Sunteți pe pagina 1din 2

STATISTICS IN MEDICINE

Statist. Med. 2001; 20:2209–2217

BOOK REVIEWS
Editor: Niels Keiding
1. J. Mark Elwood; Critical Appraisal of Epidemiological Studies and Clinical Trials. 2nd edn.
2. Christopher Jennison and Bruce W. Turnbull; Group Sequential Methods with Applications to
Clinical Trials.
3. A. C. Cullen and H. C. Frey; Probabilistic Techniques in Exposure Assessment. A Handbook for
Dealing with Variability and Uncertainty in Models and Inputs.
4. Joseph L. Gastwirth (ed.); Statistical Science in the Courtroom.
5. P. M. Fayers and D. Machin; Quality of Life. Assessment, Analysis and Interpretation.
6. David Neumann and Carole Kimmel; Human Variability in Response to Chemical Exposures.

1. CRITICAL APPRAISAL OF EPIDEMIOLOGICAL STUDIES of achieving improvement. I think the way the ar-
AND CLINICAL TRIALS. 2nd edn. J. Mark Elwood, Ox- guments are put together should make it easy for
ford University Press, Oxford, 1998. No of pages: the reader to understand that it is hard to conceive
448. Price: $29:50. ISBN 0-19-262745-7 of any other ways to provide the evidence than by
the kind of research presented in the book.
The second feature is that the book makes no
This book oBers the basic tools that are needed for distinction between, on one hand, clinical research
critical appraisal of studies providing the grounds aiming at assessing therapeutic actions and, on
for evidence-based clinical medicine and public the other hand, epidemiological research aiming
health. Although there are a number of good books at assessing either preventive actions or just pos-
available that have the same aim and scope, this sibly modiCable causes of diseases. This is a wise
particular one has several features which make strategy. There is a prevailing tendency among
it very useful for anybody with no training, or clinicians to consider public-health oriented
with only rudimentary training, in that kind of epi- epidemiology as weak, being based only on
demiological and clinical research. I suggest that ‘correlations’, as compared to both observational
it might be a suitable textbook in courses for ei- or experimental clinical research. However, often
ther medical students during the later part of their they do not realize the similarities in the basic
curriculum or for graduate students entering clin- tools and conditions of research or the obvious
ical or epidemiological research programmes. Of diBerences in opportunities that, despite the simi-
course, it may be useful in any continuing medi- larities, make it meaningless to compare the types
cal education programme for doctors who have not of research in terms of strength. On the contrary,
been engaged in this kind of research before. among public health workers there is a prevailing
The Crst feature is that the author persistently temptation to rely upon less sharp evidence than
builds up the argument that the essential core of can possibly be obtained by rigorous application
the research is its ability to produce evidence for of the tools presented in the book.
or against causal relationships. Thereby, it helps A third feature is the structure and the writing
the reader to explicitly recognize that it makes no of the book, which indicates great experience in
sense to recommend or take any action aiming at teaching in this Celd. The chapters are well deCned
improving health of the patients or of the people and in a logical sequence. Every important point is
unless there is convincing evidence that the action, clearly explained and illustrated by easily readable
either as a cause in itself or as an interference with exhibits (tables or Cgures), and well chosen exam-
the causes of ill health, does increase the likelihood ples from either clinical medicine or public health.

Copyright ? 2001 John Wiley & Sons, Ltd.


2210 BOOK REVIEWS

There is at the end of each chapter a series of appro- here, most of it being put into an appendix. It is a
priate self-test questions with the answers placed bit confusing, although of course not wrong, that
at the end of the book. There is so little mathe- stratiCed sampling or frequency matching is con-
matical statistics that the average medical student sidered equivalent to serial restricted sampling (p.
should not be scared. On the other hand, the sta- 126), and I am not sure that the recommendation
tistical methods and analysis as presented are set to disregard the stratiCcation in analysis of such
up in such a way that the students should have no data is appropriate (p. 146). In presenting the as-
diIculty in following and repeating the calcula- sumptions of the simple linear and additive multi-
tions and even understanding the basic principles variate model for a continuous dependent variable,
of the methods, of which those based upon or de- it is wrongly stated that the usual methods of cal-
rived from the Mantel–Haenszel test are chosen as culation make the assumption that this variable as
the central ones. The statistical formulae, worked such (rather than its residuals) has a normal dis-
examples and tables are put together in appendices. tribution (p. 151). Maybe the explanation of the
The last six chapters of the book present a thor- conCdence limits – ‘we can be 95 per cent conC-
ough and detailed critical appraisal – applying all dent that these limits will include the true value’
the tools just put forward and excellently summa- (p. 174) – can be debated dependent on what is
rized in Chapter 9 entitled ‘The diagnosis of cau- meant by ‘conCdent’?
sation’ – of six papers on important topics and Many of my colleagues, also among the younger
published in highly esteemed journals. generation, will strongly object to the chosen po-
I have found very little to criticize. However, sition of ‘basic physiological principles’ alongside
a few points follow, which should not be seen anecdotal experience in the bottom of the hierachy
as a reservation of my initial recommendation of of evidence relevant to human health studies (p.
the book. There is a tendency that every writer 229)! The expected opposition may be more will-
of an epidemiological textbook selects and even ing to accept the arguments if this position had
creates their own terminology, which continues to been justiCed rather than just taken. On the other
confuse the communication in the Celd. The idea hand, seeing this diminutive role of what they be-
that strength of an association, measured for ex- lieve is the ideal basis for clinical practice could
ample in relative risk, may reJect nothing but the possibly act as an eye opener.
relative frequency of the component causal fac-
tors in the population is neglected. I miss some re-
Jections about how confounding factors, which by THORKILD I. A. SHRENSEN
deCnition are part of the causal web of the outcome, Professor of Clinical Epidemiology
may actually become associated with the putative Danish Epidemiology Science Centre
causal factor, the eBect of which they confound. Institute of Preventive Medicine
In view of the role of time-to-event or counting Copenhagen University Hospital
process analysis is modern epidemiology and clin- DK 1399 Copenhagen K
ical research, there is rather little on these methods Denmark

2. GROUP SEQUENTIAL METHODS WITH APPLICATIONS classical methods for interim monitoring. Peter Ar-
TO CLINICAL TRIALS. Christopher Jennison and mitage began adapting sequential methods to ex-
Bruce W. Turnbull, CRC=Chapman & Hall, U.K., periments in human populations in the 1950s, but
2000. No. of pages: xviii + 390. Price: $ 39:00. the papers by Pocock [1] and O’Brien and Fleming
ISBN 0-849-30316-8 [2] mark the turning point when statisticians began
to think of sequential designs as more the rule than
the exception. The literature on group sequential
The use of group sequential methods in random- designs for clinical trials has grown rapidly since
ized trials is the largest change in the design and 1980. Jennison and Turnbull have been among the
analysis of clinical trials in the last 15 years. major contributors to this literature in the last 10
Even though the major properties of the sequential years, and now they have summarized the state of
probability ratio test were worked out by Abra- the art and science of group sequential designs in
ham Wald and others more than 50 years ago, this well written and comprehensive book.
the particular demands of clinical trials, espe- Group sequential designs specify early stop-
cially multi-centre trials, made it impractical to use ping boundaries for experiments in which

Copyright ? 2001 John Wiley & Sons, Ltd. Statist. Med. 2001; 20:2209–2217

S-ar putea să vă placă și