Week Seven: Infectious Disease Research: Study Designs & Measures of Association

What is a study design, and why is it important in infectious disease research?

A study design is the plan researchers use to collect and analyze data to answer a research question. It determines how participants are selected, how exposure and outcome are measured, which disease-frequency and association measures can be calculated, and what conclusions the findings can support.

What is an exposure in an epidemiologic study, and must an exposure be harmful?

An exposure is the factor being studied in relation to a health outcome. Examples include antibiotic use, smoking, vaccination, injection drug use, or an environmental factor. Exposure does not necessarily mean something harmful; it can be a treatment, preventive intervention, behavior, or other characteristic.

What is an outcome in an epidemiologic study?

An outcome is the health event or condition researchers are studying. Examples include C. difficile infection, influenza, hepatitis C infection, surgical site infection, or death. Researchers examine whether the outcome differs between exposure or intervention groups.

What is the difference between descriptive and analytic epidemiologic studies?

Descriptive studies describe how often or how commonly a health event occurs and may examine patterns by age, place, or time. Analytic studies examine whether an exposure is associated with an outcome, usually by comparing groups. Descriptive studies focus on describing patterns, while analytic studies focus on relationships between variables.

How can you identify an analytic study from its research question?

Look for a question asking whether an exposure is associated with an outcome, such as whether broad-spectrum antibiotic use is associated with C. difficile infection. The study generally compares groups with different exposure statuses or intervention conditions.

What is the difference between an observational study and an experimental study?

In an observational study, researchers observe naturally occurring exposures and outcomes without assigning the exposure. In an experimental study, researchers assign an intervention to participants. Cohort, case-control, and cross-sectional studies are observational designs; a randomized controlled trial is experimental.

What are the four major study designs covered in Week 7, and how does each begin?

A cross-sectional study begins with a population sample and measures current exposure and outcome at approximately the same time. A cohort study begins by identifying exposed and unexposed groups and follows them toward outcomes. A case-control study begins by selecting people with and without the outcome and comparing prior exposures. A randomized controlled trial randomly assigns an intervention and follows outcomes.

What is a cross-sectional study, and why is it often described as a snapshot?

A cross-sectional study samples a population and measures exposure and current outcome status at approximately the same time. It provides a snapshot of a population rather than following participants over time to observe new cases.

What can a cross-sectional study measure, and what is its main limitation?

A cross-sectional study can measure prevalence and compare prevalence between exposed and unexposed groups. Its main limitation is that the timing of exposure and outcome may be unclear, so researchers may not know whether the exposure occurred before the outcome.

Researchers survey 820 adults about recent insect bites and whether they currently have signs or symptoms of lymphangitis. Which study design is this, and why?

This is a cross-sectional study because researchers survey a population and measure exposure information and current outcome status at approximately the same time. Participants are not followed forward to observe new cases.

What is the difference between prevalence ratio and prevalence difference?

Prevalence ratio is a relative measure that divides prevalence in the exposed group by prevalence in the unexposed group. It describes how many times as high or low prevalence is. Prevalence difference is an absolute measure that subtracts prevalence in the unexposed group from prevalence in the exposed group and describes the difference in percentage points.

How do you calculate prevalence in an exposed group and an unexposed group?

Calculate prevalence separately in each group by dividing the number of people with the existing outcome by the total number of people in that group. Prevalence = existing cases divided by total group size. Multiply by 100 to express the result as a percentage.

A cross-sectional study finds hepatitis C in 40 of 100 youth reporting injection drug use and 50 of 400 youth not reporting injection drug use. What is the prevalence in each group?

Exposed group prevalence = 40/100 = 0.40 or 40%. Unexposed group prevalence = 50/400 = 0.125 or 12.5%. These values describe the proportion with existing hepatitis C in each group.

Using the hepatitis C example, how do you calculate and interpret the prevalence ratio?

