J7ournal ofMedical Ethics 1999;25:151-156

Genetic testing: a conceptual explorationR L Zimmern Public Health Genetics Unit, Cambridge

AbstractThis paper attempts to explore a number ofconceptual issues surrounding genetic testing. It looksat the meaning of the terms, genetic information andgenetic testing in relation to the definitions set out bythe Advisory Committee on Genetic Testing in theUK, and by the Task Force on Genetic Testing in theUSA. It argues that the special arrangements thatmay be requiredfor the regulation ofgenetic testsshould not be determined by reference to the nature ortechnology of the test, but by considering thosemorally relevant features that justify regulation.Failure to do so will lead to the regulation ofgenetictests that need not be regulated, and wouldfail tocover other tests which should be regulated. The paperalso argues that there is little in the nature of theproperties ofgene tests, using DNA or chromosomes,that in itselfjustifies a special approach.(7ournal ofMedical Ethics 1999;25:151-156)Keywords: Genetic testing; genetic discrimination; geneticinformation; privacy; confidentiality

IntroductionThe Advisory Committee on Genetic Testing(ACGT), in its Report on Genetic Testing for LateOnset Disorder' under the chairmanship of Profes-sor Peter Harper, and the US Task Force onGenetic Testing, in Promoting Safe and EffectiveGenetic Testing in the United States,' have eachattempted to define genetic testing. The AdvisoryCommittee defines it as "testing to detect thepresence or absence of, or alteration in, a particu-lar gene, chromosome or gene product"; the US TaskForce as "the analysis of human DNA, RNA, chro-mosomes, proteins, and certain metabolites in order todetect heritable disease related genotypes, muta-tions, phenotypes or karyotypes for clinicalpurposes". The latter explicitly states that thedefinition "excludes tests conducted purely forresearch, tests for somatic mutations as opposedto heritable mutations and testing for forensicpurposes".The ACGT carefully distinguishes between

diagnostic genetic testing and predictive genetictesting, categorising the latter into presympto-matic testing and susceptibility testing, but themain body of the text is confined to "presympto-matic testing of healthy relatives with a family his-

tory of serious late onset disorder with a cleargenetic basis and commonly following dominantinheritance". Issues connected with populationscreening, the use of genetic tests in symptomaticindividuals and susceptibility testing for disordersinvolving multiple genetic and environmental fac-tors are dealt with cursorily in three short appen-dices.A recent paper by Peter Harper entitled,What

do we mean by genetic testing?,' defines genetictesting as "the analysis of a specific gene, itsproduct or function, or other DNA and chro-mosome analysis, to detect or exclude an altera-tion likely to be associated with a geneticdisorder". This definition includes not only DNAtests but tests of its products and function. In thepaper Professor Harper also argues that theconcept of a genetic test should include not onlythe laboratory analysis itself, but the preliminarypreparation and counselling of the patient andsubsequent interpretation and support. In so faras it is possible to categorise tests as being genetic,as distinct from non-genetic, I believe hisdefinition to be correct, including not onlybiochemical tests, such as those for phenylketonu-ria and hypercholesterolaemia, but others, such asthe use of renal ultrasound in polycystic kidneydisease or computerised tomography in tuberoussclerosis. For the sake of simplicity I shall, in theremainder of this paper, use the term genetic testingto refer to any type of test that indicates that aperson is likely to have a genetic or familial disor-der, and the term gene testing to refer to tests con-fined to the analysis of DNA, RNA or chromo-somes.The purpose of the reports from both the US

Task Force and the Advisory Committee onGenetic Testing in the UK is to protect the publicfrom indiscriminate genetic testing withoutproper safeguards and controls, and to ensure thatgenetic information obtained as a result of suchtests is used in an appropriate manner, withadequate protection of its privacy and confiden-tiality. This purpose appears to be quite properand rests on the view that genetic information issomewhat special and different from other typesof health information, and requires a different andspecial approach from both health professionals

