

Editorial by Milo Schield:

Jerome Cornfield (Jerry: 19121979) is almost
forgotten in the field of statistics. His
Wikipedia entry
is minimal. He isn't listed in the
RSS Timeline of leading
statisticians. I consider him one of the greatest figures in the
history of statistics. He created several measures of
association including Relative Risk and the Odds ratio.

Cornfield's greatest statistical contribution was to deduce "the minimum
effect size necessary for a potential confounder to explain an observed
association assuming the association is totally spurious." This
is one of the most important contributions of statistics to human knowledge.
(Schield, 1999)

"Cornfield's minimum effect size is as important to observational studies as
is the use of randomized assignment to experimental studies. No longer could
one refute an ostensive causal association by simply asserting that some new
factor (such as a genetic factor) might be the true cause. Now one had to
argue that the relative prevalence of this potentially confounding factor
was greater than the relative risk for the ostensive cause. The higher the
relative risk in the observed association, the stronger the argument in
favor of direct causation, and the more the burden of proof was shifted onto
those arguing against causation. While there might be many confounding
factors, only those exceeding certain necessary conditions could be
relevant." (Schield, 1999)

Cornfield's argument was critical in supporting the claim that smoking
caused cancer. See the
1964 Surgeon General's
report. See also Hill's 1965 President's address titled "The
Environment and Disease: Association or Causation?" Note that of
Hill's nine criteria, "strength" was #1.
ARTICLES ABOUT CORNFIELD:
ABSTRACTS OF ARTICLES ABOUT CORNFIELD:

2012 Bayesian clinical trials in action by
JJ1 Lee and CT Chu. Stat Med. 2012 Nov 10;31(25):295572. doi:
10.1002/sim.5404. Epub 2012 Jun 18.
Abstract: Although the frequentist paradigm has been the
predominant approach to clinical trial design since the 1940s, it has
several notable limitations. Advancements in computational algorithms
and computer hardware have greatly enhanced the alternative Bayesian
paradigm. Compared with its frequentist counterpart, the Bayesian
framework has several unique advantages, and its incorporation into
clinical trial design is occurring more frequently. Using an extensive
literature review to assess how Bayesian methods are used in clinical
trials, we find them most commonly used for dose finding, efficacy
monitoring, toxicity monitoring, diagnosis/decision making, and studying
pharmacokinetics/pharmacodynamics. The additional infrastructure
required for implementing Bayesian methods in clinical trials may
include specialized software programs to run the study design,
simulation and analysis, and webbased applications, all of which are
particularly useful for timely data entry and analysis. Trial success
requires not only the development of proper tools but also timely and
accurate execution of data entry, quality control, adaptive
randomization, and Bayesian computation. The relative merit of the
Bayesian and frequentist approaches continues to be the subject of
debate in statistics. However, more evidence can be found showing the
convergence of the two camps, at least at the practical level.
Ultimately, better clinical trial methods lead to more efficient
designs, lower sample sizes, more accurate conclusions, and better
outcomes for patients enrolled in the trials. Bayesian methods offer
attractive alternatives for better trials. More Bayesian trials should
be designed and conducted to refine the approach and demonstrate their
real benefit in action. PubMed PMID: 22711340. PMCID:
PMC3495977

2005
'Jerome Cornfield' by SW Greenhouse in Biostatistics.
"Born in 1912 in New York City; died in 1979 in Herndon, VA. Best known
for helping develop Cornfield's inequality linking the observed risk
ratio to the prevalence of the omitted variable in smoking and
nonsmoking groups."
Excerpt1: "When epidemiologists began turning their attention to
the study of chronic diseases, prospective cohort designs for finding
causes of, or risk factors for, chronic diseases were in many instances
impractical. They therefore turned to case–control or retrospective
types of strategies. A problem with these designs, assuming they are
well planned, is that they do not yield traditional estimates of
absolute risk or relative risk. Cornfield, in 1955 at the Third Berkeley
Symposium in Mathematical Statistics and Probability [4, 18], presented
a derivation which demonstrated that under a rather strong assumption
(but rather reasonable in the case of chronic diseases) the odds ratio
or cross product ratio (in a 2×2 table) is a fairly good approximation
of the relative risk. The assumption was that the incidence of the
disease under study should be small."
Excerpt2: "the question of the effect of latent, unobservable
variables. Sir Ronald Fisher, in arguing against the smoking – lung
cancer relationship, had offered an hypothesis that postulated the
existence of some constitutional factor (latent and unobservable), e.g.
genetic, that caused cancer and that was also associated with the need
to smoke. Without giving the details of his argument here, Cornfield
demonstrated that if cigarette smokers are shown to have nine times the
risk of nonsmokers of getting lung cancer, but that this elevated risk
is due, not to cigarettes, but to some latent factor X, then the
proportion of smokers having X must be larger than nine times the
proportion of nonsmokers having X. Cornfield’s conclusion was that if X
was a causative agent of this magnitude, then the relationship between
the latent factor X and the observed agent would probably have been
detected much before that of the agent and the disease. No such factor
has been found."