Prevalence ratio = prevalence in exposed group divided by prevalence in unexposed group = 0.40/0.125 = 3.2. Hepatitis C prevalence was 3.2 times as high among youth reporting injection drug use as among youth not reporting injection drug use.

Using the hepatitis C example, how do you calculate and interpret the prevalence difference?

Prevalence difference = prevalence in exposed group minus prevalence in unexposed group = 0.40 - 0.125 = 0.275. Hepatitis C prevalence was 27.5 percentage points higher in the group reporting injection drug use. This is an absolute difference, not a relative ratio.

What is a cohort study, and what is the direction of inquiry?

A cohort study is an observational study that begins by identifying exposed and unexposed groups and then determines who develops the outcome. The direction of inquiry is exposure to outcome. Data may be collected prospectively or obtained from existing records.

What is the difference between a prospective cohort study and a retrospective cohort study?

In a prospective cohort study, researchers identify exposure groups and collect follow-up data moving forward to determine who develops the outcome. In a retrospective cohort study, researchers use existing records to identify exposure groups and determine outcomes that occurred during a prior period. Both begin conceptually with exposure status and examine outcomes.

Researchers enroll adults who report injection drug use and adults who do not, then follow both groups for five years to determine who develops hepatitis B infection. Which study design is this?

This is a cohort study because participants are grouped according to an existing exposure status and followed over time to identify new outcomes. Researchers do not randomly assign injection drug use.

What is incidence risk, and why is it commonly measured in cohort studies?

Incidence risk is the probability that initially disease-free individuals at risk develop a new outcome during a specified period. Cohort studies can calculate it because they identify groups at risk and observe which participants develop new outcomes during follow-up.

In a cohort study, 20 of 100 patients with diabetes develop a surgical site infection, compared with 10 of 200 patients without diabetes over 30 days. What is the risk in each group?

Risk among patients with diabetes = 20/100 = 0.20 or 20%. Risk among patients without diabetes = 10/200 = 0.05 or 5%. These are the 30-day risks of developing a new surgical site infection in each group.

What is a risk ratio, and how is it calculated?

A risk ratio compares the incidence risks of two groups. Formula: risk ratio = risk in exposed group divided by risk in unexposed group. A ratio greater than 1 indicates higher risk in the exposed group, a ratio of 1 indicates equal risks, and a ratio below 1 indicates lower risk in the exposed group.

Using the diabetes and surgical site infection example, what is the risk ratio, and how should it be interpreted?

Risk ratio = 0.20/0.05 = 4. The 30-day risk of surgical site infection was four times as high among patients with diabetes as among patients without diabetes.

What is a risk difference, and how is it calculated?

Risk difference is an absolute comparison of the risks in two groups. Formula: risk difference = risk in exposed group minus risk in unexposed group. A positive value means the exposed group has a higher risk, a negative value means it has a lower risk, and zero means equal risks.

Using the diabetes and surgical site infection example, what is the risk difference, and how should it be interpreted?

Risk difference = 0.20 - 0.05 = 0.15. The 30-day risk of surgical site infection was 15 percentage points higher among patients with diabetes. This is an absolute difference, whereas the risk ratio of 4 is a relative comparison.

During one month, 30 of 150 patients exposed to broad-spectrum antibiotics develop C. difficile infection, compared with 10 of 200 unexposed patients. What is the risk ratio?

Risk among exposed patients = 30/150 = 0.20 or 20%. Risk among unexposed patients = 10/200 = 0.05 or 5%. Risk ratio = 0.20/0.05 = 4. The risk was four times as high in the antibiotic-exposed group.

What are the strengths of a cohort study?

Cohort studies can establish whether exposure preceded the outcome, examine multiple outcomes from one exposure, and study uncommon exposures. They can calculate incidence risk or incidence rate when the necessary data are available.

What are the limitations of a cohort study?

