical utility at the moment but that does not mean it has no utility to the individual. A negative result can bring relief where there is a family history of the disease and a positive result can enable preparations for the increased possibility. There does seem to be demand for this sort of knowledge and there is no evidence that it is harmful, on the contrary (caveat: providing of course that there is full explanation of the consequences of the test before it is given) see Green et al and Vernarelli et al.
A further problem is what is the definition of a clinical benefit? A gene may not be associated directly with disease risk, such as myocardial infarction, but to intermediate phenotypes, e.g. lipid levels, hypertension, homocysteine, etc, which are independent risk factors for disease. Some commentators require that clinical utility is demonstrated as a reduction in disease incidence, the majority view though accepts that lowering of intermediate risk factors is acceptable (as is the case for phytosterols and their cholesterol lowering properties). However there are those who insist that clinical utility can only be proven through suitable randomised clinical trials (RCTS) even if that takes years (Ransohoff & Khoury).
It is also the most difficult category to regulate, if at all possible, and a common point of view is that the regulations should cover the analytical validity, the truth of the material used to disseminate information about the test (e.g. advertising & marketing) and the personal information itself at least in the short term.
Lifestyle vs. Medical
Adding to the difficulty of reaching clear regulations (or consensus) is the definition of a test – what is a “lifestyle” test and what is “medical”. A nutrigenetic test for example, is often described as a lifestyle test and the results are not used to give individuals a “risk assessment” for various diseases. However it is argued that you cannot talk about genes, health and lifestyle without there being a medical component and the fact that individuals are given genetic results for which disease risk information is easily available on the internet means that it should be classified as “medical”. The tests offered by 23andMe and DeCodeMe would appear to be clearly “medical” as they describe disease risk however both companies argue that they are providing information for education purposes only and not for the purposes of making health decisions. It is a very grey area but there is a good argument that any overtly medical risk assessment being sold DTC should at least require some sort of pre-market review. Note I am not referring to the actual gentoyping on which there should not be restrictions. It is the information itself, the interpretations, to which a pre-market review should apply. Who does the review is another thing entirely but if the industry took a lead from journal peer-review, while not perfect, it could be a good beginning.
In the meantime – follow the code…
What’s the use of it though – the technology is moving much more rapidly than the knowledge of the effects of genetic variation and we still only really have a few “clinically useful” applications? A problem that has not yet been fixed is the oversight of commercial services, the fact that there are not yet any new specific regulations to deal with personal genetics is probably just as well because by the time any law is passed it would be likely to be out of date. The lack of really effective clinical utility and the existence of commercial interests increases the confusion though. It’s hard to sell something that is interesting, “fun”(?), quite expensive, but not actually that useful to the majority right now. Hard to sell means sometimes over the top marketing. It’s also hard for the media to know what to do with it so we end up with confusion borne of ignorance. Many are getting worried that it’s being oversold, others worry about over-regulation. I think we have to be patient, it will become irresistibly useful sometime. The clinical applications will be clear as will the non-clinical uses. Maybe it’s like the early days of the motor car – useless (I can walk faster than that thing and when I need real transport I’ll use my horse – for transport read medicine and for horse read family history) and potentially dangerous but also interesting and “fun” (and expensive).
For now we have what we have, it’s not going away and we need to make the best of it. Since there has just been some twitter discussion about utility and since we wrote a review recently that’s about to be submitted I thought I’d post some maybe relevant excerpts here.
What to Regulate?
The types of services offered include health risk assessments, nutrigenetics, ancestry, parentage, pharmacogenetics and neonatal genotyping – the last category is particularly sensitive especially as services based on infant growth, IQ, dietary requirements etc are starting to appear. These offerings will develop whether or not the evidence exists to support the services (see for example My Child Talent Profile) and the presentations of these highly questionable services will become ever more sophisticated and convincing.
Regulations or strong guidelines are required for at least two purposes, both with the same aim of reducing the possibility of harm while trying not to be overly cautious and exclude possible benefits for the consumer. They are needed to protect the delivery and growth of serious beneficial personal genetics services, allow the existence of harmless even if not necessary beneficial services and to reduce the influence of questionable, potentially harmful services. Regulations serve to protect both the consumer from harm and the reputable service providers from disreputable competition, and they also need to protect from harm by not prohibiting potentially useful services.
