Epidemiology is about disease in populations
- PDF / 374,127 Bytes
- 3 Pages / 595.276 x 790.866 pts Page_size
- 56 Downloads / 176 Views
COMMENTARY
Epidemiology is about disease in populations Anders Ahlbom1,2 Received: 13 November 2020 / Accepted: 16 November 2020 © The Author(s) 2020
Dekkers and Mulder argue convincingly that it is not possible to predict with certainty whether a given person will develop a disease or not during a specified period of time [1]. This may serve as a reminder that epidemiology is about disease in populations, rather than in individuals or patients. This editorial has to credos: First, the intrinsic core of epidemiology is that it relates cases of disease to a source population observed during a period of time, and second, epidemiology entails both subject matter knowledge and methods to generate such knowledge. Basic descriptive epidemiological measures are taught at the beginning of every introductory course in epidemiology. These measures appear trivial on the screen in the classroom and can be taught by any teacher in 15 minutes, but it may take longer to learn them and mismatches between numerator and denominator appear again and again. This has of course been discussed extensively with respect to casecontrol studies but is a frequent issue also in other contexts. The population connection was rather obvious in the early days when textbooks were organized in chapters addressing time, place, and person and when the first schools of public health were opened, and so even in John Snow’s and James Lind’s days [2]. However, when establishment of disease causation started to become a more explicit study aim, and in particular when long latency periods and multiple causes required new study designs to be developed, the population dimension of epidemiological studies became less apparent. Indeed, it is customary to report study results in terms of only relative risks with no mentioning of the basic rates and, e.g., the case-control study does not even provide rates. Yet, although not immediately visible, the population dimension remains in study designs such as the case-control study,
* Anders Ahlbom [email protected] 1
Institute of Environmental Medicine, Karolinska Institutet, Box 210, 171 77 Stockholm, Sweden
Center for Work and Environmental Health, Stockholm Region, Stockholm, Sweden
2
family based designs, and other designs where the aim is to draw conclusions based on disease occurrence in populations [3]. The advent of the Covid-19 pandemic has not only promoted epidemiology and incorporated the word in everyday language, it has also provided numerous examples of the difficulties that may be attached to the seemingly trivial concept of identifying cases of disease and an underlying population properly. The purpose of this text is not to criticize individual studies but to illustrate problems, so reference to specific studies are not given: Any study comparing Covid-19 incidence across populations, e.g., defined by socio-economic status, must consider that differences in testing frequency will affect results. Likewise, if one plans to use a database of verified Covid-19 cases to look at risk factor
Data Loading...
