380 EPIDEMIOLOGY
population. This provides information on the distribution
of the disease and possibly allocation of resources in its
prevention and treatment. Although the use of prevalence
and incidence was demonstrated for a chronic disease,
Parkinson’s disease, it can also be used for occupational
and environmental diseases and events. Application of
preralence incidence are illustrated in Table 2 and 3.SICKNESS SURVEYSIn some cases an estimate of the amount of sickness in the
population has been made by market-research techniques,
whereby people in the street are interrogated about their
health in the last week (or month, but the shorter period is
preferred for purposes of better accuracy). This method was
used by the British government during World War II and
was described as the “Survey of Sickness.” There are several
obvious omissions that are likely to distort the fi ndings, such
as the chronically ill who cannot get out into the street, or
those who return straight to work after short illness and are
not available to be questioned in the street. Nevertheless it
is a cheap method that may have suffi cient consistency to
validate time trends of indices based on it. For more severe
illnesses, statistics of admissions to hospitals, diagnoses,
treatment methods, length of stays, etc. may serve to supple-
ment the measurement sickness in the population. Not all
hospitals may be able to provide useful fi gures, however, nor
may their catchment areas be suffi ciently clearly defi ned.
However, with the advent of computers and development
of databases for diseases, these issues of disease are better
defi ned, especially in Westernized countries. Because of the
wide range of sickness itself, the means of dealing with it,
and variation in reporting, some form of sampling from the
various sources is likely to yield the most useful results.
However, even with attempts at standardization, ICD codes,
there is often a wide variation in the incidence and preva-
lence of disease, even in the same community. One solu-
tion to this problem, in providing an accurate estimate of
disease, is to employ the capture-recapture method (CRM).
Since the problem in determining the occurrence of sickness
and disease is at the heart of counting, the CRM has been
suggested as the most accurate method for counting (Lange
et al., 2003c). However, since this method was originally
derived for counting wildlife, it is best known as an eco-
logical- and population-biology method, and has not been
widely adopted by epidemiologists. Its recent use in count-
ing hazardous-waste sites (Lange et al., 2003a) demonstrates
this method’s versatility for counting, including in the area
of epidemiology.DISEASE REGISTERSFor a number of diseases, attempts have been made to make
and maintain lists or registers of those affected. Infectious
diseases are among the most obvious to fall into this cat-
egory, since isolation of those affected was for so long theonly effective deterrent to their spread. The advent of spe-
cifi c treatment methods, or more usefully of vaccines and
immunization, has greatly reduced their impact, except per-
haps for rarities such as Lhasa fever, or, in another category,
AIDS. The pulmonary-tuberculosis register was an important
one until treatment became readily available. Notifi cation of
the local health authority of any of the range of diseases has
often been a requirement of general practitioners, in order to
obtain early warning of an impending epidemic, and to mon-
itor the occurrence of those diseases. Some heart conditions
have formed the subjects of registers, though limited usually
by time or space. Furthermore, a number of diseases, usually
rare and often genetic, have registers or societies of affected
patients, which can be useful not only for the discussion and
possible alleviation of common problems, but also for the
purposes of research. Most recently, disease registries, or
what could be included as registries, have emerged for spe-
cifi c cohorts such as migrant agricultural workers (Zahm and
Blair, 2001) as well as for specifi c occurrences of disease
(Lange et al., 2003d).CANCER REGISTERSCancer registers, or more commonly registries, form a very
distinct and important group of disease lists. They fall into
three categories: (1) special registries concerned only with
certain sites of the disease (e.g., bone cancer or gastrointes-
tinal cancers); (2) hospital-based registries, which record all
those cases seen at a particular hospital or groups of hospitals;
and (3) population-based registries, which endeavor to collect
records of every case of cancer within a specifi c population.
The last group is the most important as a source of morbidity
data about cancer. Cancers form a special group of diseases
of exceptional importance, at least historically, which attract
a great deal of interest and research, and for which epide-
miological methods are of outstanding relevance. The date of
diagnosis of cancer can be used as the basis of morbidity rates
analogously to the mortality rates, so that from a population-
based registry, rates of morbidity by sex, age, and site can be
constructed, using the sex and age structure of the population
for appropriate denominators. Cancer registries exist now in
many parts of the world, though more in developed than in the
underdeveloped countries, and their incidence rates by sex,
age, and site have been collectively published in the succes-
sive volumes of Cancer Incidence in Five Continents , begin-
ning in 1966 and subsequently at intervals of approximately
5 years. They have formed a valuable source of comparative
data about the different patterns of cancer found geographi-
cally, and in combination with other data sources can lead
to the generalization hypotheses of etiology. To their use in
a variety of other ways, such as in the evaluation of occupa-
tional and environmental carcinogenic hazards, in the conduct
of comparative clinical trails (especially in those patients not
included), and among the sequelae of certain types of chronic
disease, we shall refer later.
The pattern of cancer displayed in relation to sex, age,
and site, whether in terms of mortality or morbidity, canC005_011_r03.indd 380C005_011_r03.indd 380 11/18/2005 10:25:42 AM11/18/2005 10:25:42 AM