There are several areas that are potentially subject to regulation:
1. Analytical validity – this covers the laboratory process and is a measure of the accuracy of the genotyping. It is relatively straightforward and each country has its own laboratory accreditation procedures which cover accuracy and reproducibility although they vary from one country to another.
2. Clinical validity – the genetic results are interpreted based on current knowledge. Clinical validity concerns the accuracy of the interpretation, i.e. the accuracy with which a test predicts a clinical outcome. For example a certain set of SNPs might be predicted to influence cholesterol levels under given dietary circumstances. Clinical validity is less clear cut but generally straightforward for many genetic risks and gene-environment interactions however there is not always wide agreement on when genetic association may be considered valid.
3. Clinical utility – the measure of the likelihood that the recommended therapy or intervention will lead to a beneficial outcome. Clinical utility tends to be assessed after the test is offered and the decision is usually made by the payer such as the government or private insurance. For personal genetic testing which is not yet reimbursed by either payer, the decision maker will be the end user (practitioner and/or patient/customer).
Utility is far from clear
Clinical utility is the most controversial aspect; it is often difficult to define and has to take into consideration many factors including positive or negative psychological or motivational effects on the end user. It is generally proposed that clinical utility can only be thoroughly established through randomised clinical trials (RCT) but these are challenging for the personal genetics environment which is rapidly evolving, often includes diet, lifestyle & behavioural changes and has small cumulative effects over decades (see Gulcher & Stefansson and Ransohoff & Khoury for an example of the current debate) and benefit may extend beyond purely clinical utility (Foster, et al). For example an Alzheimer's risk assessment would not seem to have much clinical utility at the moment but that does not mean it has no utility to the individual. A negative result can bring relief where there is a family history of the disease and a positive result can enable preparations for the increased possibility. There does seem to be demand for this sort of knowledge and there is no evidence that it is harmful, on the contrary (caveat: providing of course that there is full explanation of the consequences of the test before it is given) see Green et al and Vernarelli et al.
A further problem is what is the definition of a clinical benefit? A gene may not be associated directly with disease risk, such as myocardial infarction, but to intermediate phenotypes, e.g. lipid levels, hypertension, homocysteine, etc, which are independent risk factors for disease. Some commentators require that clinical utility is demonstrated as a reduction in disease incidence, the majority view though accepts that lowering of intermediate risk factors is acceptable (as is the case for phytosterols and their cholesterol lowering properties). However there are those who insist that clinical utility can only be proven through suitable randomised clinical trials (RCTS) even if that takes years (Ransohoff & Khoury).
It is also the most difficult category to regulate, if at all possible, and a common point of view is that the regulations should cover the analytical validity, the truth of the material used to disseminate information about the test (e.g. advertising & marketing) and the personal information itself at least in the short term.
Lifestyle vs. Medical
Adding to the difficulty of reaching clear regulations (or consensus) is the definition of a test – what is a “lifestyle” test and what is “medical”. A nutrigenetic test for example, is often described as a lifestyle test and the results are not used to give individuals a “risk assessment” for various diseases. However it is argued that you cannot talk about genes, health and lifestyle without there being a medical component and the fact that individuals are given genetic results for which disease risk information is easily available on the internet means that it should be classified as “medical”. The tests offered by 23andMe and DeCodeMe would appear to be clearly “medical” as they describe disease risk however both companies argue that they are providing information for education purposes only and not for the purposes of making health decisions. It is a very grey area but there is a good argument that any overtly medical risk assessment being sold DTC should at least require some sort of pre-market review. Note I am not referring to the actual gentoyping on which there should not be restrictions. It is the information itself, the interpretations, to which a pre-market review should apply. Who does the review is another thing entirely but if the industry took a lead from journal peer-review, while not perfect, it could be a good beginning.
In the meantime – follow the code…
Nice article Keith. Assessment should be done with all care and seriousness. Even one percent of unseen consequences can put the life of the patient in utmost danger. Doctors and all medical staffs should do risk assessments with all their hearts.
ReplyDelete