Cohort studies may require substantial time and resources, loss to follow-up can affect results, and they may be inefficient for very rare outcomes because a large population or long follow-up may be needed to observe enough cases.

What is a case-control study, and how are cases and controls selected?

A case-control study is an observational study that begins with outcome status. Cases have the outcome, while controls do not have the outcome. Researchers then compare prior exposure histories between the two groups to investigate possible associations.

What is the direction of inquiry in a case-control study?

The study begins with the outcome and looks backward to compare previous exposures. The direction is outcome status to prior exposure, unlike a cohort study, which begins with exposure and follows toward the outcome.

Why is a case-control study especially useful for rare diseases or outcomes?

Researchers can deliberately select people who already have the rare outcome as cases and people without the outcome as controls. This is often more efficient than following a very large population over time to observe a small number of rare cases.

What are the limitations of a case-control study?

Risk and incidence rate usually cannot be calculated directly because participants are selected according to outcome status. Prior exposure information may be incomplete or inaccurate, and selecting appropriate controls is critical. These issues can affect the validity of the observed association.

Researchers select 30 attendees with Salmonella infection as cases and 40 attendees without infection as controls, then ask about previous raw egg consumption. What is the study design?

This is a case-control study because participants are selected according to whether they have Salmonella infection, and researchers compare prior raw egg exposure between cases and controls.

In a case-control study, 24 of 30 cases ate raw eggs and 6 did not; 20 of 40 controls ate raw eggs and 20 did not. Can incidence risk be calculated directly from these data?

No. Researchers selected participants according to outcome status, and the numbers of cases and controls were determined by the study design. The sample does not show how many people in the source population developed Salmonella infection, so incidence risk and risk ratio cannot be calculated directly.

What is an odds ratio, and why is it commonly used in case-control studies?

An odds ratio compares the odds of exposure among cases with the odds of exposure among controls. It is commonly used in case-control studies because researchers select participants based on outcome status and usually cannot calculate incidence risk or risk ratio directly.

How do you calculate an odds ratio from a 2-by-2 case-control table?

For a table with cases exposed = a, cases unexposed = b, controls exposed = c, and controls unexposed = d, the odds ratio is (a × d)/(b × c). It compares exposure odds among cases with exposure odds among controls.

Using the raw egg example with 24 exposed cases, 6 unexposed cases, 20 exposed controls, and 20 unexposed controls, what is the odds ratio?

Odds ratio = (24 × 20)/(6 × 20) = 480/120 = 4. The odds of prior raw egg exposure were four times as high among cases as among controls.

A second raw egg study includes 120 cases who ate raw eggs, 30 cases who did not, 90 controls who ate raw eggs, and 210 controls who did not. What is the odds ratio?

Odds ratio = (120 × 210)/(30 × 90) = 25,200/2,700 ≈ 9.33. The odds of raw egg exposure were approximately 9.33 times as high among Salmonella cases as among controls in this sample.

How do you interpret an odds ratio greater than 1, equal to 1, or less than 1 in a case-control study?

An odds ratio greater than 1 means the odds of exposure are higher among cases than controls. An odds ratio equal to 1 means the exposure odds are similar. An odds ratio less than 1 means the exposure odds are lower among cases than controls.

What is a randomized controlled trial (RCT), and what makes it an experimental study?

A randomized controlled trial is an experimental study in which investigators randomly assign participants to intervention groups, include a comparison or control group, and follow participants to compare outcomes. It is experimental because researchers assign the intervention rather than simply observe naturally occurring exposure.

What do randomized and controlled mean in an RCT?

Randomized means participants are assigned to groups by chance. Controlled means outcomes in the intervention group are compared with outcomes in another group, such as a placebo group or a group receiving standard treatment.

Why does random assignment matter in a randomized controlled trial?

Random assignment helps balance known and unknown baseline characteristics between groups on average, reducing systematic baseline differences that could distort comparisons. Groups can still differ by chance, so randomization does not guarantee perfectly identical groups.

