how clear is the crystal ball?

LONDON The genome landed amid a mass of orchestrated press releases and international media coverage. Some commentators prophesied the dawn of a new and glorious era — others poured forth scorn or doom-laden predictions. Now that a few months have passed, the British Medical Journal has decided it is time for another bout of more careful reflection, and today published a series of papers that together give a less than euphoric view of the future, much less the current utility of predictive genetic testing.Add to this the British House of Commons' recently published report Genetics and Insurance, which roundly condemns the way that the UK's insurance industry has already started to use predictive testing, and the need for careful consideration becomes all too apparent.Predictive genetic testing, says director of cancer genetics services at the University of North Carolina, James Evans, and co-authors (BMJ 2001, 322:1052-1056), represents a new and growing class of medical tests, which differ fundamentally from conventional diagnosis. Whereas conventional tests assess the patient's current condition, predictive genetic tests aim to provide a technological crystal ball, allowing us to glimpse the future.The tests, however, have a "substantial component of uncertainty, not only about whether a specific condition will develop, but also about when it may appear and how severe it will be."Even if the tests are good at predicting the future, there is still a debate over the value of using them if there is no advice or treatment that could be given to enable the person to at least mitigate the effects of the disease.For a few conditions, such as the rare multiple endocrine neoplasia 2, predictive tests seem to be highly effective and presymptomatic intervention is available. This would be a good candidate for predictive testing. Alzheimer's disease, Evans et al maintain, is an example of the other end of the spectrum, where predictive testing has the potential to cause harm. The currently available tests looking at variations in apolipoprotein E genotype "give imprecise measures of risk and could result in anxiety, stigmatisation, or discrimination," they say. "The principle of avoiding harm suggests that currently such testing would generally be unethical because no effective prevention is available."A brief paper by Jaakko Kaprio, professor of epidemiology at the University of Helsinki, Finland, points out that the more we know, the more complex the picture becomes (BMJ 2001, 322:1019). Monogenic diseases had seemed to be the ones with most hope for a valid use of genetic testing. Kaprio is not now so sure. "There is increasing recognition and evidence that even the monogenic diseases are not simple and that the relation between genotype and phenotype is modified by both other genes and environmental effects," he says.To support his argument he comments on the study of familial hypercholesterolaemia (BMJ 2001, 322:1019-1022). The simple story is that if you have a mutated low density lipoprotein receptor gene, you are at increased risk of coronary heart disease. But look at the data more carefully and a different picture emerges. This mutation has only been associated with increased mortality since 1915. Prior to that it had a neutral effect, or was even beneficial, as it seems to offer protection from infectious disease.It is the interaction between this monogenic alteration and the environment, not the mutation per se, that has led to familial hypercholesterolaemia currently being a disease risk. It is an important lesson that needs to be remembered, lest people place too much certainty on data from genetic tests.Misplaced certainty and the creation of a genetic underclass were at the heart of the concerns raised by the House of Commons Science and Technology Committee, when they viewed the recent actions of the UK insurance industry.The committee found that three years ago, the industry had started making use of genetic information relating to predictive tests for seven different conditions. These test results had the possibility of causing some people to pay higher premiums and making others uninsurable.It wasn't just the chance of discrimination, but the use of poorly substantiated tests, that worried the Committee, as it noted that four of the original seven tests have already been abandoned as sources of information for underwriters. "The Association of British Insurers' decision that four of the tests it recommended insurers use three years ago are now no longer relevant or reliable casts the gravest possible doubts on the validity of all the tests not explicitly approved by the GAIC [Genetics And Insurance Committee], being currently used by insurers. Insurers have given the test results a predictive significance that cannot, at present, be justified."The three tests that are still allowed look for the HD gene for Huntington's chorea, the APP, PS1 and PS2 genes indicating early onset familial Alzheimer's Disease and BRCA1 and BRCA2 that predisposes a person to hereditary breast and ovarian cancer.The insurance industry maintains that discrimination has always been a part of their business. Around 5% of the population have always been denied insurance. The potential that genetic tests taken for medical purposes might expand the pool of uninsurable people is, however, a cause for concern and the committee maintained that "The Government and the insurance industry must collaborate to provide an alternative form of insurance for those who would be denied it because of their genetic make-up. Alternative insurance arrangements may help to calm public fears about the risk of genetic discrimination."In a press release following the Committee's report, the ABI said that it is currently reviewing its policy. And the pressure on the industry is increasing. This week UK Health Secretary Alan Milburn said that the government would be willing to outlaw the use of genetic tests by insurance companies if the Human Genetics Commission recommended it.According to Mark Hall, Professor of Law at Wake Forest University, Winston-Salem, NC, the situation in the US is different, but equally complex. The 1996 Federal law entitled the Health Insurance Portability and Accountability Act prohibits insurance companies from using genetic information to refuse to issue coverage, or to limit coverage, for health insurance. This prohibition does not apply to other types of insurance. This federal law applies only to group insurance obtained from employers and has no affect on individuals seeking their own policies. Thirty-nine States have additional laws giving some additional protection for group health insurance, but protections for individual insurance are somewhat less common, although even that protection can be found in many of the most populous states (eg CA, HI, MD, OR, VA).A key argument given in favour of predictive testing, even if it is imprecise, is that it gives people the chance to alter their behaviour and lifestyle. For example, showing someone that they are particularly at risk of cancer if they smoke could persuade them to stop. Even this is not so simple.According to Theresa Marteau, a geneticist working at Guy's, King's and Thomas' Medical School, London and Caryn Lerman of Georgetown University Medical Centre, Washington DC, in some cases, providing genetic information on risk can decrease motivation to change. Although some people will adapt their behaviour in the light of new information, others take a fatalist view believing that this is the way that they were made and there's nothing that they can do to improve the situation. "The current evidence suggests that providing people with DNA derived information about risks to their health does not increase motivation to change behaviour beyond that achieved with non-genetic information," they say (BMJ 2001, 322:1056-1059).So where does that leave predictive testing? Undoubtedly it is here to stay. The need now is to train those at the medical front line so that they know its strengths and weaknesses. Jon Emery of the University of Cambridge and Susan Hayflick of Oregon Health Sciences University, point out that those at the front line are the members of the primary care team (BMJ 2001, 322:1027-1030). Even there, these authors see a problem: "The impact of genetics on primary care is currently small and primary care practitioners may not yet perceive genetic education as relevant. It will require increased public demand for genetic advice or the development of genetic tests of direct clinical value to drive the acquisition of genetic skills."Enter Alan Milburn. He announced this week a £30m ($42m) package aimed at bringing genetics into medical practice. It sounds good, but it only aims to double the number of genetic consultants to 140 by 2006 and double the number of support staff to 450. Not a lot, if everyone in the country starts seeking tests.What does predictive genetic testing really need? It needs tests that are strongly predictive, that relate to diseases for which there is a worthwhile treatment, and it needs a pool of people who know how to apply the tests and their results so they do good, not harm.

how clear is the crystal ball?

, and a recent report from the UK's House of Commons Science and Technology Committee raise serious concerns about the widespread use of predictive genetic testing.

Interested in reading more?

Become a Member of