Disclaimer: This article is intended for my students, but may be beneficial to others as well.
Over the years, I have observed several students writing about epidemiological study designs in response to a question regarding investigation of an outbreak. This article will attempt to explain why the conduct of a formal epidemiological study is not the right answer to such questions.
Background Information:
Epidemiological study designs include various types of studies conducted on individuals or groups of people. The basic purpose of these studies is (generally) to investigate the relationship between one or more risk factor and a health condition of interest.
Epidemiological studies may be observational or experimental in nature (I am using ‘nature’ as a synonym of ‘design’ since it is easier to understand).
Observational studies are further classified into descriptive studies and analytical studies. Descriptive studies are generally used to investigate conditions that are unfamiliar. For instance, Zika virus disease was initially reported using descriptive studies that provided details of the social, temporal, geographic, demographic, and health profile of patients (time, place and person). This served to distinguish the disease from other conditions and also helped generate hypotheses regarding cause and effects. The observation of microcephaly was initially suspected to be due to maternal pesticide exposure and lead to studies investigating that possibility. Eventually, the mechanism by which the infection causes microcephaly was described and the pesticide hypothesis died a natural death.
Analytical studies go one step further than descriptive studies and try to quantify the association (or lack of it) between suspected risk factors, characteristics, or exposures and the outcome (condition) of interest. This is done through case control studies or cohort studies that yield measures of risk for exposure (odds ratio for case-control studies, and risk ratio for cohort studies). Analytical studies are much more focused and refined than descriptive studies, but due to the observational nature of such studies, they cannot fully isolate the effect of an exposure on an outcome.
Experimental studies are conducted to test a hypothesis based on evidence generated by other (observational) studies. They may be randomized (Randomized controlled trial) or non-randomized (Non-randomized controlled trial) and try to isolate the effect of exposure on outcome by controlling for characteristics that could potentially influence the outcome.
Key Messages:
Epidemiological studies are employed to investigate the relationship between suspected exposure and outcome(s) or health condition(s). Usually, the health condition(s) are important public health problems that either affect the general population or a specific subset of the population (like individuals belonging to a particular occupation, etc.).
The fundamental motivation behind the conduct of epidemiological studies is the potential identification of ways to control the health condition. (Once risk factors or important exposures are identified as associated with occurrence of disease, measures to reduce them could be initiated. If successful, this will eventually decrease the disease burden.)
Elucidation of the relationship between suspected exposure(s) and outcome(s) typically takes several studies conducted over multiple years in a variety of settings. As knowledge of this relationship may form the basis for treatment, prevention, and health policy, precision of study results is prioritised over other considerations.
When there is a disease outbreak, the first priority is to control the outbreak using the fastest means available. Precision of diagnosis is secondary to rapid control– even if a precise diagnosis is unavailable, control measures based on the clinical picture will be implemented. For instance, if there is an outbreak of acute gastroenteritis, limited effort is directed to identifying the exact pathogen responsible (only the first few cases are evaluated thoroughly, with subsequent cases assumed to have the same condition if the clinical profile is similar). Since the control measures for gastroenteritis are well known (disinfection of water, hand hygiene, oral rehydration, etc.), those are implemented without delay to rapidly end the outbreak.
During an outbreak one uses descriptive epidemiology and careful pattern recognition to rapidly decide what kind of illness is responsible for the outbreak. Based on this, control measures are initiated without delay. This is an efficient way to contain the outbreak and public anxiety.
Conducting a formal epidemiological study during an outbreak is often impractical and not feasible. Formal epidemiological studies require ethical approval and reasonable understanding of the health condition under investigation. At the beginning of an outbreak there may be little clarity about the cause- especially if it is a new or uncommon disease. Therefore, it makes sense to use common epidemiological tools and approaches to investigate the outbreak without conducting a formal study. This provides enough flexibility to yield rapid clues about the disease, as well as the nature of interventions that could help control the outbreak.
Findings from the initial epidemiological investigation could be used to conduct a subsequent epidemiological study, but during the outbreak this is not a priority.
Useful Links:
Links to articles on how to approach problem-solving questions:
communitymedicine4all 