152 Genetic testing: a conceptual exploration

and society. One of the most important reasonsbehind moves to protect the public from indis-criminate genetic testing is a fear of geneticdiscrimination, and of the negative actionsemployers, insurers, mortgage and other lendersmight take against people known to have a geneticdisease or susceptibility to a multifactorialdisorder.4 5 I shall attempt in this paper to look atfour issues: first, to explore the meaning of geneticinformation and to ask if it can be adequately setapart from other types of health information; sec-ond, to explore in a similar fashion, the concept ofgenetic testing as a separate, identifiable categoryof medical test; third, to argue that the notion ofregulating genetic testing as the means of address-ing the problems of potential genetic discrimina-tion and of protecting the patient and the public ismisconceived, and that the morally relevantfeatures that require certain aspects of medicalpractice to be regulated are not those which relateto the categorisation of a test as being genetic (asdistinct from non-genetic), and fourth, to arguethat the ACGT is mistaken in confining theirdefinition of genetic testing to what I refer to asgene testing, that is tests which only involve DNA,RNA or chromosomes.

Genetic informationThe term genetic information, may be used in oneof two ways. First it may be regarded asinformation about the genetic constitution ofindividuals, their genes or chromosomes, andtheir inheritance. Second, and by contrast, geneticinformation may be taken to refer to anyinformation from which we may infer knowledgeabout the genetic constitution of individuals. Inclinical practice geneticists, concerned classically(at least until recently) with Mendelian disorderssuch as Huntington's disease or DuchenneMuscular Dystrophy, or with chromosomal de-fects such as Down's syndrome, have used adetailed analysis of the family history and of chro-mosomes or DNA, as the information base fromwhich to make a diagnosis or to predict the level ofrisk in their patients. These disorders have tendedto manifest clear-cut patterns of inheritance andto confer a very high risk on family members.They may, without any confusion, be referred toas genetic diseases, and may also be distinguishedrelatively easily from other disorders where thepatterns of inheritance are significantly less clear-cut.

Notwithstanding this distinction, it is also thecase that genetic factors play a part in the patho-genesis of common disorders such as Alzheimer'sdisease, diabetes, breast and colo-rectal cancersand ischaemic heart disease. This has been known

for many years. In these disorders the genetic fac-tors range from rare single genes which accountfor a tiny fraction of familial disease, for example,BRCA1 or BRCA2 in breast cancer or PS-1 ininherited Alzheimer's disease, to a number of asyet unidentified low-penetrance susceptibilitygenes or polymorphisms, which, interacting withenvironmental factors, account for the muchlarger proportion of familial cases of thesecommon disorders. A woman with a first degreerelative with breast cancer will be twice as likely asa woman without a family history to develop thedisease, but would still be very unlikely to harbourgene mutations in BRCA1 or BRCA2.6

Infectious diseases such as tuberculosis andmeningitis, behavioural traits, such as hyperactiv-ity, and psychiatric disease, such as schizophrenia,have also been shown to have a significant geneticcomponent. For these (non-Mendelian) disordersit would be more correct to speak of genetic sus-ceptibility rather than genetic disease, but thegenetic component nevertheless confers somedegree of increased risk for the individual andsome degree of heritability and implications forother family members. In these instances the con-cept genetic information is more problematic, atleast if used in the first sense, unless it. can beshown directly either (a) that genes have somepart to play in the determination of the clinicalfeatures of a disease or trait, or (b) that somedegree of heritability is conferred on family mem-bers.

In relation to the second use of the term, wherephenotypic manifestations allow us to infer thepresence of genetic factors, we may reflect on thefact that the most medically unsophisticated of usis able to elicit the difference between men andwomen, a clear-cut genetic trait. It is equally pos-sible to imagine that an astute observer with clini-cal knowledge may make a diagnosis after thedetection of scoliosis, cafe au lait spots andneurofibromata in a fellow bather on a beach.These features, visible to all, allow the inference ofa genetic trait by simple observation. A physicianin his surgery may elicit a family history of bowelcancer in one patient or of deaths from heart dis-ease at a young age in another. In all these exam-ples, whether through simple observation outsidea medical setting, or the use of clinical history-taking and examination during a consultation,health information passes from one to anotherperson; and in each a genetic component may beinferred.The reality is that all human variation and all

disease have both genetic and environmentaldeterminants. The expression of the disease thatresults in illness is a consequence ofboth gene and

Zimmern 153

environment. A positive family history confers onan individual an enhanced risk of disease. In com-bination with harmful environmental factors, suchas smoking, the risk may be further increased.Knowledge of these risks provides the physicianwith the clues that enable diagnoses to be madeand prognoses to be discussed. All form part ofthe data-set that we call health information; andfrom each, to a greater or lesser extent, we mayinfer knowledge about the genetic make-up of anindividual.

