NU505 Chapter 7 – Risk from Disease to Exposure

NU505 Chapter 7 – Risk from Disease to Exposure

Student Name

Purdue University Global

NU505 Clinical Epidemiology and Population Health Promotion

Prof. Name:

Date

Understanding Risk by Looking Back at Exposure

Case-control studies are an efficient epidemiological research design used to identify whether a past exposure is associated with a disease or health outcome. Unlike cohort studies, which follow individuals over time from exposure to disease, case-control studies begin with people who already have the disease (cases) and compare them with similar individuals without the disease (controls). Researchers then examine prior exposure to potential risk factors. This approach is especially valuable for studying rare diseases or conditions with long latency periods because it is faster, less expensive, and requires fewer participants than cohort studies.

Case-control studies provide an estimate of the association between exposure and disease, usually expressed as an odds ratio, which approximates relative risk under certain conditions. Although these studies are highly efficient, they are more vulnerable to bias and cannot directly measure disease incidence or absolute risk.

Understanding Risk and the Need for Case-Control Studies

Many chronic diseases develop slowly. The latency period—the time between exposure to a risk factor and the appearance of disease—often spans several decades.

For example:

• Smoking may lead to lung cancer or coronary heart disease after 20 years or more.

• Poor diet and lack of physical activity during early adulthood may contribute to osteoporosis later in life.

• Environmental and occupational exposures may not produce detectable health effects until many years after exposure.

Because of these long delays, cohort studies require extensive follow-up, making them expensive, time-consuming, and difficult to conduct. Case-control studies overcome these challenges by starting with individuals who already have the disease and investigating their previous exposures.

How Case-Control Studies Work

Case-control studies compare two carefully selected groups:

• Cases: Individuals who have the disease or outcome of interest.

• Controls: Similar individuals who do not have the disease.

Researchers then determine how frequently each group was exposed to a suspected risk factor in the past.

The process follows these basic steps:

1. Identify individuals with the disease.

2. Select appropriate controls from the same source population.

3. Measure previous exposure in both groups.

4. Compare exposure frequencies.

5. Calculate the odds ratio to estimate the strength of the association.

Instead of following participants forward through time, researchers work backward from disease to exposure.

Cohort Studies vs. Case-Control Studies

Both study designs investigate relationships between exposures and diseases, but they differ significantly in methodology and outcomes.

Feature	Cohort Study	Case-Control Study
Starting point	Exposure	Disease
Direction of study	Forward in time	Backward in time
Best for	Common diseases	Rare diseases
Follow-up required	Yes	No
Measures incidence	Yes	No
Measures absolute risk	Yes	No
Estimates relative risk	Directly	Indirectly using odds ratio
Cost and duration	High	Lower
Efficiency	Lower for rare diseases	Higher for rare diseases

Advantages of Case-Control Studies

Case-control studies remain one of the most widely used observational research methods because they offer several practical benefits.

Faster Than Cohort Studies

Researchers do not need to wait years for disease development because cases already exist.

Cost-Effective

Fewer participants require detailed exposure assessment, reducing overall research costs.

Ideal for Rare Diseases

Rare diseases often affect very few individuals, making cohort studies impractical. Case-control studies focus only on affected individuals and an appropriate comparison group.

Useful for Long Latency Diseases

Diseases that develop decades after exposure can be studied without prolonged follow-up.

Allows Evaluation of Multiple Risk Factors

Researchers can investigate numerous past exposures that may contribute to a single disease.

Limitations of Case-Control Studies

Despite their efficiency, case-control studies have several important limitations.

Greater Risk of Bias

Because researchers collect exposure information after disease has occurred, several biases can affect study validity.

Common biases include:

• Recall bias

• Selection bias

• Information bias

• Confounding

Cannot Measure Disease Incidence

Case-control studies cannot calculate:

• Absolute risk

• Disease incidence

• Attributable risk

• Population risk

They primarily estimate the association between exposure and disease.

Temporal Relationships May Be Less Certain

Although exposure generally occurs before disease, retrospective data collection may make establishing the exact timing more challenging.

Real-World Example: Bisphosphonates and Atypical Femoral Fractures

Researchers in Sweden investigated whether bisphosphonates, medications commonly prescribed for osteoporosis, increased the risk of atypical femoral fractures.

