Preventive Medicine: Epidemiology 6

Causal Research in Epidemiology

• Causal research in epidemiology requires two fundamental distinctions:
  • distinction between people who have and people who do not have exposure to the risk factor (or protective factor) under study (independent variable)
  • distinction between people who have and people who do not have the disease (or other outcome) under study (dependent variable)

Complications of Epidemiologic Research

• Measurements of exposure and outcome are subject to random errors and biases
• Analysis may be complicated by the need to analyze several independent (possibly causal) variables at the same time, including interactions
• Measuring different degrees of strength of exposure to the risk factor, duration of exposure, or both, may be necessary
• Determining the time of onset of exposure may be difficult for risk factors such as sedentary lifestyle and excess intake of dietary sodium
• Measuring different levels of disease severity is necessary as exposure and outcome may vary across a range of values

Definition of Study Groups

• Causal research depends on the measurement of differences in cohort studies and case-control studies
• In cohort studies, the difference is between the frequency of disease in persons exposed to a risk factor and the frequency of disease in persons not exposed to the same risk factor
• In case-control studies, the difference is between the frequency of the risk factor in case participants (persons with the disease) and the frequency of the risk factor in control participants (persons without the disease)

Comparison of Risks in Different Study Groups

• Differences in risk can be measured in absolute terms or in relative terms
• Absolute differences in risks or rates can be expressed as a risk difference or as a rate difference
• Relative differences in risks or rates can be expressed as a relative risk (RR) or an odds ratio (OR)

Absolute Differences in Risk

• The risk difference is the risk in the exposed group minus the risk in the unexposed group
• The rate difference is the rate in the exposed group minus the rate in the unexposed group
• The risk difference is also known as the attributable risk because it is an estimate of the amount of risk that can be attributed to, or is caused by, the risk factor

Relative Differences in Risk

• The relative risk (RR) can be expressed in terms of a risk ratio or an odds ratio
• The risk ratio is the ratio of the risk in the exposed group to the risk in the unexposed group
• The odds ratio (OR) is the ratio of the odds of exposure in the diseased group to the odds of exposure in the nondiseased group

Data Collection and Analysis

• Clinical medicine requires constant collection, evaluation, analysis, and use of quantitative and qualitative data.
• Data are used for diagnosis, prognosis, and choosing and evaluating treatments.
• Errors in data can occur and are difficult to eliminate, and can be categorized into differential errors (bias) and nondifferential errors (random errors).

Promoting Accuracy and Precision

• Two distinct goals of data collection are accuracy and precision.
• Accuracy refers to the ability of a measurement to be correct on average, while precision refers to the ability of a measurement to give the same result or a similar result with repeated measurements.
• Both accuracy and precision are essential for data collection, as either one alone is not sufficient.

Reducing Errors

• There are two types of errors to avoid in data collection: differential errors (bias) and nondifferential errors (random errors).
• Differential errors result from systematic or consistent errors that tend to be inaccurate in a particular direction.
• Nondifferential errors result from random errors that can produce both high and low values.

Intraobserver and Interobserver Variability

• Intraobserver variability refers to the differences in measurements or interpretations obtained by the same clinician when measuring or interpreting the same data multiple times.
• Interobserver variability refers to the differences in measurements or interpretations obtained by different clinicians when measuring or interpreting the same data.
• Reducing intraobserver and interobserver variability is essential to ensure accurate and reliable data.

Studying the Accuracy and Usefulness of Screening and Diagnostic Tests

• The accuracy and usefulness of screening and diagnostic tests can be evaluated by assessing their performance in two groups of individuals: those with the disease and those without the disease.
• Factors that influence the accuracy and usefulness of tests include the stage of the disease, the spectrum of disease in the study population, and the characteristics of the population being tested.

False-Positive and False-Negative Results

• False-positive results occur when a test result is positive in a person without the disease.
• False-negative results occur when a test result is negative in a person with the disease.
• False-positive and false-negative results can be influenced by the stage of the disease, the spectrum of disease in the study population, and the characteristics of the population being tested.