Genetic information is therefore a term which iscoherent and sound, and refers to informationabout an individual's genes and his or her inherit-ance. But in as far as it may in practice be used inone of two ways, either directly, to refer toinformation about a person's DNA or chromo-somes, or indirectly, to infer from some pheno-typic feature, clinical, biochemical or radiological,the use of the term is often ambiguous. Any state-ment must show precisely in which of these twosenses it is used.

Genetic testingThe same ambiguities of usage arise when we tryto distinguish genetic from other medical tests. Atest, whether in the form ofinformation elicited bytaking a history or a physical sign, a physical pro-cedure such as a sigmoidoscopy, or a biochemicaltest or radiological procedure, is an interventiondesigned to alter the pre-test assessment of theprobability of the subject having a particulardisorder to a different post-test (whether greateror lesser) probability. A genetic test is no differentfrom any other test in this regard.By convention, the term, genetic test, is usually

taken to refer to tests made directly on geneticmaterial such as DNA or on chromosomes, andwhich, in this paper, I have referred to as a genetest. This usage suggests that there are differencesbetween genetic tests and, by inference other tests,and applies the distinction between that which iscalled "genetic" or "non-genetic" to the test itself.By analogy with the discussion of the use of theterm, genetic information, it is possible to conceiveof another use of the term, genetic test, and toregard any clinical, haematological, radiological orbiochemical test, from which information aboutthe gene or the inheritability of a disorder to beinferred, as a form of genetic testing.The reason why in conventional usage the term

genetic test is used in the narrower sense, the sensewhich I call a gene test, is that it is argued that theregulation of gene testing will protect patients andsociety from certain ethical and social implica-tions of the diagnosis of genetic disease. Thisobjective, in my view, cannot be achieved by refer-

ence to the nature and technology of the test pro-cedure, but must relate to morally relevantfeatures which are more likely to be associatedwith the nature of the disorder or disease itself. Iwill deal with this issue in the next section of thepaper, where I will attempt to argue that neitherthe regulation ofgenetic testing nor ofgene testingis likely to achieve the desired result.

Genetic testing and the prevention ofgenetic discriminationI start from the premise that the prevention ofgenetic discrimination and of adverse ethical con-sequences to families and individuals, includingthose of privacy and confidentiality, of the abilityto predict in advance of its clinical appearance theonset of genetic disease, is worthwhile. A societywhich supports institutions and structures whichadversely affect people with genetic disease morethan those with disease that is not primarilygenetic in origin is clearly undesirable and shouldbe discouraged. To the extent that there isevidence for the contention that certain groups insociety suffer discrimination, one may argue thatsteps should be taken, through voluntary codes ofpractice or through regulation and legislation, tominimise or prevent such discrimination. Withinthe UK, evidence of racial and of sex discrimina-tion led to the Race Relations Act 1976 and theSex Discrimination Act 1975; the advent of HIVand AIDS led to codes of practice regarding theuse of HIV-testing and a more stringent require-ment to protect confidentiality that did not extendto other medical tests.7 In each of these instancesthere was evidence or a prima facie reason tobelieve that, without explicit action, discrimina-tion, with documented adverse consequences forindividuals, would occur.

In the case of late onset Mendelian disorderssuch as Huntington's disease the role played bygeneticists and the ethical issues surroundinggenetic testing, whether through linkage analysisor direct gene testing, has been exemplary8; but itis very much to be doubted whether a blanketextension of those principles through codes ofpractice or regulation of genetic tests to non-Mendelian disorders would be ofuse. The morallyrelevant circumstances in which regulation is war-ranted, whether voluntary or statutory, are, in myview: (a)where there is already de facto evidenceof significant discrimination or (b)where the pre-dictive value of a test and the probability of devel-oping the disease, or phenotypic manifestations,attributable to the genetic defect is high enough togive employers, insurers or others in society a rea-son to justify discriminatory policies.

154 Genetic testing: a conceptual exploration

These morally relevant conditions apply to alltests from which a genetic predisposition might beinferred, not just gene tests. However, it does seemto me that a practical distinction for decidingwhether regulation is required, using the condi-tions specified above, might be made by distin-guishing between those tests which confer veryhigh risks on individuals and a high probability ofbeing inherited by other family members, such asMendelian disorders, and tests for other diseases.

