BMJ

Diagnostic Kits And The Clinical ChemistAuthor(s): Mary WarnerSource: The British Medical Journal, Vol. 280, No. 6210 (Feb. 2, 1980), pp. 329-330Published by: BMJStable URL: http://www.jstor.org/stable/25438723 .

Accessed: 28/06/2014 12:41

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact support@jstor.org.

.

Digitization of the British Medical Journal and its forerunners (1840-1996) was completed by the U.S. NationalLibrary of Medicine (NLM) in partnership with The Wellcome Trust and the Joint Information SystemsCommittee (JISC) in the UK. This content is also freely available on PubMed Central.

BMJ is collaborating with JSTOR to digitize, preserve and extend access to The British Medical Journal.

http://www.jstor.org

This content downloaded from 91.213.220.138 on Sat, 28 Jun 2014 12:41:59 PMAll use subject to JSTOR Terms and Conditions

BRITISH MEDICAL JOURNAL 2 FEBRUARY 1980 329

Rehabilitation and manual medicine

Sir,?May I join the discussion on re

habilitation (12 January, p 111), and make a

plea for the linkage of rehabilitation and

manual medicine in the United Kingdom in

1980? The manual medicine physician concentrates his skill on attention to posture, exercise, and manipulative therapy. He

restores a normal range of mobility, thereby

satisfying an essential requirement of re

habilitation.

"Country house" rehabilitation is wholly

admirable, but there is a gap in our nation's

health care system with regard to the specialist examination and treatment of back pain.

Rehabilitation as a subject is incomplete without acknowledging the part manual

medicine has to play, just as the service to

back pain sufferers will remain second class if

manual medicine expertise is ignored.

Norman Healey Honorary Secretary, British

Ost?opathie Association

London WIN 1PE

Abortion (Amendment) Bill

Sir,?Mr D B Paintin, of the department of

obstetrics and gynaecology at St Mary's

Hospital (26 January, p 248), has every right to oppose my Abortion (Amendment) Bill.

He does not, however, have the right to

publish inaccuracies about the Bill. The

words "serious" and "substantial," contrary to Mr Paintin's claim, do appear in the

Abortion Act 1967.

Mr Painton's main concern appears to be

the effect that clause 4 of my Bill will have on

private practice, but he has misunderstood the

point. Before the last war doctors practised what was called "fee splitting." This was an

arrangement whereby the referring doctor

received money from the consultant to whom

he had referred a private paying patient. As a

result of BMA disapproval, this practice

disappeared shortly after the war. The Select

Committee on Abortion in 1975 objected to

the practice of certain abortion counselling services receiving cash payments in respect of

patients sent to abortion clinics. The purpose of clause 4 is to end this modern version of

fee splitting. The Standing Committee on my Bill was

told that a charge for counselling of ?16 was

made by the British Pregnancy Advisory Service, for which the counsellor's fee was

?6-60 per session. This counselling is normally done by lay persons. Mr Paintin's reference to

counselling by two doctors must be a relatively rare practice. The purpose of my clause 4 is to ensure that the pregnant woman receives

unbiased counselling, free from pressures of a

financial nature.

John Corrie

House of Commons, London SW1A OAA

The clinical chemist and the future

Sir,?I read the letter by Mary Warner

(15 December, p 1581) with interest; as I

reread it I was assailed with a growing feeling of dismay. As an ardent anti-letter-writer I

quelled my stirring breast, filed the journal and

returned to the less arduous task of writing Christmas cards. The letter of Dr R D

Eastham (12 January, p 116) has, however,

awakened my dormant literary instincts: the two letters taken together demand a riposte.

Miss Warner regards self-monitoring of blood glucose as "the tip of an enormous

iceberg" and appears to be doing her best to melt it or at least push the iceberg back

under the water. She expresses, quite rightly, a certain distrust of "diagnostic kits" and

reagent strips, but equally correctly points out that the DHSS plays an important role

in evaluating such kits and strips. She then,

however, takes a quantum jump in logic and

makes the basically false assumption that the

clinical chemistry laboratory as the home of

chemical expertise should be the place where tests are performed and the aforementioned kits used.

Sadly, this defensive territorial reaction

bedevils modern clinical chemistry. Obviously clinical chemists are trained to perform chemical tests reproducibly and accurately, but is it in the best interest of the patient to

have all these tests performed in a recognised

laboratory ? In many cases it would be to the

advantage of the patient to have tests per formed at the bedside with results immediately available rather than two to 24 hours later.