Researchers enroll adults with bacterial pneumonia and randomly assign them to a new antibiotic or current standard treatment, then follow them for one month. Which study design is this?

This is a randomized controlled trial because the researchers randomly assign the treatment and follow the groups to compare outcomes.

What is the main difference between an RCT and a cohort study?

Both can follow groups over time and measure incidence. In an RCT, researchers randomly assign the intervention. In an observational cohort study, researchers identify existing exposure groups and do not assign the exposure.

In a randomized trial, the vaccine group has 15 new infections during 1,000 person-months and the control group has 30 new infections during 1,000 person-months. What are the incidence rates?

Vaccine group rate = 15/1,000 = 15 infections per 1,000 person-months. Control group rate = 30/1,000 = 30 infections per 1,000 person-months.

Using the vaccine trial data, what is the rate ratio, and how should it be interpreted?

Rate ratio = intervention rate divided by control rate = 15/30 = 0.5. The infection rate in the vaccine group was half the rate in the control group during the observed period.

What is a rate difference, and how do you interpret a negative rate difference?

Rate difference = intervention or exposed group rate minus comparison group rate. A negative value means the first group's rate was lower. In the vaccine example, 15 - 30 = -15 infections per 1,000 person-months, meaning the vaccine group had 15 fewer infections per 1,000 person-months.

The treatment group has 5 infections per 1,000 person-months and the control group has 10 infections per 1,000 person-months. What is the rate ratio?

Rate ratio = 5/10 = 0.5. The treatment group's infection rate was half the control group's rate during the measured period.

What is the difference between a risk ratio, rate ratio, and prevalence ratio?

A risk ratio compares incidence risks over a specified period. A rate ratio compares incidence rates using person-time denominators. A prevalence ratio compares the proportion of people with an existing outcome at a particular time or during a specified period. The appropriate measure depends on the study design and available data.

What is the difference between a ratio measure and a difference measure?

A ratio uses division to compare groups relatively and answers how many times as high or low an outcome is. A difference uses subtraction to compare groups absolutely and answers how much higher or lower the outcome is. Risk ratios, rate ratios, and prevalence ratios are relative measures; risk differences, rate differences, and prevalence differences are absolute measures.

How do you interpret a ratio of 2.0, 1.0, and 0.5 for risk, rate, or prevalence?

A ratio of 2.0 means the measure in the first group is twice as high as in the comparison group. A ratio of 1.0 means the measures are equal. A ratio of 0.5 means the measure in the first group is half as high as in the comparison group.

How should a risk difference or prevalence difference of 0.10 be interpreted?

A difference of 0.10 represents an absolute difference of 10 percentage points when comparing proportions. It does not mean the outcome is 10% higher in relative terms; relative comparisons require a ratio.

How do you choose the correct comparison measure for a study?

First identify the study design, then identify the disease-frequency data available. If the data provide risk, use a risk ratio or risk difference. If they provide incidence rates, use a rate ratio or rate difference. If they provide prevalence, use a prevalence ratio or prevalence difference. For case-control sampling, an odds ratio is typically used because risk and rate usually cannot be calculated directly.

What study design and association measure fit a study of 150 antibiotic-exposed patients and 200 unexposed patients who are all infection-free at baseline and followed for one month for C. difficile infection?

This is a cohort study because researchers identify exposed and unexposed groups and follow them for new outcomes. Incidence risk can be calculated for each group, and a risk ratio or risk difference can compare them. If the study measures person-time and calculates incidence rates, a rate ratio or rate difference may also be appropriate.

Why does association not prove causation?

An observed association may reflect a true effect, confounding, bias, chance, or unclear timing between exposure and outcome. A relationship between an exposure and an outcome does not by itself prove that the exposure caused the outcome.

What is confounding, and how can it affect study findings?

