Limitations of epidemiology
(go to Outline)
Every epidemiological tool has limitations. These limitations are discussed in detail for each data collection method and health outcome included in this web course. Awareness of these limitations is crucial to making the correct decisions about what to ask of epidemiologists, which studies to commission, and when and how to interpret findings.
Of course, epidemiology depends on valid data. Often in emergencies, the ability to gather data is severely restricted. This may be due to insecurity preventing survey workers from carrying out data collection or lack of resources preventing health workers from submitting surveillance data. Lack of access may also be due to difficulties in communication and transport to remote areas.
Epidemiology is also constrained by the rapid changes in the health and nutritional status of many emergency-affected populations. By the time appropriate data and collected and analysed, the conclusions and recommendations derived from these analyses may be out of date.
Another major limitation of epidemiology is in programme evaluation. Policy-makers sometimes expect that organizations delivering a certain intervention will be able to demonstrate an "impact" on health that is unequivocally traceable to that intervention alone. This is often nearly impossible to achieve without very complex and expensive studies, such as randomised trials, because many different factors can simultaneously affect a specific health outcome or indicator. For example, imagine trying to establish a causal link between an agency's distribution of hygiene kits and deaths due to diarrhoeal disease in a camp where many other factors are changing simultaneously, such as the prevalence of malnutrition, water quantity and quality, excreta disposal, individual behaviour, etc.
Finally, perhaps the most important limitation of epidemiology is that epidemiology and the data gathered by epidemiologic methods are routinely ignored. Many major decisions during humanitarian relief are not made on the basis of data or evidence; political concerns, resource limitations, personal priorities, public relations and many other non-scientific factors may result in irrational policy or programme decisions. Although perhaps not a limitation of epidemiology itself, one must keep in mind limitations in the need for epidemiologic evidence. Most interventions recommended by the Sphere Guidelines (Click here to open document).and other standard-setting publications have been demonstrated effective by good scientific studies or by years of experience. These interventions are not experimental. Their potential for impact need not be reconfirmed every time such interventions are implemented. For example, we know that measles vaccination protects against measles disease and its deadly complications. This need not be demonstrated each time measles vaccination is implemented. On the other hand, we can never assume that target beneficiaries are indeed receiving the given interventions. As a result, coverage (i.e. the proportion of people in need of an intervention who actually receive it) is what primarily needs to be documented in emergency situations.