Pre-test risks associated with Mendelian disor-ders in first degree relatives, parents, children andsiblings, are usually either of the order of 1 in 2 or1 in 4. Genetic testing may either reduce the riskof having the disorder to zero (or almost so) in aparticular individual or increase it to a certainty(or almost so). For them, the very high predictivevalue of the test could well give society reasons tojustify discriminatory policies. It is for this reasonthat Peter Harper has so cogently argued for theneed to conceptualise a genetic test as one whichincludes pre- and post-test information andadvice. It is for the same reason that I believe thatin these circumstances regulation might well berequired, but, if required, it should be applied toall genetic tests, and not just gene tests, and whereethical considerations do not dictate the need forregulation, it should not be used at all.By contrast, individuals with multifactorial dis-

eases such as diabetes or heart disease have lowerpre-test risks than patients with Mendelian disor-ders. The test may also lead to the formulation ofa post-test risk which, whether higher or lowerthan the pre-test risk, will nevertheless notincrease to near certainty or decrease to near zero.For this and other reasons, the ACGT reportattempts to distinguish (a) "pre-symptomatictesting", defined as "testing carried out in asymp-tomatic individuals to provide definitive infor-mation about that individual's future health",from (b) "susceptibility testing", defined as "test-ing which provides information about the geneticcomponent in a multifactorial disorder".2 In rela-tion to "pre-symptomatic testing" the report goeson to say: "Such a test result may indicate that theindividual has a high likelihood of developing thedisorder or excluding it. Predictive testing is mostfrequently used in late onset autosomal dominantdisorders such as Huntington's disease".A case might therefore be made for having dif-

ferent principles and a different approach to deal-ing with tests and information obtained from peo-ple with a predisposition to a Mendelian disorderfrom those with genetic susceptibility in multifac-torial disease. However, it is not the distinctionbetween genetic tests, however defined, as distinctfrom non-genetic tests, that confers the need for a

different approach in the two instances. The con-tention that all genetic tests should be subjected toa regulatory framework, irrespective of whetherthe morally relevant indications outlined aboveexist or not, seems far too wide and unnecessary.It would theoretically include within its remit thetaking of a family history, the measurement of ablood pressure, and the use of biochemical andother pathological tests, such as cholesterol orHLA typing, as well as gene tests, as narrowlydefined. However, an approach which requiresonly gene tests to be regulated is equally withoutfoundation. It would lead to a situation in whichmany tests, not based on gene technology, escapedregulation when they ought to be regulated; whilesome gene tests would be regulated, for which nomorally relevant reason exists.

Gene testingThe social and ethical issues that are raised ingenetic disease, including the potential for dis-crimination by insurers and others,9 10 arise as aconsequence of making a diagnosis in a patientwith symptoms, or in establishing an increasedrisk in pre-symptomatic individuals, or in familymembers. The social and ethical consequencesarise from the making of the diagnosis of a famil-ial disease. In Peter Harper's words: "The relevantfactor is not the technology but the fact that thetest is detecting a change directly related to aninherited disorder".3 This statement alone isadequate to raise in my mind the difficulties ofaccording special status, including the need forcounselling, when a diagnosis or an increased riskof genetic disease is predicted through the use ofDNA technology but not when it comes about asa result of a biochemical or a radiological test. Anasymptomatic adolescent from a family withknown polycystic kidney disease is at the same riskof discrimination, and the diagnostic procedure isjust as worthy of regulation (or not) whether it isan ultrasound scan or a DNA test.The ACGT report outlines a number of differ-

ences between genetic tests (meaning gene tests) andother medical tests, stating that while these others"may detect changes in those at risk before symp-toms occur, they still reflect an early stage of thepathological process".' I list some of these. We aretold that DNA is extremely stable, that DNAanalysis can be carried out at any point from con-ception to old age, and that the presence orabsence of an abnormality in a gene test isunaffected by whether the individual has symp-toms or not. Testing for inherited disorders is saidto be different from other clinical tests becausesuch testing may reveal important informationabout relatives. Another important feature of