1982 The contributions of
Jerome Cornfield to the theory of
statistics by M. Zelen. Biometrics. 1982 Mar;38 Suppl:115.
Abstract: This paper is a review of the contributions of Jerome
Cornfield to the theory of statistics. It discusses several highlights
of his theoretical work as well as describing his philosophy relating
theory to application. The three areas discussed are: linear
programming, urn sampling and its generalizations to the analysis of
variance, and Bayesian inference. It is not widely known that Jerome
Cornfield was perhaps the first to formulate and approximately solve the
linear programming problem in 1941. His formulation was made for the
famous "Diet Problem". An early publication introduced the method of
indicator random variables in the context of urn sampling. This simple
method allowed straightforward calculations of the low order moments for
estimates arising from sampling finite populations and was later
generalized to the twoway analysis of variance. The application of the
urn sampling model to the analysis of variance served to illuminate how
one chooses proper error terms for making tests in the analysis of
variance table. Jerome Cornfield's philosophy on applications of
statistics was dominated by a Bayesian outlook. His theoretical
contributions in the past two decades were mainly concerned with the
development of Bayesian ideas and methods. A brief survey is made of his
main contributions to this area. A particularly noteworthy result was
his demonstration that for the twosample slippage problem of location,
the likelihood function under a permutation setting is uninformative for
the slippage parameter. However, the posterior distribution differs from
the prior distribution despite the fact that the likelihood is
uninformative. Source:
http://www.ncbi.nlm.nih.gov/pubmed/7046817

1982 Jerome Cornfield's contributions to
epidemiology by SW Greenhouse. Biometrics. 1982 Mar;38
Suppl:3345.
Abstract: This paper reviews the contributions Jerome Cornfield made to
epidemiologic methodology. Section 2 discusses his development of the
odds ratio obtained in a casecontrol study as an estimate of the
relative risk of the disease under study. Section 3 presents Cornfield's
introduction of the multiple logistic risk function as a smoothing
function for data classified in a multiway contingency table in order
to determine the joint effects of several risk factors on the incidence
of a disease. Section 4 gives a brief description of his work in the
analysis of contingency tables. In Section 5, there is a summary of his
views on a number of issues relating to the research, mostly
casecontrol studies, on the relationship between smoking and lung
cancer. The discussion in this section is selective and undoubtedly does
not reflect all the important things he had to say on the subject.
Finally, in Section 6, there is a discussion, based on only one of his
papers on the subject, of some very significant thoughts on intervention
studies in coronary disease. Source:
http://www.ncbi.nlm.nih.gov/pubmed/7046823

1982 Jerome Cornfield's contributions to the
conduct of clinical trials by F. Ederer. Biometrics. 1982 Mar;38
Suppl:2532.
Abstract: Jerome Cornfield's important contributions to the conduct of
clinical trials are summarized here. They include consultative advice in
the planning of many national trials, active collaboration in the
conduct of many others, discussions of the role of classical and
Bayesian methods of statistical inference in clinical trials,
recommendations on data monitoring, contributions to the analysis of
results of the University Group Diabetes Project, and efforts to assist
the planning of coronary intervention trials with quantitative
assessments of possible reductions in disease rates due to intervention
on smoking and diet. An attempt is made to evaluate the impact of
Cornfield's contributions to clinical trials. Source:
http://www.ncbi.nlm.nih.gov/pubmed/7046822
ARTICLES ON SMOKING and LUNG CANCER:
 Hill A. Bradford (1952). The Clinical
Trial. New England Journal of Medicine 247. No 4 P.
113119.
 Hill A. Bradford (1953). Observation and experiment.
New England Journal of
Medicine 248:9951001.

US Surgeon General's
Report: Smoking causes lung cancer. Associated Press Story
1964
 Hill A. Bradford (1965). The environment and disease: association or
causation? Proceedings of the Royal Society of Medicine 58:295300.
ARTICLES BY CORNFIELD:

Cornfield, J (1951). Method of estimating comparative rates from
clinical data. Applications to cancer of the lung, breast and cervix.
Journal of the National Cancer Institute 1951; 11, 126975

Cornfield J (1954). Statistical relationships and proof in
medicine. The American Statistician 8(5):1921.

Cornfield
J (1956). A statistical problem arising from retrospective studies.
Proceedings 3rd Berkeley Symposium on Mathematical Statistics 4:135–48.