Would it not be a clinical improvement to

have serum amylase values available im

mediately when one is faced with a patient with possible acute, pancreatitis, or cardiac

enzymes for the patient with suspected myocardial infarction ? Those who, like me, do a Monday morning ward round would also, I am sure, appreciate Monday morning tests rather than the stale results of tests from the

preceding Friday. The potential time saving and increased speed in diagnosis and manage

ment, with lessening of patient distress, are

obvious. Dr Eastham, in his mildly icteric attack on

blood glucose meters (which in the North-east of England cost less than ?100?inflation

must be worse in the South-west), is guilty of a similar logical fallacy. One of the benefits of

home monitoring of glucose is to enable the

patient to make decisions about his or her

therapy and diet immediately. Using filter

paper methods removes this benefit?there is an inevitable delay of 24 hours or more

before the result is available. It is also worth

commenting that we are no longer routinely

dispensing meters in that the new Boehringer test strip can be read visually with acceptable accuracy. This last development shows

incidentally that the best of the manu

facturers exist not only to make a profit but also to provide the best service for the patient (rather than the contrary, as suggested by

Miss Warner). So where should the clinical chemist stand

if tests are to be done at the bedside ? I am

far from suggesting that we should climb out on to a metaphorical limb and then hand a

well-honed axe to our Scrooge-like ad

ministrators. Instead we should be doing our

best to provide the necessary foolproof methods for our junior (and even senior?) clinical colleagues to use. Clinical chemistry is in the process of being revolutionised by the new solid-phase chemical methods, as

used in test strips, which are eminently suitable for side-ward testing. Miss Warner's

salutary experience with glucose test strips in the South-western Region should serve

not as a bar but as a challenge. I am convinced

that if a diabetic can perform self-monitoring of glucose accurately then it should not be

beyond our wit to devise methods that even

a sleepless house surgeon can perform

reproducibly and with clinically acceptable accuracy. Our staff can service the necessary

machines and perform the appropriate quality control tests. Having devised such

methods for the more routine tests, what joy we can have in our laboratories developing new tests, working on tissue biochemistry, and performing the more complex assays, uncluttered and unhampered by the burden of our present routine workload.

At present clinical chemistry is like a camel with an ostrich head?a beast of burden with its head firmly implanted in the mire. Is it not

time for us to stop pretending to be a collective Duke of Plaza Toro, and t? lead from the front for a change ?

K G M M Alberti

Department of Clinical Biochemistry, Royal Victoria Infirmary, Newcastle upon Tyne ?E1 4LP

Diagnostic kits and the clinical chemist

Sir,?As Dr M S Walker (12 January, p 115) mentions my name and accuses me of un

balanced statements I feel challenged to reply even though he merely picks a few nits off the

main body of my letter (15 December, p 1581). He states that manufacturing companies

have made a significant improvement in the

services that hospital laboratories can offer. Is this true ? Well, yes and no. Taking diagnostic products away from us would leave my

laboratory offering a very much reduced service in that people would be pottering away in the balance room rather than actually

doing tests. However, the example he gives of

radioimmunoassay kits do not in fact benefit our patients as we cannot justify a gamma counter and so these tests still have to be sent

away. In any case, some results are uninter

pretable just as an isolated number; we need

the big population groups and the interpreta tive expertise lurking in those teaching centres, no matter how the results are pro duced.

Dr Walker also states that significant scientific advances come from manufacturing

companies. This I do support and am con

vinced that this percentage of total advances

will increase as NHS scientists become more

and more starved of funds. One could even say that the wasting of money on "side room"

testing will make matters worse. Our hospital

spent ?2400 on stick tests last year; our total

allocated budget for the whole pathology

laboratory chemicals and gases was ?9000.

Also, of course, money spent on ward testing could mean less for companies such as Serono

producing valid and important products for a

laboratory environment.

Dr Walker's paragraph bringing in the

quality control schemes I am not certain I

completely understand, except that he would

appear to be saying that even laboratories can

fail?and this too I support. If we take a

blunder rate of about the national average of

1 % the number of tests we send out that are

wrong is 1200; on the other hand, of course, we do get 118 800 about right. But my point would be that heads of departments are

extremely aware of the virtual impossibility of

scientific accuracy: just minimum bias is our

aim?we know we are imprecise and by how

much, and we know we blunder. The whole

atmosphere of the laboratory is geared to

reduce and control these errors; clinicians

laugh as we worry over 0-3 mmol/1 out on a

glucose control test but we know that this may

This content downloaded from 91.213.220.138 on Sat, 28 Jun 2014 12:41:59 PMAll use subject to JSTOR Terms and Conditions

330 BRITISH MEDICAL JOURNAL 2 FEBRUARY 1980

be the first step on a slide to unacceptable

inaccuracy and clinically useless results.

