APPENDIX 29
Memorandum submitted by the Centre for
Exploitation of Science & Technology (CEST)
INTRODUCTION
1. CEST is a not-for-profit organisation
whose role is to encourage and enable the appropriate application
of science, technology and knowledge. It acts as an independent
facilitator bringing together industry, Government and third parties
to explore and identify blockers and enablers of technology. Most
importantly participants focus on cross-sectoral technologies
that can lead to new applications and opportunities.
2. CEST is based in the UK, but has an extensive
European network, which brings together groups of opinion formers
(business planners, marketing specialists, scientists and technologists;
policy analysts, legislators and regulators and consumer representatives).
BACKGROUND INFORMATION—CEST'S
EXAMINATION OF
ISSUES RELATED
TO GENETICS
AND INSURANCE
3. Between September 1999 and December 2000,
CEST ran a collaborative programme examining business opportunities
and regulatory challenges in the "New Genetics".
The consortium included diverse players including: Enterprise
Ireland; Glaxo Wellcome; Nycomed Amersham; Human Genetics Commission;
Department of Health; NHS Executive; Siemens; Deutsche Telekom;
Forensic Science Service; IBM; Unilever; Amersham Pharmacia; and
Scottish Enterprise. The comments below reflect the key points
that may be of relevance to the House of Commons Science and Technology
Committee and do not reflect any particular policy stance by either
CEST or the participants.
4. As part of CEST's New Genetics initiative
a meeting was held in September 2000 to explore issues in Genetics,
Insurance and Employment. At the meeting specialist briefings
were heard from:
— Professor Sandy McCall Smith (Vice-Chair
of the Human Genetics Commission);
— Dr Jan von Overbeck (Chief Medical
Officer of Swiss Reinsurance);
— Dr Angus Macdonald (Director, Genetics
and Insurance Research Centre, Heriot Watt University); and
— Mr Iain Bourne (Office of the Data
Protection Commissioner).
In addition to the consortium members and speakers,
a representative from the Association of British Insurers also
attended this meeting. In preparation for the meeting, Professor
Sandy Raeburn (Genetics Adviser to the Association of British
Insurers) and Mr Alastair Kent (Genetics Interest Group) were
also consulted.
KEY ISSUES
IDENTIFIED IN
THE COURSE
OF PREPARATION
AND RUNNING
THE MEETING
Unexpected consequences of legislation—defining
"genetic tests"
5. There are three levels at which genetic
tests can be used: predictive, diagnostic or prognostic. Furthermore
there is no single form of "genetic test"—they
can be based on analysis of specific genes, examination of gene
products or based on broader screens or "profiles" of
Single Nucleotide Polymorphisms (SNPs). SNP profiles are already
being developed by major drug companies, tailored to providing
information as to a patient's likely response to a given medicine
(ie a prognostic profile). Such tests can be tailored so that
they do not give collateral information other than the likely
safety and efficacy of a particular treatment course. The application
of the specific type of genetic test being applied and its use,
therefore determines the type of information generated and its
wider utility.
6. Will legislation to address questions
of genetic testing prevent current medical examination data being
used in the underwriting decision? This question is relevant as
the underlying results of many existing medical tests, although
not using molecular technologies, are determined by the underlying
genotype. Where there is either a direct Mendelian-linked genetic
basis for a trait (eg blood type) or where penetrance of a susceptibility
gene or genes is high, then the results of an existing biomedical
test give rise to information that is either directly or strongly
linked to an individual's underlying genotype. For example, high
levels of cholesterol are usually caused by a particular genotype,
although this does not necessarily mean that all with this genotype
will have an increased risk of heart disease. Even ultrasound
scanning, which may pick out genetically determined foetal kidney
defects, provides information that can be extrapolated to a person's
genetic make-up.
7. Care must be taken therefore when precisely
defining: what constitutes a genetic test; when a genetic test
can be of benefit; and when a genetic test could add value to
medical or underwriting processes.
Access?
8. Is access to life assurance a right?
Since life assurance is provided under voluntary individual contracts
(mutuality) rather than as a compulsory scheme (solidarity, eg
UK health insurance) can private insurers be compelled to provide
universal coverage?
How deterministic are the results of a test?
9. Much of the concern about genetic testing
is based on the premise that it is determinist—ie that if
you have the mutation you will get the condition, in other words
inheritance is Mendelian in fashion and penetrance of the genotype
is high. However this is clearly not the case in many instances
and every disease or human condition is influenced to a lesser
or greater extent by both the genotype and the environment.
10. Diseases such as asthma, diabetes and
heart disease are caused by complex interactions of multiple genes
(and the various alleles) as well as environmental influences.
Even infectious diseases are influenced by an individual's genotype,
although to a much lesser extent—this is typified by people
suffering to different degrees when they catch the same cold or
influenza infection.
11. For a single gene defect with high penetrance,
such as cystic fibrosis, mutations in different parts of the gene
sequence can give rise to different levels of disease and mutations
in other genes may make the condition milder or more severe. In
addition there is an environmental influence, although this is
minimal.
12. In terms of predictive genetic tests,
these may be useful where they suggest a lifestyle modification
that is clinically relevant. For example, if it is known that
an individual has haemochromatosis early in life and lifestyle
modifications are made, insurance risk may be decreased. If genetic
factors for diseases are declared could premiums be adjusted to
reflect subsequent lifestyle changes, perhaps backed by regular
health checks? The key point is that genetic testing is only one
way of assessing an individual's total risk. Even if genetic test
information becomes more deterministic, an assessment of the risk
must encompass environmental factors, such as smoking, occupation
and hobbies.
