The greatest minds of history have contemplated the aetiology, causation and treatment of disease. Hippocrates writing in 400 BC accepted that ‘the science of medicine makes use of principles which can be of real assistance … but it would not be fair to expect medicine to attempt cures that are all but impossible nor be unfailing in its remedies’ (Chadwick and Mann 1950). Implicit in this statement is that even 2500 years ago there was the realization that there was a limit to one’s knowledge of disease, not all treatments were effective in all people at all times, and those that were to be used needed to undergo some scientific process of evaluation.

It is true today, as it was in the time of Hippocrates, that the overwhelming cause of death is infectious disease, and it is only in the developed world that diseases of degeneration and ageing have become dominant. This led Robert Koch and Friedrich Loeffler to develop in 1884 a set of criteria, later known as Koch’s postulates (Box 4.1) for infectious disease aetiology (Grimes 2006).

Epidemiology is a science that is interested in the cause, occurrence, distribution, natural history and treatment of disease. Some 30 years before Koch produced his seminal work, John Snow in 1854 demonstrated these principles. During August and September 1854 London experienced a severe epidemic of cholera resulting in nearly 100 deaths. Snow found that by mapping the place of residence of the deaths that had occurred during this period, he was able to identify that the hand water pump in Broad Street was at the epicentre of the deaths and likely to be related in a causal way. His action in getting the local authorities to remove the handle on the water pump resulted in a curtailment of the epidemic and saved many lives. This event is often hailed as one of the first triumphs of modern epidemiology with its impact on public health policy (Johnson 2007).

Then, as now, Koch’s postulates were never seen as absolutes in that many infectious agents can be present in humans without causing disease, e.g. polio and HIV viruses, in violation of Koch’s first postulate. Similarly his third postulate recognizes the fact that many organisms experimentally inoculated with an infectious agent do not necessarily develop disease; the lack of disease in many of the residents of Broad Street in 1854 would support this, as not all residents exposed to infected water succumbed to cholera.

One of the founding fathers of modern medical statistics, Sir Austin Bradford Hill, in his Presidential address to the Royal Society of Medicine in 1965 addressed the issue of causation of disease from the perspective of occupational exposure; however, the tenet of his talk is just as relevant for wider discussions of causality (Hill 1965). Starting from the point of view that cause must precede effect if there is in fact a causal relationship between an agent and a disease, Bradford Hill suggested a number of factors that need to be taken into account in order to judge whether any such association was likely to be causal (Box 4.2).

Using the example of cigarette smoking and its association with lung cancer, which will be discussed in more detail later in this chapter, it is well known today that this association is causal. If we apply the criteria in Box 4.2 to this agent/disease combination we are able to see how we might gain confidence in the conclusion that there does in fact exist a causal chain at work.

The strength of an association between factors and a disease manifests itself if the disease occurs much more frequently, or exclusively, in those exposed. For example, in this case, heavy smokers have a significantly higher rate of lung cancer deaths than non-smokers.

Webster’s New World Medical Dictionary (2001) describes surveillance as ‘the ongoing systematic collection and analysis of data and the provision of information which leads to action being taken to prevent and control a disease, usually one of an infectious nature’. Worldwide, a considerable amount of time and effort goes into the collection, collation and interpretation of data on disease occurrence, by which healthcare organizations and nations can plan healthcare services and their response to emerging threats.

In most westernized democracies, surveillance systems exist to monitor infectious disease, particularly those that can be prevented by vaccination. For example, both the United Kingdom and the United States monitor the common diseases of childhood, measles, mumps and rubella, which are notifiable to public health bodies, the data from which are used to monitor disease control and vaccine effectiveness. In addition the World Health Organization (WHO) has a network of sentinel centres around the world for specific diseases. Influenza, a disease which has the potential to kill millions of people worldwide, as it did in 1918, is one example of international collaboration between states and institutions to monitor, collect, collate, identify and characterize disease in order to detect any emergent threat immediately.

Severe acute respiratory syndrome (SARS) is a respiratory disease in humans which is caused by the SARS coronavirus; between November 2002 and July 2003, 8096 individuals in 29 countries were infected with this virus and 774 deaths occurred (Vijayanand et al 2004). The initial disease, from a smouldering start in mainland China, where it had probably spread to humans from bats who are asymptomatic to the disease, either directly, or through civet cats, then spread first to Hong Kong and then across the developed world. After China, Canada and Singapore were the countries worst affected, and because of the nature of the disease, healthcare workers, in particular, in the early stages of the epidemic were very badly affected. This single event was a wake-up call to the developed world that new and emergent diseases still pose a significant threat to humankind and that eternal vigilance is required to quickly identify and contain disease. In the years following the SARS epidemic, there has been a realization that a greater threat to humankind may emerge from the highly pathogenic strain of avian influenza, H5N1, that has, since it first appeared in 1997, devastated the world’s bird populations and has infected over 300 people up to 2008, with a mortality rate of over 50%. Worldwide surveillance systems are now in place across the world linked into major laboratories in countries such as the UK, USA and Australia so that influenza and severe respiratory illness can be quickly identified and preventative measures, such as antiviral drugs in the case of influenza, can be rapidly deployed.

In the developing world the old killers of malaria, TB and the newer and more deadly, threat, HIV, are devastating the populations of sub-Saharan Africa. In these countries disease surveillance is less well developed, or even non-existent as in countries such as Zimbabwe, Iraq and Afghanistan.

In the previous section we looked at the role of surveillance in the ongoing monitoring and detection of infectious disease; in addition to this, national governments also collect vital statistics on births, deaths and disease occurrence within their borders. Surveillance information is often combined with population data, obtained from national census, in the case of the UK once every 10 years, to create a powerful epidemiological database. Census data provide information on the health, wealth and well-being of a nation’s citizens and also demonstrates ongoing trends in births and deaths, areas of increased mortality to be identified and the general slow change in disease importance to be monitored. It is beyond the scope of this chapter to cover, in detail, the use of routine data for epidemiological purposes; however, in order for the reader to gain some knowledge for further self-directed learning, a number of terms will be defined.

The health of a nation can be measured in a number of ways, which broadly fall into two distinct categories. One relates to deaths in the population, i.e. mortality, and the other to the amount of illness present, morbidity, and these in turn give rise to more complex measures such as, life expectancy at birth and quality-adjusted life years (QALYs), respectively, which will not be covered here.

While knowing the number of deaths that occur is in itself useful, additional details such as cause of death, age, sex, date and place of death, place of residence and occupation (or last employed position) all add much more to our ability to turn routine data into information with which to plan and target services.

Box 4.3 presents an example of how data on the number of deaths occurring in a population can be combined with data on age and population demographics to produce much more meaningful measures that may be used to compare the health of a population of interest with another, for example the health of one region of the UK with the health of the UK as a whole, as measured by deaths occurring.