The study included:

• 160 women aged 55 years or older with atypical femoral fractures (cases)

• 774 women with ordinary femoral fractures (controls)

Investigators adjusted for several potential confounding variables, including:

• Age

• Osteoporosis

• Previous fractures

• Corticosteroid use

• Estrogen therapy

The findings showed:

• Women using bisphosphonates had approximately 29 times higher risk of atypical fractures.

• Risk increased with longer duration of treatment.

• Women taking the medication for four years or longer had substantially higher estimated risk than shorter-term users.

This study demonstrated how case-control designs can identify important medication-related risks more efficiently than cohort studies.

Real-World Example: Helmet Use in Skiers and Snowboarders

Researchers in Norway evaluated whether helmets reduce head injuries among skiers and snowboarders.

The study included:

• 578 individuals with head injuries (cases)

• A comparison group of uninjured skiers and snowboarders (controls)

Researchers adjusted for several variables, including:

• Age

• Sex

• Nationality

• Skiing ability

• Equipment type

• Previous ski school attendance

• Rental versus owned equipment

Results showed:

• Helmet use was associated with approximately 60% lower risk of head injury.

• A later study found helmet use remained protective, particularly against severe head injuries.

This example demonstrates how case-control studies evaluate preventive interventions under real-world conditions.

Selecting Appropriate Cases

Choosing appropriate cases is essential for producing reliable results.

Researchers generally prefer incident cases, which represent newly diagnosed disease, rather than prevalent cases, which include individuals who have lived with the disease for varying lengths of time.

Incident cases are preferred because they:

• Better represent disease occurrence.

• Reduce survival bias.

• Minimize distortion caused by disease duration.

• Improve interpretation of exposure-disease relationships.

Researchers should also use clear diagnostic criteria and confirm diagnoses using standardized methods whenever possible.

Selecting Appropriate Controls

Controls should accurately represent the population from which the cases originated.

An ideal control group should:

• Come from the same source population.

• Be capable of developing the disease.

• Have similar opportunities for exposure.

• Differ only in disease status.

Proper control selection is one of the most critical aspects of minimizing bias.

Types of Control Selection

Population-Based Case-Control Studies

In population-based studies, both cases and controls are randomly selected from a clearly defined population.

Advantages include:

• Better representation of the target population

• Reduced selection bias

• Greater generalizability

Nested Case-Control Studies

Nested case-control studies are conducted within an existing cohort.

Researchers identify cases from the cohort and select controls from participants who remain disease-free.

Benefits include:

• Higher-quality exposure data

• Reduced selection bias

• Lower research costs compared with analyzing the entire cohort

• Ability to combine strengths of cohort and case-control designs

Hospital and Community Controls

When population-based sampling is not feasible, researchers may use hospital or community controls.

Hospital controls are convenient but may introduce bias because hospitalized patients often differ from the general population.

Community controls may better represent the source population, although ensuring that cases and controls come from the same underlying population remains challenging.

Multiple Control Groups

Researchers sometimes include more than one control group to evaluate potential selection bias.

If different control groups produce similar findings, confidence in the study’s conclusions increases.

If findings differ substantially, investigators should examine possible sources of bias or reconsider the study design.

Multiple Controls Per Case

Increasing the number of controls assigned to each case can improve statistical power, particularly when studying rare diseases.

Benefits include:

• Improved precision of risk estimates

• Greater ability to detect true associations

• More efficient use of limited case numbers

However, adding many controls eventually produces only small improvements in statistical power.

Common Sources of Bias in Case-Control Studies

Researchers must carefully minimize bias throughout study design and analysis.

Major sources include:

• Recall bias: Cases may remember past exposures differently than controls.

• Selection bias: Cases and controls may not represent the same population.

• Confounding: Other variables may influence both exposure and disease.

• Information bias: Inaccurate exposure measurement can distort results.

• Overmatching: Excessive matching can eliminate genuine differences between cases and controls.

Careful participant selection, standardized exposure assessment, and statistical adjustment help reduce these biases.

Measuring Association: Odds Ratio

Case-control studies typically calculate an odds ratio (OR) rather than relative risk.

Interpretation includes:

• OR = 1: No association.

• OR > 1: Exposure is associated with increased disease risk.

• OR < 1: Exposure may be protective.

When diseases are uncommon, the odds ratio closely approximates the relative risk.

