Get Adobe Flash player

Main Menu

(8) Epidemiology

 MERS Virus Outbreak: Everything You Need to Know

Middle East respiratory syndrome coronavirus


The boom in genetic case-control studies

Genes inherited from the parents are fixed characteristics of a living organism. They cannot be altered by the occurrence of a disease and they are not subject to recollection errors, unlike exposures ascertained by questioning cases and controls. For this reason, genes represent an ideal exposure to be measured accurately in case-control studies. Large series of cases, uniformly diagnosed, can be assembled from many clinical centres, providing adequate numbers for detecting associations between gene variants and disease. With the availability of techniques that permit testing a million `single nucleotide polymorphisms' (SNPs), the gene variants, the current trend is to first throw the net wide and explore SNPs distributed over all 23 chromosomes. These studies are labeled `GWAS', or Genome Wide Association Studies, and after the first phase are followed by confirmatory phases to check that the associations found in the first phase are not false positive results arising simply by chance. Several GWAS studies are in progress and many more are starting. The first results of one large study of 14,000 cases of seven common diseases and 3,000 controls has identified more than 20 associations, involving a mental disorder, coronary artery disease, type 1 and type 2 diabetes, an intestinal inflammatory disease, and rheumatoid arthritis.

Confirmed associations open the way to the investigation of the physiological mechanisms leading from the gene variant to the disease, a task beyond the scope of ordinary case-control studies. In principle, blood could be taken from cases and controls and biochemical studies carried out to probe these mechanisms. However, the presence of the disease makes the meaning of any physiological finding questionable: would it really represent a step leading to the disease or instead be a consequence of the disease itself? Only case-control studies conducted within cohorts where blood was collected and stored before the disease occurred, like the EPIC or the British Biobank cohorts, are free of this problem. Laboratory studies on experimental animals and on cells, including fresh white blood cells from human volunteers, complement these epidemiology-based studies on the physiological paths linking genes to disease. Understanding these paths also helps to clarify the role of the external factors at play. Drugs and other means (e.g. changes in diet) that can interfere with the paths and prevent disease development are the ultimate objective of this research.

Space, time, and individuals

Intervention studies, randomized or non-randomized, and observational studies (cohort and case-control) form the core of epidemiology. As they aim to test hypotheses about causal relationships between exposures and effects, they are often collectively called analytical studies (cross-sectional and correlation studies, of which more later on in this chapter, also belong to the group). Analytical studies are generally both preceded and followed by descriptive studies of how health and disease, as measured by rates of deaths,

new cases of disease, or hospital admissions, are distributed in space by geographical area, in time by week or month or year, and in categories of people of different age, gender, and socio-economic status. Observing a disease distribution in space, time, and categories of people provides useful indications of which factors need to be explored in depth through analytical studies as possible determinants of the distribution. Visually friendly tools, like graphs and maps (sometimes collected in atlases), convey a pictorial view of disease burden and evolution. They facilitate the examination of descriptive data and the generation of causal hypotheses.

Cholera made headlines in 2008 when a lethal epidemic hit Zimbabwe, but in the 19th century it had often ravaged Europe, producing waves of high mortality, particularly among the poorest sections of the population living in squalid conditions. From foci in India, it spread to Europe, an early example (1832) of a health map, illustrates its progression westward.

Today, cholera has practically disappeared from Europe while cardiovascular diseases head the league of causes of mortality. In particular, death rates from heart attacks, tend to be markedly higher in northeast than in southwest Europe.

Sixty years ago, the marked difference in the occurrence of heart attacks between southern Europe and other regions, especially the United States, prompted American investigators to develop a comparative cohort study in seven countries with very different rates (Finland, Greece, Italy, Japan, the Netherlands, the United States, Yugoslavia). Together with other cohorts, especially the Framingham study, this `Seven Country study', follow-up of which has continued until recently, has provided essential information to establish high cholesterol, tobacco smoking, and high blood pressure as three major determinants of heart attacks.