At least Dr Walker does not deny that some

kits in the past have failed; if he does dis

believe this he can call on our laboratory and

I'll show him my files of kits we could not use

for one reason or another. If this is accepted as a fact, how do these kits get stopped when

there is a blank acceptance of manufacturers'

statements by the scientifically and technically naive? These "side room" kits will not be

tested for amount of bias, interference, im

precision, or run in parallel with a referee

technique by actual users. Even if they are, how

about continuing control ?

And so to return to my original letter and

the major question in it. Manufacturing com

panies are leading us down a path of testing outside main laboratories to bring profit to

them (Dr Walker said so). Do we tread this

path? Sometimes we may have to if the

clinician decides that the product offers real

benefit to patients ; and in these circumstances

I say "Involve the hospital biochemist even if

he has to be dragged screaming from his ivory tower." Without manufacturing companies we

would presumably still be at the wattle and

daub stage so obviously I accept advance?

but not when that advance brings detriment

to established order, as in the selling of milk

feeds to primitive societies. In my mind the

state of play for tests performed outside

laboratories is just that.

Mary Warner

Department of Clinical Pathology, Yeovil District Hospital, Yeovil, Somerset BA2I 4AT

The incidence of biochemical

emergencies

Sir,?As a result of the "industrial dispute" over emergency duty payments involving

medical laboratory scientific officers (MLSO), the Torbay District Management Team

asked that requests for emergency bio

chemical investigations outside normal work

ing hours should be generated at consultant

level. Thus instead of the request being

generated by a junior doctor and passed to an

MLSO the route became junior doctor to

clinical consultant to laboratory consultant to

MLSO. Throughout the preceding year, the

average number of requests for out-of-hours

emergency biochemical investigations had

been 66 per week. During the month of

consultant monitoring the average number of

such requests was two a week.

Is this the true incidence of biochemical

emergencies ? As Professor Joad would have

said, it all depends what you mean by "emer

gency." Gerald Manley

Department of Chemical Pathology, Torbay Hospital, Torquay, Devon

Berkson's fallacy in case-control studies

Sir,?In the review by Dr Charles du V

Florey (17 November, p 1283) of an epidemio

logical textbook designed for "the uninitiated"

I was very glad to see a reference to Berkson's

fallacy in case-control studies. The demon

stration of this potential bias was first pub lished1 with probability symbols in a journal

known to very few clinical investigators, but a simple arithmetical demonstration, submitted

to Dr Berkson before publication,2 was easily grasped by my students and clinical colleagues.

During the subsequent quarter century statisticians and epidemiologists have paid curiously little attention to the possible risk.

The frequent defence has been that it is

theoretical, never shown on real data. Appar ently it was not until 19783 that there was

published a search for confirmation, by comparison of frequencies of diseases in a

hospital with the frequencies in the parent

community; but now that this report has revealed the effect of Berkson-type biases there is no excuse for ignoring the risk. Some technical details in the report would be difficult for "the uninitiated," but any of us

when trying to evaluate the report of a case

control study can ask whether the author has

ignored Berkson's fallacy or has dismissed it

casually as of no consequence.

In order to ask that question, however, we must know what we are looking for, and simple arith metic still gives me the clearest picture. In a certain

community are three features, A, B, and X, that cause some of their possessors to be admitted to a certain hospital (or clinic). The features may be

diseases, injuries, signs, symptoms, or disabilities such as ocular refractive errors in the control group in Berkson's original paper. In a study conducted in the hospital the As found there will be the "cases," the Bs found there will be the "controls," and the Xs possess some feature that we are

considering as possibly associated with A (perhaps as a clue to cause). In the community the respective populations are 1000 As, 1000 Bs, 300 Xs in the

As, and 300 Xs in the Bs?that is, there is no closer association of X with A than with B. (For simplicity we omit the As that are also Bs.)

Ten per cent of the As are admitted to the

hospital because they are As?that is their "in herent" admission rate. The inherent admission rate for B is 50% and for X is 20%. Of course, in the study we shall be ignorant of these admission rates; we postulate them here to see the effect.