Cornfield J (1959).
Principles of research. American Journal of
Mental Deficiency 64:240252. Reprint
2012 Statistics in Medicine P1
 Cornfield J, Haenszel W, Hammond EC, Lilienfeld AM, Shimkin MB,
Wynder EL (1959). Smoking and lung cancer: Recent evidence and a discussion of some questions. Journal of the National Cancer
Institute 22 (1):173203.
TOC.
Copy in 2009 Int. J. Epidemiology.
Summary: "This report reviews some of the more recent epidemiologic
and experimental findings on the relationship of tobacco smoking to lung
cancer, and discusses some criticisms directed against the conclusion
that tobacco smoking, especially cigarettes, has a causal role in the
increase in bronchogenic carcinoma. The magnitude of the excess
lungcancer risk among cigarette smokers is so great that the results
can not be interpreted as arising from an indirect association of
cigarette smoking with some other agent or characteristic, since this
hypothetical agent would have to be at least as strongly associated with
lung cancer as cigarette use; no such agent has been found or suggested.
The consistency of all the epidemiologic and experimental evidence also
supports the conclusion of a causal relationship with cigarette smoking,
while there are serious inconsistencies in reconciling the evidence with
other hypotheses which have been advanced. Unquestionably there are
areas where more research is necessary, and, of course, no single cause
accounts for all lung cancer. The information already available,
however, is sufficient for planning and activating public health
measures."
 Cornfield, Haenszel and Hammond (1960).
Some aspects of retrospective studies. Journal of Chronic
Diseases 11:523534.
 Cornfield, Gordon and Smith (1961).
Quantal response curves for experimentally uncontrolled variables.
Bulletin of the International Statistical Institute 38:
97115.
 Cornfield J (1962). Joint dependence
of risk of coronary heart disease on serum cholesterol and systolic
blood pressure: a discriminant function analysis. Federation
Proceedings 21:5861.
 Cornfield J (1966a). A Bayesian test of some classical hypotheses,
with applications to sequential clinical trials. Journal of the
American Statistical Association 61:577594.
 Cornfield J (1966b). Sequential trials, sequential analysis and the
likelihood principle. The American Statistician 20:1823.
 Cornfield J, Greenhouse SW (1967). On certain aspects of
sequential clinical trials. In Proceedings of the Fifth Berkeley
Symposium on Mathematical Statistics and Probability, Vol. 4. (eds.
Neyman and LeCam) pp. 813829.
 Cornfield J (1969). The Bayesian outlook and its applications (with
discussion). Biometrics 25:617657.
 Cornfield J (1970a). Fixed and floating sample size trials. In
Symposium on Statistical Aspects of Protocol Design. Engle RL, Jr.
(Symposium Chairman). Bethesda, Maryland: Clinical Investigation Review
Committee, Clinical Investigations Branch, National Cancer Institute,
National Institutes of Health, pp 181187,with discussion on pp 197204.
 Cornfield J (1970b). The frequency theory of probability, Bayes'
theorem, and sequential clinical trials. In Bayesian Statistics (eds. Donald L. Meyer, Raymond 0. Collier, Jr.) Itasca, Illinois:
Peacock Publishers Inc., pp 128.
 Cornfield J (1970c). Discussion by J. Cornfield, B.M. Jill,
D.V.Lindley, S. Geisser, and C.M. Mallows. In Bayesian Statistics (eds.
Donald L. Meyer, Raymond 0. Collier, Jr.) Itasca, Illinois: Peacock
Publishers Inc., pp 85125.
 Cornfield J (1971). The University Group Diabetes Program. A further
statistical analysis of the mortality findings. Journal of the
American Medical Association 217:16761687.
 Cornfield J (1974a). Statement of Dr. Jerome Cornfield, Chairman,
Department of Statistics, The George Washington University, Washington,
D.C. In Subcommittee on Monopoly (1974), pp 1077810794.
 Cornfield J (1974b). Interrogation of Holbrooke S. Seltzer, M.D. In
Subcommittee on Monopoly (1974), pp 1088910895.
 Cornfield J (1974c). Correspondence between Senator Gaylord Nelson
and Neil L. Chayet, Dr. Jerome Cornfield, Dr. Christian R. Klimt, and
Dr. Jeremiah Stamler. In Subcommittee on Monopoly (1974), pp
1150711523.
 Cornfield J (1975). A statistician's apology.
Journal of the
American Statistical Association 70:714.
 Cornfield J (1976). Recent methodological contributions to clinical
trials. American Journal of Epidemiology 104:408421.
 Cornfield J (1978). Randomization by group: a formal analysis.
American Journal of Epidemiology 108:100102.

Jerome Cornfield Papers: Historical Note. Special Collections
Department,
Iowa State University.

Jerome Cornfield's Bayesian approach to
assessing interim results in clin... http://www.jameslindlibrary.org/articles/jeromecomfieldsbayesianapp
.. 10 of 13
