Although communicable diseases have declined in industrialised societies, outbreaks of disease such as influenza, gastroenteritis, and hepatitis are still important. During the 1957-8 influenza epidemic, for example, the death rate in England and Wales was 1 per 1000 population above the seasonal average; an estimated 12 million people developed the disease; and the workload of general practitioners increased fivefold. From time to time new communicable diseases such as Lassa fever, legionnaires' disease, and, most recently, AIDS appear in epidemic form.
In outbreaks of common communicable diseases such as gastroenteritis and hepatitis appropriate investigations must be initiated. The routine for these investigations is also the model for studying non-infectious disease epidemics.
At the outset it is necessary to verify the diagnosis. Three patients with halothane induced hepatitis were referred to one university hospital. Investigation of an outbreak of infectious hepatitis was begun, presumably because the clustering of cases gave an impression of infectivity and unduly influenced the physician's diagnosis. With some diseases - Lassa fever, for example - urgency demands that immediate action is taken on the basis of a clinical diagnosis alone. But for most diseases there is less urgency and the doctor should remember that clusters of cases of uncommon noninfectious diseases sometimes occur in one place within a short time simply by chance.
From time to time errors in collecting, handling, or processing laboratory specimens may cause "pseudo epidemics". The Centers for Disease Control in Atlanta, Georgia, USA, have reported several such pseudo epidemics. In one, an apparent outbreak of typhoid occurred when specimen contamination produced blood cultures positive for Salmonella typhi in six patients.
If a disease is endemic (habitually present in a community) it is necessary to estimate its previous frequency and thereby confirm an increase in incidence above the normal endemic level. Pseudo epidemics may arise from sudden increases in doctors' or patients' awareness of a disease, or from changes in the organisation of a doctor's practice. When the endemic level has been defined from incidences over previous weeks, months, or years the rate of increase of incidence above this level may indicate whether the epidemic is contagious or has arisen from a point source. Contagious epidemics emerge gradually whereas point source epidemics, such as occur when many people are exposed more or less simultaneously to a source of pathogenic organisms, arise abruptly.
To build up a description of an epidemic it will be necessary to take case histories to identify the characteristics of the patients . Patients whose diseases are notified or otherwise recorded are often only a proportion of those with the disease, and additional cases must be sought. Thereafter it is necessary to define the population at risk , and relate the cases to this. This will require mapping of the geographical extent of the epidemic.
Defining the population at risk enables the extent and severity of the epidemic to be expressed in terms of attack rates-which may be given either as crude rates, relating the numbers of cases to the total population, or as age and sex specific rates. It may be possible to identify an experience that is common to people affected by the disease but not shared by those not affected; and, from this, a hypothesis about the source and spread of the epidemic may be formulated.
There are several examples of large scale epidemics due to chemical contaminants. Outbreaks of mercury poisoning, with resulting deaths and permanent neurological disability, have been reported from non-industrial countries as a result of ingestion of flour and wheat seed treated with methyl and ethyl mercury compounds. In 1981 in Spain 20 000 people were affected by a new disease, named the "toxic allergic syndrome", the most striking feature of which was a pneumonitis. During the first four months of the epidemic more than 100 people died and 13 000 were treated in hospital. Epidemiological and clinical investigation showed that the cause was ingestion of olive oil adulterated with contaminated rape seed oil.
Widespread environmental contamination is a new agent of epidemic disease. During the 1980s, 26 epidemics of hospital admission for asthma occurred in the city of Barcelona. Epidemiological investigations eventually established that the cause was allergy to soya bean dust released into the atmosphere when cargoes of beans were unloaded in the harbour.
Increasing recognition of environmental hazards from substances introduced by man into his environment, as a result of the application of new technology, has led to a demand for large scale monitoring systems based on automated record linkage. Whether or not such systems come into Operation, clinicians' awareness of changes in disease frequency or of the appearance of clusters of unusual cases will continue to be crucial to the early detection of new epidemics. Clinicians have a special responsibility in the early detection of epidemics caused by medication. The rise in mortality during the 1960s among asthmatic patients who used pressurised aerosols, and the Occurrence of corneal damage, rashes, and various other adverse effects of practolol are two of many examples of epidemics resulting from prescription of new drugs.
New diseases continue to appear. The name legionnaires' disease was given to an outbreak of pneumonia at a convention of American Legionnaires in Philadelphia, Pennsylvania, USA, in 1976. There were 29 deaths. This stimulated an intensive epidemiological investigation whose successful outcome was the identification of a Gram negative bacillus as the causative agent.
From 1981 to 1983 some 2000 cases of AIDS were reported in the USA. The ratio of men to women was 15 to 1, and the epidemiology suggested an infectious agent usually transmitted by homosexual intercourse. AIDS seemed to be a new disease. Subsequent studies, however, showed it to be endemic in central Africa but with a sex ratio of around 1 to 1, which suggested spread by heterosexual contact. Investigations of this kind are a dramatic application of epidemiology.