Diseases vary both in the short and long term. Depicts the rise and fall of an outbreak of a viral disease, mumps, communicable from person to person through airborne droplets or saliva: it was a small (37 cases), self-limiting, and non-lethal outbreak localized particularly in a school.

Epidemic curves are not only useful as visual summaries of past epidemics. When constructed day by day or week by week while an epidemic is in progress, they allow, combined with information on the mechanism of transmission of the disease, the building of mathematical models, deterministic or probabilistic, predicting the likely evolution of the disease. Descriptive studies not only stimulate and guide the development of analytical epidemiological studies but are also an ultimate check that the results of the latter `make sense'. If tobacco smoking is an important cause of lung cancer, the pattern of lung cancer rates by geographical area and over a period of years should reflect, or at least be compatible with, the pattern of tobacco consumption, as has in fact been found again and again in many countries. And if effective measures of prevention have been taken, for instance reducing tobacco consumption or removing a pollutant or treating a disease, they should be reflected in a change in the rates of the disease.

Preventive measures and improved treatment in fact show up in the halving of the risk of

heart attacks between the late 1970s and the late 1990s; at the same time, the persistent difference between higher and lower social classes is a stimulus for new research.

Sources of data

The description of health and disease in populations relies on data that are collected on an ongoing and systematic basis or through special enquiries. In principle, all countries in the world (192 are members of the World Health Organization) have a system of records of basic life events: births, deaths, and causes of deaths. The organization and, more importantly, the coverage and quality of the data collected are very variable. Only 20% of the world's population living in 75 countries, mostly economically developed, is in fact covered by cause-of-death statistics judged (for instance in respect to accuracy, completeness, and other requirements) as high or medium-high quality. For as many as 25% of the world population, WHO does not receive any data on cause of death. Several years, from three to ten or more, may be needed for cause of death to be reported to WHO and made available for international comparisons. As to births, 36% go unregistered, with vast differences between countries, from 2% in industrialized countries to 71% in the very least developed. In the decade 1995-2004, only 30% of the six billion world population lived in countries with complete registration of births and deaths. This percentage had not changed much from the 27% figure of the period 1965-74 (when the world population was less than 4 billion).

Morbidity statistics are disease-oriented, and range from hospital discharge records, in principle available wherever a hospital exists but in practice of vastly variable quality, to registries intended to cover all cases of a disease, or of a group of diseases, occurring within a population. In selected and limited areas of a number of developed countries, registries are operational for malformations, myocardial infarctions, diabetes, stroke, and other conditions. Cancer or, better, cancers (because the heading embraces several hundred different diseases), is the condition best covered. Cancer registries started in the 1930s in North America and the first nationwide registry was established in Denmark in 1942. By the end of the last century, there were close to 200 good-quality cancer registries, mostly local or regional, in 57 countries.

Because of the danger of contagion, several infectious diseases have been the object of compulsory and rapid notification at national level since the late 19th century. Today, cases of diseases such as smallpox, SARS, poliomyelitis, and cholera fall within the larger scope of the `International Health Regulations' and also require notification to the World Health Organization. Surveillance systems of communicable diseases have evolved in promptness and coverage, yet even the best systems based on doctors' diagnoses can hardly report a rising epidemic in less than one or two weeks. Recently, and somewhat surprisingly, just counting an increasing number of daily queries on influenza in Google has proved capable of detecting the rise of the disease in just one or two days. This simple and cheap approach might also work for other epidemic diseases in areas with a large population of Web users. An accurate estimate of the propagation and severity of a potentially fatal disease demands, however, a complete enumeration both of cases and of deaths. When many mild cases are not recorded, as it may be for the A(H1/N1) influenza, the extent and speed of the epidemic is underestimated and the fatality rate, i.e. the ratio of deaths to cases in a time interval, is overestimated (as deaths are less likely to go unrecorded).



googleplus sm


ar bg ca zh-chs zh-cht cs da nl en et fi fr de el ht he hi hu id it ja ko lv lt no pl pt ro ru sk sl es sv th tr uk

Verse of the Day

Global Map