To start your own research study, you need to have a basic understanding of epidemiology and study design. Briefly, epidemiology is the study of how diseases and health conditions are distributed in populations, and the factors that influence their occurrence. The goal is to understand patterns of health in populations of people rather than individuals.

Key Definitions

Population: The group of people being studied. This could be a specific community, patients with a certain disease or within a certain age range, or any defined group.

Exposure: The factor being investigated as a potential cause or influence on health. In observational studies, we call this factor an 'exposure' (something that naturally occurs). In experimental studies, we call it an 'intervention' (something researchers deliberately apply). These can be protective, harmful, or neutral.

Outcome: The health event or condition of interest and sometimes called the "effect" that occurs after the "cause". Examples might include death, recovery, or some other measure of health.

Building Your Research Question

When you put these three components together, you get a research question. Take the following example:

"In low-income adults over age 18, are audio telehealth visits associated with greater visit completion rate?"

• Population = Low-income adults over age 18
• Exposure = Audio telehealth visits
• Outcome = Visit completion rate

It is critical you know how to define a good research question. Vague questions are ambiguous, hard to study, and nearly impossible to publish. Here are the steps to creating a solid research question.

Defining a Strong Study Population

Clear and specific criteria - You must be able to precisely describe inclusion and exclusion criteria so another researcher could identify the same population.

Justified selection rationale - Your population definition should be based on accepted medical or scientific reasons. For instance, use standard age cutoffs (18 for adulthood, 65 for elderly) rather than unconventional boundaries like 27 years old.

Creating a Strong Exposure Variable

Objective and measurable - The exposure must be quantifiable through reliable methods rather than subjective assessment.

Appropriate timing - Exposure must precede the outcome (for causal inference) and occur within a biologically or theoretically relevant timeframe.

Variability between groups - Your study group must include people with and without the exposure so you can compare outcomes between the two groups.

Biological plausibility - The exposure should have a reasonable scientific explanation through which it could influence your outcome.

Developing a Strong Outcome Variable

Precise definition and measurement - Specify exactly what constitutes the outcome using objective, reproducible criteria.

Meaningful significance - The outcome should matter to patients, clinicians, or public health practitioners.

Adequate frequency - The outcome must occur often enough in your study population to allow for statistical analysis within practical constraints.

Practice Examples

I'll write some examples of research questions. You decide the strengths and weaknesses of each. (Answers below).

Examples of Research Questions

1. Among adults aged 40-75 with newly diagnosed hypertension (population), does adherence to a Mediterranean diet for 12 months (exposure) reduce cardiovascular events compared to standard dietary counseling (outcome)?
   
   
2. In patients with cancer (population), does treatment (exposure) make them feel better (outcome)?
   
   
3. Among people with brown hair (population), does stress (exposure) affect health (outcome)?
   
   
4. In veterans over 65 years and older (population), is noise pollution (exposure) associated with greater anxiety as scored by a standardized assessment (outcome)?
   
   

Answers to the Examples

Example 1: Excellent - This is a well-crafted research question. The population is specific with clear age boundaries and a defined health condition. The exposure (Mediterranean diet adherence) is measurable and has a defined timeframe. The outcome (cardiovascular events) is objective, clinically important, and commonly occurs in this population. The comparison to standard care makes this actionable.

Example 2: Poor - This fails on multiple criteria. While "patients with cancer" defines a population, it's very broad - cancer type, stage, and prognosis matter enormously. "Treatment" is completely unspecified - chemotherapy? surgery? radiation? The outcome "feel better" is entirely subjective and unmeasurable. There's no timeframe and no way to define clearly any of these variables.

Example 3: Poor - This is problematic on several levels. "People with brown hair" is a bizarre population choice with no biological rationale for why hair color would matter for the research question. "Stress" is vague and difficult to measure objectively. "Health" is far too broad to be a meaningful outcome. There's no plausible connection between these variables, making this question scientifically meaningless.

Example 4: Excellent - The population (veterans over 65) is well-defined and represents a specific group with shared characteristics. "Noise pollution" needs more specificity (decibel levels? duration of exposure?) but is measurable. The outcome using a "standardized assessment" for anxiety is objective and measurable. The association being tested is biologically plausible. Minor improvement would be specifying the exposure measurement more precisely.

Next Steps

Mastering research questions is one of the most important steps toward meaningful publications. If you're ready to take the next step in your research journey, sign up for the Lumono newsletter to receive exclusive research tools and mentorship guides.

‍