Sensitivity and Specificity

• Sensitivity refers to the ability of a test to detect a disease when present, and is calculated as the proportion of true-positive results among all diseased individuals.
• Specificity refers to the ability of a test to indicate nondisease when no disease is present, and is calculated as the proportion of true-negative results among all nondiseased individuals.
• Both sensitivity and specificity are essential measures of a test's performance.

Predictive Values

• Predictive values are used to answer two important clinical questions: what is the probability that a person has the disease if the test result is positive, and what is the probability that a person does not have the disease if the test result is negative.

• Positive predictive value (PPV) is the proportion of true-positive results among all positive test results.

• Negative predictive value (NPV) is the proportion of true-negative results among all negative test results.

• Predictive values are influenced by the prevalence of the condition being tested, and can be difficult to interpret in the presence of false-positive or false-negative results.### Screening Tests and Confirmatory Tests

• When clinicians test for rare conditions, most positive test results are likely to be falsely positive

• Additional testing is necessary to determine if the disease is present in individuals with positive results

• Screening tests are still worthwhile for conditions with low prevalence, as the number of individuals requiring follow-up diagnostic tests may be small

Principles of Screening Tests and Confirmatory Tests

• One test does not make a diagnosis, unless it is a pathognomonic test (a "gold standard")
• Box 7-1 summarizes key principles concerning screening tests and confirmatory tests

Likelihood Ratios, Odds Ratios, and Cutoff Points

• Likelihood ratios are not influenced by the prevalence of the disease
• Likelihood ratio positive (LR+) is the ratio of sensitivity to false-positive error rate
• LR+ = [a/(a + c)] ÷ [b/(b + d)]
• A higher LR+ indicates a better test, with a ratio much larger than 1
• Sensitivity and false-positive error rate are independent of disease prevalence, and their ratio is also independent
• Likelihood ratio negative (LR-) is the ratio of false-negative error rate to specificity
• LR- = [c/(a + c)] ÷ [d/(b + d)]
• A smaller LR- indicates a better test, with a ratio closer to 0
• If LR+ is large and LR- is small, it is likely a good test
• Experts in test analysis sometimes calculate the ratio of LR+ to LR- to obtain a measure of separation between positive and negative tests

Characteristics of Tests Needed to “Rule Out” and “Rule In” a Diagnosis

• A clinician must order various tests to screen or “rule out” (discard) false hypotheses when a patient presents with complaints.
• These tests should be highly sensitive tests, which have a low false-negative error rate, to ensure that not many true cases of the disease are missed.
• After most of the hypothesized diagnoses have been eliminated, the clinician begins to consider tests to “rule in” (confirm) the true diagnosis.
• These tests should be highly specific tests, which have a small false-positive error rate, to ensure that not many patients are misdiagnosed as having a particular disease when in fact they have another disease.

Principles of Testing

• A screening test, used to rule out a diagnosis, should have a high degree of sensitivity.
• A confirmatory test, used to rule in a diagnosis, should have a high degree of specificity.

Concepts of Proportions and Odds

• A proportion is a ratio of the form a/(a + b), whereas an odds is a ratio of the form a/b.
• Odds can only describe a variable that is dichotomous (i.e., has only two possible outcomes).
• The odds of a particular outcome can be converted to the probability of that outcome, and vice versa, using the formula: Probability of outcome X = Odds of outcome X / (1 + Odds of outcome X).

Receiver Operating Characteristic (ROC) Curves

• ROC curves are used to decide on a good cutoff point for a clinical test that measures continuous variables.
• The curve plots the sensitivity of a test against the false-positive error rate for several possible cutoff points.
• The y-axis shows the sensitivity of a test, and the x-axis shows the false-positive error rate (1 - specificity).
• The ROC curve can be considered a graph of the Likelihood Ratio (LR+).
• The ideal ROC curve for a test would rise almost vertically from the lower left corner and move horizontally almost along the upper line, indicating a high sensitivity and a low false-positive error rate.
• The ROC curve for most clinical tests is somewhere between the ideal and the no benefit line, which represents a diagonal straight line from the lower left to the upper right corner.