Zimmern 155

testing for inherited disorders is said to be thepower of such testing to predict the potential futurehealth of the individual and that the test result maycause anxiety in an otherwise healthy individual, sothat the consultation before and after a gene test maybe different from that needed in many other typesof medical test. These properties, said to be thosewhich relate to analysis of DNA itself, may bedescribed by the terms stability, heritability, predic-tiveness and sensitivity. I find it difficult to convincemyself that any of these are properties of DNA orgene tests that are not shared by other diagnosticinterventions in medical practice.The stability of DNA is a feature which does

give rise to concern, particularly when specimensare stored and subsequently used for purposes notcovered by the original consent. However, thisparticular ethical issue is about regulating the useof DNA, principally by third parties, and its rela-tion to the law of consent and of proprietary rightsto human tissue, rather than to diagnostic or pre-symptomatic testing in the clinical situation. Her-itability is the defining feature of familial andinherited disorders. But conceptually, the degreeof heritability is a feature of the condition understudy, highly heritable in Mendelian disease, lessso in multifactorial disorders. The phenotypicmanifestations of heritability, referred to as thepenetrance of a gene, is as likely to be determinedby environmental as gCeletic factors. Whatever themechanisms, heritability is certainly not a featureof the technology used to diagnose the disorder.

Predictiveness, the ability to predict the risk ofdisease in asymptomatic individuals is a feature ofall tests. The likelihood of myocardial infarctioncan be predicted from knowing the risk factors towhich the subject is exposed. DNA informationdoes allow an individual to know if he or she iswith or without a deleterious gene mutation, in afamily where the mutation exists, and may confera degree of knowledge that cannot be obtainedfrom phenotypic information directly. In thatsense gene testing, in certain circumstances, maygive important data about the heritability of thedisease in an individual. In many other instances,phenotypic information is much more useful as apredictor of disease manifestation. The predictivepotential of a family history in genetic disease iswell known to geneticists, whereas informationabout the presence of a mutation in a gene, or ofits exact DNA sequence, may not always allow oneto know if the mutation will give rise to manifesta-tions of disease. Put in a different manner, theability of a test based on genotypic information, tochange the probability that a patient is at risk ofdisease, may, in most cases, be lower than that of

an appropriate test based on phenotypic infor-mation.

It is said also that the use of DNA testing isethically much more sensitive and will lead toanxiety, both for the patient tested and potentiallyfor other family members. The sensitivity whichleads to anxiety is there whenever a test result isconveyed to a patient, and is not a specific featureof gene testing. It is also a feature of participationin screening programmes. The capacity to createanxiety is not a specific feature of gene tests.

ConclusionThe fact that I have attempted to explore the con-cepts of genetic information and genetic testing,and to suggest that the use of a regulatory frame-work to control such information and tests, asconventionally and narrowly defined, is illogicaland without foundation, does not mean that I amagainst the construction of a regulatory frame-work in circumstances where discrimination islikely to exist. The purpose of my analysis hasbeen to demonstrate that:

(a) the special arrangements that may be requiredfor the regulation of tests which lead to thediagnosis of genetic disease should be deter-mined not by reference to the nature of thetest but by morally relevant features whichshould apply to all genetic tests and not just togene tests

and

(b) a regulatory framework which applied to allforms ofgene tests, and only to gene tests, wouldlead to the regulation of tests that should notbe subject to special regulation, and would failto cover other tests, from which a diagnosis ofgenetic disease may be inferred, which shouldbe so regulated.

The attempt by both the US Task Force and theACGT to prevent the abuse of genetic infor-mation by regulating gene testing is the result of thevery clear and genuine concerns of geneticists,based on their experience of Mendelian disorders.I suggest an alternative practical basis for the pre-vention of discrimination might lie in differentiat-ing Mendelian (and other disorders where thedegree of genetic risk is of the same order of mag-nitude as Mendelian traits) from other disorders,or whenever there is prima facie evidence that dis-crimination occurs or is likely to occur; and thatthis distinction should be irrespective of the tech-nology or the type of test used to make thediagnosis or to categorise the risk.