Confounding occurs when another factor related to both the exposure and the outcome distorts the observed association between them. The association may appear stronger, weaker, or different from the actual relationship because the groups differ in an additional relevant factor.

What is bias in epidemiologic research?

Bias is systematic error in participant selection, measurement, information collection, or another part of the research process. Unlike random variation, bias can consistently distort the results away from the truth.

How can unclear timing limit conclusions about an exposure and an outcome?

If researchers cannot determine whether exposure occurred before the outcome, they cannot confidently establish the temporal order required for a causal explanation. This is an important limitation of cross-sectional studies.

What participant protections are important in infectious disease research?

Research should receive appropriate institutional review board (IRB) review, provide clear information about purpose, procedures, risks, and benefits, obtain voluntary informed consent when required, protect privacy and confidentiality, minimize risks, monitor safety, and use an ethically appropriate comparison group.

What is informed consent, and what rights should participants have?

Informed consent means participants receive clear information about the study's purpose, procedures, risks, and benefits before agreeing to participate. Participation should be voluntary when consent is required, and participants should have the right to decline or withdraw without penalty.

Why is confidentiality important in infectious disease research?

Confidentiality protects participants' private health information and reduces the risk of unauthorized disclosure. Researchers must protect information appropriately while conducting the study and reporting findings.

What are the major steps for interpreting an infectious disease research scenario?

Identify how participants were selected, determine whether researchers assigned an intervention, establish when exposure and outcome were measured, identify the available disease-frequency measure, select the appropriate association measure, calculate it correctly, interpret whether occurrence is higher, lower, or similar between groups, and recognize that association alone does not prove causation.

Study Design

The plan for how researchers collect and analyze data to answer a research question; determines participant selection, measurement of exposure and outcome, available calculations, and conclusions the study can support.

Exposure

The factor being studied in relation to a health outcome, such as antibiotic use, smoking, vaccination, injection drug use, or an environmental characteristic. Exposure does not necessarily mean a harmful factor.

Outcome

The health event or condition being studied, such as C. difficile infection, influenza, hepatitis C infection, surgical site infection, or death.

Descriptive Epidemiologic Study

A study that describes how often or how commonly a health event occurs and may examine patterns by person, place, or time. It focuses on describing disease patterns rather than primarily comparing exposures and outcomes.

Analytic Epidemiologic Study

A study that examines whether an exposure is associated with an outcome, generally by comparing groups with different exposure statuses or intervention conditions.

Observational Study

A study in which researchers observe naturally occurring exposures and outcomes without assigning the exposure. Cross-sectional, cohort, and case-control studies are observational designs.

Experimental Study

A study in which researchers assign an intervention and examine the resulting outcomes. A randomized controlled trial is a major example.

Cross-Sectional Study

An observational study that samples a population and measures exposure and current outcome status at approximately the same time. It is often described as a snapshot and can measure prevalence.

Snapshot

An analogy for a cross-sectional study because exposure and current outcome are measured at approximately the same time rather than by following participants forward.

Prevalence

The proportion of a population with an existing outcome at a particular point in time or during a specified period. Formula: existing cases divided by total population in the group.

Prevalence Ratio

A relative measure comparing prevalence in two groups. Formula: prevalence in exposed group divided by prevalence in unexposed group. A ratio greater than 1 indicates higher prevalence in the first group; a ratio of 1 indicates equal prevalence; a ratio below 1 indicates lower prevalence.

Prevalence Difference

An absolute measure comparing prevalence between two groups. Formula: prevalence in exposed group minus prevalence in unexposed group. For proportions, a result of 0.10 represents a difference of 10 percentage points.

Cohort Study

An observational study that begins by identifying exposed and unexposed groups and follows or reconstructs their follow-up to determine who develops the outcome. The direction of inquiry is exposure to outcome.

Prospective Cohort Study

A cohort study in which researchers identify exposure groups and collect follow-up data moving forward to determine who develops the outcome.

