An introduction to epidemiology
Epidemiology is the study of how often diseases or health-related conditions occur in different groups of people (populations) and why.
Epidemiological methods are used provide scientific evidence to understand and guide decisions on:
How big is the problem of a disease or condition in a population?
Where is it occurring and who is it affecting?
What are the risk factors?
Can be controlled it through a public health approach?
Epidemiological information is used to implement public health control strategies to:
Prevent a disease or health condition from occurring or spreading
Provide health systems with priorities for planning and targeting essential health services
And promote the use of evidence for effective clinical care and policy development.
Case definition, incidence and prevalence are basic concepts in epidemiology which we refer to throughout this course.
It is important to accurately define and describe any disease or condition we are studying before we do any ‘counting’. This is called the case definition. A case definition is a set of standard criteria classifying whether a person has a particular disease, syndrome, or other health condition.
In diabetic eye disease, many epidemiological studies have adopted, or adapted, the case definition for classifying (or grading) diabetic retinopathy as developed by the Early Treatment Diabetic Retinopathy Study (EDTRS).
Prevalence is a measure of how many people have a disease, or a health condition (cases), in a given population at a specific time. A good analogy for this is a photograph of the population and knowing how many people in the photograph have a disease or health condition.
To calculate the prevalence of a disease or condition we divide the number of people with the disease (the cases) at the specified time by the total number of people in the population we are studying. We then multiply by 100 to get a percentage.
Prevalence data helps us understand how big the problem of a disease or condition is in a population, who is affected and – using the case definition - what level of problem it is. We can use prevalence data to make key planning decisions, for example to prioritise allocation of for health service resources.
Let’s look at an example of a prevalence calculation. Health workers at diabetes clinic X examine 300 people with type 2 diabetes. Using the EDTRS classification, they find that 112 have some diabetic retinopathy (DR) during the survey period.
To calculate the prevalence of diabetic retinopathy amongst the people with type 2 diabetes cared for by clinic X we divide 112 by 300 and multiply by 100. This equals 37%.
Incidence is a measure of how many people get a disease or health condition, in a given population during a specified period of time. It estimates the possibility, or risk, that an individual will develop a disease during that time.
Incidence uses three numbers:
- The time period being studied
- The population at risk but disease free at the start of the study
- The number of new cases with the disease or health condition identified within the population at risk
To calculate incidence, we divide the number of new cases identified (the numerator) by the at risk population (denominator) that was followed up over the specified period of time.
It is important to include only those people who are disease free at the start of an incidence study. They are then followed up over the specified time. For example in order to understand the incidence of diabetic retinopathy, it is important to select persons with diabetes but only those without any retinopathy at the start.
Incidence allows us to determine a person’s risk of being diagnosed with a disease during the specified period. We can use incidence data to see how patterns of a condition within a population, or between populations, change over time.
This example shows how incidence data helps to predict the risk of a disease complication developing. The study population is 10 000 insulin dependent people with diabetes over the age of 30. None of them has any diabetic retinopathy at the start of the study. We follow them up for a period of six years and 800 new cases of diabetic retinopathy (using the EDTRS classification) are identified in that time.
To calculate the risk, or incidence, of people with diabetes in this population developing retinopathy, we divide the number of new cases, 800, by 10 000, the population being followed up. This gives us a risk of 0.08, or 8 new cases in every 100 people with diabetes, who will develop background changes in their retina in six years.
Epidemiology provides the clues for the magnitude, distributions, and determinants of diseases affecting a population. Prevalence data is very important for planning and investing in public health interventions.
Incidence data identifies the probability that a health outcome will occur and provides clarity about the type of action that is required e.g. screening. In this course, we will use these terms to describe the epidemiology of diabetic eye diseases.
Case definition is very important for epidemiological studies. Do you think the method of examination used to identify cases (when undertaking research) is important or relevant? For example, does it matter if we use fundus photography or ophthalmoscopy to identify and classify (grade) diabetic retinopathy?
© London School of Hygiene & Tropical Medicine CC BY-NC-SA 4.0