Accuracy and relevance
13. How accurate are the genetic tests on
which underwriting is to be based? Are they relevant to actuarial
calculations? These questions are being addressed by the Genetics
and Insurance Committee, which was set up in 1999 to examine the
"scientific and actuarial relevance of specific genetic
tests" to insurance underwriting. GAIC are currently
completing a review of 10 tests (for seven conditions) identified
as relevant by the Association of British Insurers (ABI). The
wider social aspects of genetic testing and insurance will be
considered by the Human Genetics Commission. The first test (for
Huntington's Disease) was approved for use in underwriting decisions
in October.
Association of British Insurers Code
14. The ABI have developed a "Code
of Practice" that is binding on their members in the UK.
The code prohibits companies from asking applicants to take a
genetic test when applying for insurance. They only ask applicants
to share knowledge of certain test results they have already taken.
Clearly there is also the potential to benefit from negative test
results for genes associated with a disease where there is a strong
familial history—the problem is people are reluctant to
take such tests if there is perceived to be an additional penalty
either in admitting to the test or testing positive.
Inevitability
15. Currently, although we know we are all
genetically different we insure ourselves against uncertain events
by regarding all people as essentially the same, the uncertainty
as random and by consolidating the risks across a large number
of people. Underwriting is used to adapt the "randomness"
principle by determining the degree of risk of individual applicants.
Depending on where on the spectrum of risk people fall, individuals
are assigned to different risk pools and charged different premiums.
The traditional risk assessment model relies on shared information
between the applicant and the underwriter. In essence private
insurers assess risks by calculating how often an uncertain event
takes place, pooling the risks and setting the premiums at a level
where they are able to pay claims.
16. In the rare case where a test result
predicts a sure event (eg death from Huntington's Disease where
repeat length is known) insurance is not a sensible mechansim
to cover the event. Insurance is based on predicting uncertainty
and if the uncertainty is removed there is no need for insurance.
Adverse selection and liquidity of insurance companies
17. If an applicant for insurance knows
they have a disadvantageous medical history (eg a highly predictive
test result) but does not disclose it there will be an asymmetry
of information and they will be quoted a lower premium than the
risk they represent to the insurance pool. This is adverse selection
and will lead to the premiums of the other members of the pool
having to be increased to cover the adverse selection loss. Quite
rightly insurance companies must limit adverse selection to remain
competitive and to prevent bankruptcy—hence the requirement
to declare the results of particular tests or indications relating
to medical history that may be predictive in nature.
18. There are three issues to consider in
terms of adverse selection:
— First, in the case of life insurance
there will be few significant problems arising from single gene
disorders unless the sums assured are above average (this represents
the real adverse selection risk). This is because the adverse
selection loss is spread over a large pool. The increased premiums
in these cases are unlikely to be above 10 per cent and may be
negligible.
— Second, multifactorial disorders
are unlikely to be a problem because the additional level of risk
added by each mutation is small.
— Third, there will be more difficulties
in the area of long-term care and private health insurance. This
is mainly because the size of these markets is so much smaller
and the individual claims larger and more open-ended. In addition
the risk cannot then be shared across such a large pool as for
life insurance.
Will testing people with a family history of genetic
disease increase their chance of getting insurance cover?
19. In the future, difficulty could arise
when complex predictive genetic tests or profiles are more commonly
available. These are unlikely to provide information regarding
an individual's likely susceptibility to long-term disease above
and beyond that which can be gleaned from familial history. Indeed,
genetic tests or profiles may actually benefit those whose family
history suggests they may be at risk of longer-term problems,
if they have not inherited the particular mutation or particular
gene mutation profile in question.
20. Back in 1997 The Wellcome Trust carried
out a survey of certain groups of people with genetic diseases
or with a family history of disease and sampled their experience
of getting a range of insurance cover. The main finding was that
there was a lot of confusion from the insurance industry and applicants
that were only carriers were being discriminated against. The
discrimination identified by The Wellcome Trust was not systematic
and much should have improved since 1997 when the survey took
place.
A case for state support for uninsurable people?
21. In the case of a highly predictive genetic
test or gene profile, consideration should be given as to whether
additional or alternative state support should be provided. Clearly,
in a free-market economy this is not an option and even in a country
where there is a welfare state, it is arguable whether any additional
insurance or medical support above and beyond that provided through
the welfare system is appropriate.
Where is further research required?
22. It was suggested by a number of consortia
participants that research into the actuarial consequences of
use of test results is required to improve the quality of decision-making.
The purpose of this research should not be seen as to allow more
discrimination, but rather to provide quantitative data to illuminate
policy debate, identify potential problems, allay unnecessary
fears and achieve fairness in provision.
CONCLUDING COMMENTS
23. These comments were drawn together from
the recent New Genetics consortia organised by CEST. It
does not reflect any policy recommendation by CEST or the policy
of those companies that participated, but is meant to summarise
the key points raised.
24. Perhaps the most important issue surrounds
the need for clarity of definition. The term "genetic tests"
covers a multitude of tests based upon different methodologies
and with widely differing purposes. It is essential therefore
that the House of Commons Science and Technology Committee adopts
consistent and clear definitions to understand the range of issues
that could arise with respect to genetics and insurance.
29 March 2001
| 