Key Takeaways

Case-control studies provide an efficient method for investigating relationships between exposures and diseases, particularly when diseases are rare or have long latency periods. Their retrospective design allows researchers to identify potential risk factors quickly while requiring fewer resources than cohort studies. However, careful selection of cases and controls, accurate exposure assessment, and appropriate adjustment for confounding variables are essential to produce valid and reliable findings.

Case-control studies continue to play a central role in epidemiology, clinical research, pharmacovigilance, and public health by identifying disease risk factors and evaluating preventive interventions.

 

Frequently Asked Questions

What is a case-control study?

A case-control study is an observational study that compares individuals with a disease (cases) to individuals without the disease (controls) to determine whether previous exposure to a potential risk factor differs between the two groups.

Why are case-control studies important?

They are particularly useful for studying rare diseases, diseases with long latency periods, and situations where cohort studies would be expensive or impractical.

What is the main advantage of a case-control study?

The greatest advantage is efficiency. Researchers can investigate disease risk without waiting years for outcomes to occur.

What is the main limitation of a case-control study?

The major limitation is the increased potential for bias, especially recall bias and selection bias. Additionally, case-control studies cannot directly measure disease incidence or absolute risk.

What is an odds ratio?

An odds ratio measures the strength of the association between an exposure and a disease. An odds ratio greater than one indicates increased risk, while an odds ratio less than one suggests a protective association.

What is a nested case-control study?

A nested case-control study is conducted within an existing cohort, allowing researchers to use high-quality cohort data while collecting detailed exposure information only for selected cases and controls.

Why are incident cases preferred over prevalent cases?

Incident cases represent newly diagnosed disease and reduce survival bias, providing a more accurate assessment of the relationship between exposure and disease development.

Case-Control Studies: Matching, Exposure Measurement, and Odds Ratio Explained

Case-control studies identify factors associated with disease by comparing individuals with a disease (cases) to those without it (controls). Researchers improve study accuracy by carefully selecting controls, matching important characteristics, accurately measuring exposure, and calculating the odds ratio to estimate the strength of the association between exposure and disease. However, poor matching or biased exposure assessment can reduce study validity and lead to misleading conclusions.

Matching in Case-Control Studies

Matching is a research technique used to make cases and controls as similar as possible for characteristics other than the exposure being studied. This helps reduce confounding, making it easier to determine whether an exposure is truly associated with a disease.

Researchers most commonly match participants based on:

• Age

• Sex

• Geographic location

• Disease severity

• Risk profile

By ensuring these characteristics are comparable between groups, researchers obtain a more accurate estimate of the relationship between exposure and disease.

Why Matching Is Important

Matching improves study quality because it:

• Reduces the influence of confounding variables.

• Increases statistical efficiency.

• Improves the ability to detect genuine exposure-disease relationships.

• Creates comparable case and control groups.

For example, age and sex frequently influence both disease risk and exposure. Matching these variables helps isolate the effect of the exposure under investigation.

Umbrella Matching

Researchers sometimes use umbrella matching, where participants are matched using a broader characteristic that represents multiple underlying variables.

Examples include matching by:

• Hospital

• Community

• Healthcare system

These factors indirectly account for variables such as:

• Socioeconomic status

• Education level

• Income

• Race and ethnicity

• Access to healthcare

• Local healthcare practices

Umbrella matching simplifies study design when measuring every possible confounder individually would be impractical.

Risks of Overmatching

Although matching improves study validity, excessive matching can introduce bias. This problem is known as overmatching.

Overmatching occurs when researchers match participants on variables that are closely related to the exposure itself. As a result, cases and controls become artificially similar, making it difficult to detect genuine differences.

For example, consider a study investigating whether nonsteroidal anti-inflammatory drugs (NSAIDs) increase the risk of kidney failure. If researchers match cases and controls based on arthritis symptoms, both groups are likely to have similar NSAID use because arthritis is commonly treated with NSAIDs. This reduces observable differences in exposure and weakens the measured association.

Overmatching may also occur when:

• Matching variables lie within the causal pathway between exposure and disease.

• Matching variables are highly correlated with one another.

• Diseases with similar treatments are selected for comparison.

Limitations of Matching

Matching has several disadvantages despite its benefits.

Once researchers match a variable, they can no longer evaluate whether that variable independently influences disease risk. Additionally, finding appropriately matched controls becomes increasingly difficult as more matching criteria are added.

Researchers often overcome this challenge by relaxing matching criteria. For example, instead of matching participants by exact age, they may match individuals within a five-year age range.