156 Genetic testing: a conceptual exploration

Although my motivation in undertaking thisanalysis has been to stimulate debate there are twoimportant practical consequences which arise ifmy arguments are accepted. First, in relation tothe multifactorial or polygenic disorders, thequestion of whether a specific regulatory frame-work will be required will depend on whether defacto there is likely to be discrimination, or othersocial and ethical implications, as a consequenceof making clinically and ethically relevant predic-tions in either a symptomatic or an asymptomaticindividual. I do not at present see much evidencethat conventional risk prediction in medical prac-tice has resulted in significant problems. Second,in relation to Mendelian disorders, I believe thatregulation, if deemed necessary, should cover allforms of diagnostic tests and not just gene testsbased on DNA analysis. In these cases, arrange-ments for counselling and pre- and post-test con-sultation and for specific consent should be avail-able irrespective of whether the consultation ismade with a geneticist who uses DNA technologyor a physician who uses biochemical or radiologi-cal techniques in reaching a diagnosis.

In spite of my analysis there may be some whowould argue that, notwithstanding the logic ofmyarguments, the practical reality of the fear andmistrust of DNA technology by the publicrequires that society acts to regulate the technol-ogy itself. For these reasons they might argue thatthe narrow definition of genetic testing used by theACGT should receive support. In favour of thisview is the contention that the stability of DNAand the possibility of third parties using it incircumstances outwith the original consent givesrise to important ethical concerns which need tobe addressed and for which codes of practice haveto be established. I would not necessarily dissentfrom this argument, but I would wish those whouse it to be clear that the motives for applying adistinction based on technology are pragmatic. Iwould also subscribe to the view that DNAtesting, like all other pathological tests, should besubject to national quality control mechanisms.However, whether the use of gene tests needs anygreater degree of regulation than any other formof medical test, and whether such regulationwould in itself prevent discrimination is question-able. The experience of HIV-testing suggests that

the procedural restrictions on the collection andthe dissemination of information about HIVstatus have not proved very effective in eliminatingdiscrimination against HIV-positive individualsand against those afflicted with AIDS within oursociety. These comments are an attempt to exposewhat I believe to be an analytical flaw in the debateabout genetic testing. I doubt very much if thepractical outcome which I seek to achieve differsgreatly from that desired by those who argue forthe distinction between genetic and other tests. Ihope that this paper might provide an interestingalternative viewpoint and generate some debateabout the definition and nature of genetic testing.

AcknowledgementsI thank very warmly, Professor Martin Bobrow, DrTony Hope, and Dr Onora O'Neill for their veryhelpful comments on the first draft of this manu-script.

DrR L Zimmern, MA, FRCPJ FFPHM, is Director ofthe Public Health Genetics Unit, Cambridge, Consult-ant in Public Health Medicine, Cambridge &Huntingdon Health Authority, and Associate Lec-turer, University of Cambridge.

References1 Advisory Committee on Genetic Testing. Report on genetic test-

ing for late onset disorders. London: Health Departments of theUK, 1998.

2 Task Force on Genetic Testing of the NIH-DOE WorkingGroup on Ethical, Legal and Social Implications of HumanGenome Research. Holtzman NA, Watson MS, eds. Promotingsafe and effective genetic testing in the united states. Bethesda:National Institute of Health, 1997.

3 Harper PS. What do we mean by genetic testing? J7ournal ofMedical Genetics 1997;34;749-52.

4 Harper PS, Clarke AJ. Genetics, society and clinical practice.Oxford: Bios Scientific Publishers, 1997.

5 Holtzman NA. Proceed with caution: predicting genetic risk in therecombinant DNA era. Baltimore: Johns Hopkins UniversityPress, 1989.

6 Pharoah PDP, Day NE, Duffy S, Easton DF, Ponder BAJ.Family history and the risk of breast cancer: a systematic reviewand meta-analysis. InternationalJournal of Cancer 1997;71;800-9.

7 McGleenan T. Rights to know and not to know: is there a needfor a genetic privacy law? In: Chadwick R, Levitt M, Shickle D.The right to know and the right not to know. Aldershot: AveburyPress, 1997.

8 Harper PS, Solden J, Tyler A. Predictive tests in Huntingtondisease. In: Harper PS, ed. Huntington disease. London: W BSaunders, 1996.

9 Association of British Insurers. Genetic testing: ABI code ofprac-tice. London: Association of British Insurers, 1997.

10 Human Genetics Advisory Commission. The implications ofgenetic testingfor life insurance. London, 197.9