Search tips
Search criteria 


Logo of bmjThis ArticleThe BMJ
BMJ. 2007 June 9; 334(7605): 1196.
PMCID: PMC1889993
Head to Head

Should genetic information be disclosed to insurers? Yes

Søren Holm, professor

UK insurers have said that they may seek approval to use the results of genetic tests for cancer from next year. Søren Holm believes there is no reason not to pass the results on to insurance companies, but Richard Ashcroft argues that the risks of disclosure justify privacy in most cases

The main argument for disclosing genetic information to insurers is that there are no good reasons for not disclosing it. If we accept that life or health insurers can legitimately seek and obtain other kinds of health information that predicts insurance risk, then we should also accept that they can seek genetic information that is predictive in the same way. There is no reason for treating genetic information differently.

What is the purpose of insurance?

The main purpose of life or health insurance is to spread the costs of expensive but unpredictable events across the pool of the insured, thereby converting a possibly catastrophic loss to a predictable regular expenditure. The actuarially fair price is the price that adequately reflects my risk plus the administration costs. This is, for instance, the reason that the price of car insurance reflects the value and type of the car, the age and sex of the driver, and the postcode of the owner, since all of these predict the chance of theft or damage and the cost of the potential loss.

In life insurance the actuarially fair price is the price that accurately reflects my likelihood of dying within the insured period and the amount I am insured for. Because my personal risk cannot be estimated with precision, insurers set premiums for groups of people with broadly similar risks.

What information should insurers be allowed?

We could have a system for life and health insurance that denied insurers access to any kind of health information. This would mean that they could differentiate premiums only according to very general risk markers (age, sex, place of residence, occupation, etc). If we think that life or health insurance is a basic social good that is essential for citizens in modern societies, such a system is attractive. It will mean that healthy people subsidise unhealthy people, but that may be acceptable as an expression of social solidarity or equality. If we choose to have such a system, or a hybrid system where you can get basic insurance without surrendering any health information but have to give this information for more extensive policies (along the lines of the current UK policy on genetic information1), we choose it not because health information is special but because we think that justice or solidarity demands risk sharing between healthy and unhealthy people.

If, however, we allow insurers to obtain some kinds of health information (body mass index, cholesterol concentration, results of a physical examination, etc) we no longer have any principled reasons for excluding genetic information. Genetic information is not special. It is not inherently more specific, predictive, sensitive, or private than other kinds of health information.2

It is also extremely difficult to define what counts as genetic information. Genetic information can be obtained without anything we would usually classify as a genetic test (is taking a family history a genetic test?). The most common genetic test is probably routine blood typing in hospitals, but does that mean that knowledge of my blood type is one of the pieces of knowledge that an insurer may not seek?

We may have good reasons to allow insurers access only to information that is properly validated and for which there is sufficient evidence that it predicts risk, but this is again a consideration that applies across the whole range of health information. It is true that many so called genetic risk factors are not well validated, but the same is true of other risk factors measured by non-genetic means.

Other considerations

It is often argued that if we allow insurers access to genetic information it will deter people from having genetic tests that are relevant to their health care. This may well be true, but the same is true for other health information (similar discussions were had about HIV testing) and it does not provide a reason to treat genetic information differently.

Another worry is that insurers may not interpret genetic information correctly and deny people insurance or levy inappropriate premiums based on faulty calculations of risk.3 This is again an obvious risk but is no reason to single out genetic information. Genetic information is not more inferentially fertile than any other kind of information and not more liable to misinterpretation. Again the earlier debates about HIV are instructive. It was claimed, probably correctly, that residence as a single man in certain areas of major cities was interpreted as a risk factor for homosexuality and HIV infection. Even if the only information we allowed insurers was the name of the person seeking insurance, surely a rather restrictive requirement, sound (but not necessarily true) probabilistic inferences could be made concerning age, sex, ethnicity, social status, etc.4 5 Because so many kinds of information can be interpreted wrongly or in discriminatory ways a better solution to the problem is to allow people to challenge decisions to deny coverage for life or health insurance, forcing insurers to make their reasoning transparent.


Competing interests: None declared.


1. HM Government, Association of British Insurers. Concordat and moratorium on genetics and insurance London: Department of Health, 2005
2. Low L, King S, Wilkie T. Genetic discrimination in life insurance: empirical evidence from a cross sectional survey of genetic support groups in the United Kingdom. BMJ 1998;317:1632-5. [PMC free article] [PubMed]
3. Holm S. There is nothing special about genetic information. In: Thompson AK, Chadwick RF, eds. Genetic information—acquisition, access, and control New York: Plenum Press, 1999:97-103.
4. Bertrand M, Mullainathan S. Are Emily and Greg more employable than Lakisha and Jamal? A field experiment on labor market discrimination Cambridge, MA: National Bureau of Economic Research, 2003
5. Fryer RG Jr, Levitt SD. The causes and consequences of distinctively black names Cambridge, MA: National Bureau of Economic Research, 2003

Articles from The BMJ are provided here courtesy of BMJ Publishing Group