Measuring Exposure in Case-Control Studies

Accurate exposure assessment is essential because incorrect classification can produce biased results.

The most reliable exposure information comes from records collected before disease develops, including:

• Pharmacy records

• Surgical records

• Laboratory databases

• Stored biological specimens

• Electronic medical records

These sources reduce bias because exposure information was documented before participants knew their disease status.

However, many exposures cannot be obtained from medical records alone.

NU505 Chapter 7 – Risk from Disease to Exposure

Examples include:

• Physical activity

• Dietary habits

• Recreational drug use

• Over-the-counter medication use

• Smoking history

• Alcohol consumption

Researchers often collect these exposures through interviews or questionnaires.

Example: Cannabis Use and Psychotic Disorders

A multicenter European and Brazilian case-control study investigated whether cannabis increases the risk of first-episode psychotic disorders.

Researchers included:

• 901 patients with newly diagnosed psychosis

• 1,237 population-based controls

Cannabis exposure was assessed by:

• Frequency of use

• Duration of use

• Cannabis potency

• THC concentration

After adjusting for other risk factors, findings showed:

• Daily cannabis users had approximately three times greater risk of psychotic disorders.

• Daily users of high-potency cannabis had approximately five times greater risk compared with non-users.

This study illustrates how detailed exposure measurement strengthens case-control research.

Recall Bias

One of the greatest challenges in case-control studies is recall bias.

Recall bias occurs when cases remember previous exposures differently than controls because they know they have the disease.

Examples include:

• Parents of children with Reye syndrome recalling aspirin use more readily.

• Patients with prostate cancer remembering prior vasectomy after media reports.

• Individuals with environmental illnesses recalling exposure more frequently following public health campaigns.

Researchers reduce recall bias by:

• Asking standardized questions.

• Limiting disclosure of the study’s specific hypothesis.

• Using validated questionnaires.

• Confirming exposure through independent records whenever possible.

Interviewer Bias

Exposure assessment can also be influenced by healthcare providers or interviewers.

If investigators already suspect a particular exposure causes disease, they may question cases more thoroughly than controls or document exposures more carefully.

For example:

• Physicians may record family history more thoroughly in cancer patients.

• Clinicians may ask detailed questions about birth control use in women with deep vein thrombosis.

Researchers minimize interviewer bias by:

• Blinding interviewers to the study hypothesis.

• Using standardized interview protocols.

• Collecting information from multiple independent sources.

Reverse Causation

Sometimes disease develops before the exposure under investigation.

This phenomenon, called reverse causation, may incorrectly suggest that exposure caused disease when early disease symptoms actually led to the exposure.

For example:

• Early symptoms of schizophrenia may increase cannabis use before diagnosis.

• Chest pain may lead physicians to prescribe beta-blockers before myocardial infarction develops.

Researchers address reverse causation by excluding participants with early disease manifestations or pre-existing symptoms before measuring exposure.

Multiple Exposure Assessment

Case-control studies efficiently evaluate multiple exposures simultaneously.

Researchers may examine:

• Lifestyle factors

• Medication use

• Environmental exposures

• Dietary patterns

• Early disease symptoms

This flexibility makes case-control studies particularly valuable for identifying potential disease risk factors.

Example: Early Symptoms of Ovarian Cancer

A case-control study conducted in England compared women diagnosed with ovarian cancer with matched controls.

Researchers identified several symptoms associated with ovarian cancer, including:

• Abdominal distension

• Increased urinary frequency

• Abdominal pain

• Postmenopausal bleeding

• Loss of appetite

• Rectal bleeding

• Abdominal bloating

After excluding symptoms occurring shortly before diagnosis, three symptoms remained independently associated with earlier disease:

• Abdominal distension

• Urinary frequency

• Abdominal pain

These findings demonstrate how case-control studies can identify early clinical indicators that support earlier diagnosis.

Understanding the Odds Ratio

The odds ratio (OR) is the primary measure of association in case-control studies.

It compares the odds of exposure among cases with the odds of exposure among controls.

Interpretation is straightforward:

• OR = 1: No association.

• OR > 1: Exposure is associated with increased disease risk.

• OR < 1: Exposure appears protective.

The farther the odds ratio is from 1, the stronger the observed association.

Odds Ratio Versus Relative Risk

Unlike cohort studies, case-control studies cannot directly calculate disease incidence because researchers select participants based on disease status rather than exposure.