Retrospective Cohort Study

A cohort study that uses existing records to identify exposure groups and outcomes that occurred during a prior period. The conceptual direction remains exposure to outcome.

Incidence Risk

The probability that initially disease-free individuals at risk develop a new outcome during a specified period. Formula: new cases divided by the population initially at risk.

Risk Ratio

A relative measure comparing incidence risks between two groups. Formula: risk in exposed group divided by risk in unexposed group. A value above 1 indicates higher risk in the first group; 1 indicates equal risk; below 1 indicates lower risk.

Risk Difference

An absolute measure comparing incidence risks between two groups. Formula: risk in exposed group minus risk in unexposed group. A positive value indicates higher risk in the first group; a negative value indicates lower risk.

Case-Control Study

An observational study that begins by selecting cases who have the outcome and controls who do not have the outcome, then compares their prior exposure histories.

Case

A participant in a case-control study who has the outcome or disease being investigated.

Control

A participant in a case-control study who does not have the outcome or disease being investigated and is selected for comparison with cases.

Prior Exposure History

Information about whether participants were exposed to a factor before the outcome occurred; case-control studies compare these histories between cases and controls.

Odds

The number of individuals with a characteristic divided by the number without that characteristic within a group.

Odds Ratio (OR)

A relative measure comparing the odds of exposure among cases with the odds of exposure among controls. For a 2-by-2 table, OR = (a × d)/(b × c). An OR above 1 means exposure odds are higher among cases; 1 means similar odds; below 1 means lower exposure odds among cases.

2-by-2 Table

A table that organizes two categorical variables, such as disease status and exposure status, into four cells. In a case-control table, a and b represent exposed and unexposed cases, while c and d represent exposed and unexposed controls.

Why Risk Usually Cannot Be Calculated Directly in a Case-Control Study

Participants are selected according to outcome status, so the numbers of cases and controls are determined by the study design rather than by the occurrence of disease in the source population. Incidence risk and risk ratio therefore usually cannot be calculated directly.

Randomized Controlled Trial (RCT)

An experimental study in which investigators randomly assign participants to intervention groups, include a comparison or control group, and follow participants to compare outcomes.

Random Assignment

The allocation of participants to study groups by chance. It helps balance known and unknown baseline characteristics across groups on average, although differences can still occur by chance.

Control Group

The comparison group in an experimental study, such as participants receiving a placebo or standard treatment, used to evaluate outcomes relative to the intervention group.

Intervention Group

The group assigned to receive the intervention being studied, such as a new medication or vaccine.

Randomized Clinical Trial

Another name for a randomized controlled trial in which participants are randomly assigned to treatment or intervention groups and outcomes are compared.

Incidence Rate

The speed at which new outcomes occur relative to total person-time at risk. Formula: new cases divided by total person-time at risk.

Person-Time

The combined amount of time participants contribute while at risk of developing the outcome. It is used as the denominator when calculating incidence rates.

Rate Ratio

A relative measure comparing incidence rates between groups. Formula: rate in intervention or exposed group divided by rate in the comparison group. A ratio above 1 indicates a higher rate in the first group; 1 indicates equal rates; below 1 indicates a lower rate.

Rate Difference

An absolute measure comparing incidence rates between groups. Formula: rate in intervention or exposed group minus rate in the comparison group. A negative value indicates a lower rate in the first group.

Ratio Measure

A relative comparison calculated using division. It answers how many times as high or low a measure is. Risk ratios, rate ratios, and prevalence ratios are ratio measures; odds ratios compare odds.

Difference Measure

An absolute comparison calculated using subtraction. It answers how much higher or lower a measure is. Examples include risk difference, rate difference, and prevalence difference.

Relative Comparison

A comparison describing how many times as high or low one group's measure is relative to another group's measure; calculated with a ratio.

Absolute Comparison

A comparison describing the amount by which one group's measure differs from another group's measure; calculated with a difference.