Thus, of the 700 Anot-Xs we put 10% ( =

70) into the hospital and of the 300 AXs 10% (

= 30),

leaving 270 AXs; but 20% of these ( =

54) come in because they are Xs. Total AXs in hospital

= 84. After doing likewise. with the Bs we have the

following :

X Not-X Total % Xs A 30 + 54= 84 70 154 54-5 B 150 + 30=180 350 530 340

The A ?B difference in Xs is 20-5 percentage points, although in the community there was no difference whatever. It appears as if the greater admission rate of the Bs had pushed the Xs into the As (the "cases" under study). Under the

postulated conditions the demonstration is as

cogent as, say, the theorem of Pythagoras; and in the real world the crucial condition, a difference in the admission rates of As and Bs, is probably much commoner than identity of the two rates.

Using the same procedure we can4 ring the

changes on the numbers of cases and admission rates, even showing how a real association between

X and A could be masked by the interplay of admission rates. We can demonstrate fictitiously and facetiously a greater prevalence of bunions

among bronchitics than among coronary disease

patients. More plausibly, we can call X a sex difference in admission rates, and we can show that the admission rate bias could distort the results of a prospective study of health care that is

dependent on voluntary attendance.

The Berkson admission rates bias is, of

course, only one of many biases that can

affect case-control studies?your leading article (13 October 1979, p 884) alluded to 35

of them. The production of methods of coping with them is an enormous task, but the

producers (epidemiologists and statisticians) can be guided and helped by consumers

(clinicians), who are not easily hoodwinked by

"significance" tests and "confidence" limits but are very sensitive to the multifarious

pitfalls in the collection of data.

Donald Mainland

Kent, Connecticut, USA

1 Berkson J. Biometrics Bull 1946;2:47-53. 2 Mainland D. Am Heart J 1953;45:644-54. 3 Roberts RS, Spitzer WO, Delmore T, Sackett DL. J Chron Dis 1978;31:119-28. 4 Mainland D. Elementary medical statistics. Phila delphia: W B Saunders, 1963/4:117-25.

Whooping-cough immunisation

Sir,?Even though some of the reasons given for not having children immunised against

whooping cough are manifestly absurd (19 January, p 179), it does not mean that the real reasons were absurd, still less that all mis

givings about the value and safety of im munisation are thereby irrational. Absurdities are to be found on both sides of the contro

versy which has grown out of what should have been treated as a serious, necessary, and difficult scientific inquiry. The truth lies in what is actually known, and where people agree rather than where they differ.

The contraindications to pertussis vaccination first proposed by Kong in 1953, after he had

surveyed 82 cases with cerebral complications following the use of the vaccine reported in the

world literature, were : a family history of nervous

disease, a history of convulsions in the child, allergic diseases, poor general condition, and evidence of acute infectious diseases.1 He also advised that injections should be discontinued after a severe reaction to the first or second injection.

Almost immediately, the need to exclude children with a record of allergy, nervous disease, or con vulsions was strongly challenged23; and indeed if

K?ng's criteria were strictly applied it would be difficult to attain the proportion of immune

people in the population demanded by the theory of herd immunity. Despite being scientifically questioned, however, his list of contraindications has been retained, with emphasis on convulsions, to reassure the public that the risks of immunisation can be avoided by not vaccinating individuals who can be identified as vulnerable?but with little attention to the importance of reactions to previous injections.

Agreement does exist, however, on the need to withhold immunisation during periods of high poliomyelitis prevalence. Not many people doubt that pertussis vaccine can provoke clinical polio

myelitis in infections that might otherwise have remained latent. But what about its effect on other viruses ? Can a vaccine which provokes activity in latent poliomyelitis not also provoke activity in other latent virus infections ? Bordetella pertussis is known for its unexplained properties, useful and

otherwise?enhancing antibody response to diph theria and tetanus toxoids, being the only micro

organism that approaches mycobacteria as an

adjuvant for inducing experimental allergic en

cephalitis.4 With such a repertoire is it not time to reflect on what else it might do ?

Ten years ago Pittman wrote: "It is likely that in certain cases encephalopathy following pertussis vaccination is actually a viral infection which was in the incubation stage at the time of vaccination or

was latent and provoked by the vaccine." After

referring to Baird and Borofsky's work implicating pertussis vaccine with viral and bacterial encephalo pathy as possible causes of infantile myoclonic seizures,5 and to Feldman and Schwartz's report of a possible association between such seizures and

cytomegalovirus,6 she added: "Had these infants with latent virus been vaccinated, there would have been a chance of provoking a severe or fatal reaction."7 This association between myoclonic seizures and cytomegalovirus has not been con firmed but the infection is now "the commonest known viral cause of mental retardation in in

fancy"8 and is linked with school failure, behaviour

This content downloaded from 91.213.220.138 on Sat, 28 Jun 2014 12:41:59 PMAll use subject to JSTOR Terms and Conditions