Instead, they calculate the odds ratio, which estimates relative risk.

The odds ratio closely approximates relative risk when:

• The disease is uncommon.

• Disease incidence is relatively low.

• Controls accurately represent the source population.

However, when disease becomes more common:

• Odds ratios greater than 1 tend to overestimate relative risk.

• Odds ratios less than 1 tend to underestimate the protective effect.

Researchers should therefore interpret odds ratios carefully, particularly for common outcomes.

A commonly accepted guideline is that the odds ratio closely approximates relative risk when disease incidence is below approximately 1–5% in the unexposed population.

Case-control studies remain one of the most efficient observational designs for studying rare diseases because they require fewer participants and less follow-up than cohort studies while still providing valuable estimates of exposure-disease relationships.

• Matching improves the validity of case-control studies by reducing confounding but should be used cautiously to avoid overmatching.

• Overmatching can bias results by making exposure rates artificially similar between cases and controls.

• Recall bias occurs when individuals with disease remember past exposures differently than healthy controls.

• Blinding interviewers and using objective records reduce exposure measurement bias.

• The odds ratio is the standard measure of association in case-control studies and estimates relative risk when disease incidence is low.

• Case-control studies are particularly useful for investigating rare diseases and multiple potential risk factors simultaneously.

FAQs

What is matching in a case-control study?

Matching is a method of selecting controls with characteristics similar to cases, such as age or sex, to reduce confounding and improve study validity.

What is overmatching?

Overmatching occurs when researchers match participants on variables closely related to the exposure, making cases and controls too similar and potentially masking true associations.

Why is recall bias important?

Recall bias can lead to inaccurate exposure measurement because participants with disease may remember past exposures more clearly than those without disease.

What is the odds ratio?

The odds ratio measures the association between exposure and disease by comparing the odds of exposure among cases with the odds among controls.

When does the odds ratio approximate relative risk?

The odds ratio provides a close estimate of relative risk when the disease being studied is uncommon, generally with an incidence below approximately 1–5%.

Why are case-control studies useful?

Case-control studies are efficient for studying rare diseases, evaluating multiple exposures, and generating evidence about potential disease risk factors with fewer participants than cohort studies.

References

Hennekens, C. H., & Buring, J. E. (1987). Epidemiology in medicine. Little, Brown and Company. https://archive.org/details/epidemiologyinme0000henn

Grimes, D. A., & Schulz, K. F. (2002). Bias and causal associations in observational research. The Lancet, 359(9302), 248–252. https://doi.org/10.1016/S0140-6736(02)07451-2

Schulz, K. F., & Grimes, D. A. (2002). Case-control studies: Research in reverse. The Lancet, 359(9304), 431–434. https://doi.org/10.1016/S0140-6736(02)07605-5

Fletcher, R. H., Fletcher, S. W., & Fletcher, G. S. (2021). Clinical epidemiology: The essentials (6th ed.). Wolters Kluwer.

Pearce, N. (2016). Analysis of matched case-control studies. BMJ, 352, i969. https://doi.org/10.1136/bmj.i969

Cornfield, J. (1951). A method of estimating comparative rates from clinical data: Applications to cancer of the lung, breast, and cervix. Journal of the National Cancer Institute, 11(6), 1269–1275. https://academic.oup.com/jnci/article/11/6/1269/897361

NU505 Chapter 7 – Risk from Disease to Exposure

Rothman, K. J., Greenland, S., & Lash, T. L. (2021). Modern Epidemiology (4th ed.). Wolters Kluwer. https://shop.lww.com/Modern-Epidemiology/p/9781975161756

Grimes, D. A., & Schulz, K. F. (2002). Compared to what? Finding controls for case-control studies. The Lancet, 359(9302), 248–252. https://doi.org/10.1016/S0140-6736(02)07449-8

Schilcher, J., Michaëlsson, K., & Aspenberg, P. (2011). Bisphosphonate use and atypical fractures of the femoral shaft. New England Journal of Medicine, 364(18), 1728–1737. https://www.nejm.org/doi/full/10.1056/NEJMoa1010650

Sulheim, S., Holme, I., Ekeland, A., & Bahr, R. (2006). Helmet use and risk of head injuries in alpine skiers and snowboarders. JAMA, 295(8), 919–924. https://jamanetwork.com/journals/jama/fullarticle/202047