CPH [2026/2027] | UPDATED ACTUAL Exam
Question and Answer | Frequently Tested
Questions
• Ecological Fallacy -✓✓ Bias that may occur because an association observed
between variables or an aggregate level does not represent the association that
exists at an individual level
• Confidence Interval -✓✓ 95% confident that the true value of a variable is
contained within the interval.
-used to account for sampling variability
-it is a point estimate +_ margin of error, where the point estimate is the best
estimate of teh unknown parameter and the margin of error is the product of the
confidence level and the standard error.
if a 95% CI for the differences in mean does not include 0 (the null value) then
there is eveidence of a statistically significant difference at sigma=0.05
• Clinical Trial Phases -✓✓ 1. Safety and Pharmacologic profiles
2. pilot efficacy studies
3. extensive clinical trials
4. after the FDA approves, look at specific effects to establish incidence of adverse
reactions, etc. longterm use effects.
• interpretation of studies -✓✓ temporality: cause precedes effect
Specificity: important in assessing the possibility of biases.
Consistency: several studies showing similar results. homogeneity statistically.
,• Confounders -✓✓ -non-causal association between exposure and outcome as a
result of a third variable.
-distortion of effect by other factors
-must be related to exposure AND outcome
-not an intermediate variable on causal pathway
• Controlling for confounders -✓✓ before data collection: random collection,
individual matching, frequency matching
After data collection: direct adjustment, indirect adjustment, mantel-haenszel,
regression techniques
• Quality Assurance vs. Quality Control -✓✓ QA: ensure quality before data
collection
QC: monitor and maintain quality during study
• reliability vs. validity -✓✓ R: precision, reproducibility
V: accuracy, absence of bias
• systematic error -✓✓ (lack of validity) if there's a difference between what is
actually being estimated and what is intended to be measured. Increasing sample
size doesn't help.
• Random Error -✓✓ (lack of precision) occurs, but increasing sample size helps.
• RCT studies -✓✓ Tests efficacy or effectiveness of healthcare services. random
allocation of participants to different treatments. Includes blinding, placebo. gold
standard for evidence.
,• Community Intervention/cluster RCT -✓✓ community-wide basis or groupwide
• Case-Crossover RCT design -✓✓ -cases serve as their own control
-exposure has transient effect
• Cross Sectional Studies -✓✓ SNAPSHOT! at a single point in time. tells the
prevalence and association. causation cannot be implied. a study that examines the
relationship between diseases and other variables as they exist in a defined
population at one particular time.
• Matching -✓✓ used to make cases and controls as similar as possible to avoid
confounding. ex: race, gender, age. +Maybe the only way to control confounding.
increases statistical power, straightforward. -requires use of special analytical
techniques, residual confounding can occur if you match continuous variables by
category.
• types of matching -✓✓ individual matching: case and control matched
individually
frequency matching: a group of controls
Minimum Euclidean Distance measure: match to closest person.
• Cohort Studies -✓✓ RISK RATIO, RELATIVE RISK, INCIDENCE RATE,
RATE RATIO
-rare exposures
-group of subjects who shared experiences during a particular time. Determines if
incidence is related to exposure.
, • Concurrent/longitudinal cohort studies -✓✓ starts now (with a baseline exam)
and goes into future. expensive and time intensive.
• non-concurrent/retrospective cohort studies -✓✓ assembled in past based on
existing records. faster and quicker, but records can be limited or biased. follow up
can be hard.
• Prevalence of disease -✓✓ measure of the burden of disease in a community
(new and existing cases). the number of events in a given population at a
designated time.
-obscures causal relationships
• point prevalence -✓✓ -proportion of pop that is diseased during a single point of
time.
-at a specific point in time
number of cases at a particular moment/
number in population at that moment
• period prevalence -✓✓ -proportion of pop that is diseased during a specified
duration of time.
-during a specific period of time
number of cases during a specified time period/
number in population at midpoint of period
• incidence of disease -✓✓ measure of risk (new cases)
Question and Answer | Frequently Tested
Questions
• Ecological Fallacy -✓✓ Bias that may occur because an association observed
between variables or an aggregate level does not represent the association that
exists at an individual level
• Confidence Interval -✓✓ 95% confident that the true value of a variable is
contained within the interval.
-used to account for sampling variability
-it is a point estimate +_ margin of error, where the point estimate is the best
estimate of teh unknown parameter and the margin of error is the product of the
confidence level and the standard error.
if a 95% CI for the differences in mean does not include 0 (the null value) then
there is eveidence of a statistically significant difference at sigma=0.05
• Clinical Trial Phases -✓✓ 1. Safety and Pharmacologic profiles
2. pilot efficacy studies
3. extensive clinical trials
4. after the FDA approves, look at specific effects to establish incidence of adverse
reactions, etc. longterm use effects.
• interpretation of studies -✓✓ temporality: cause precedes effect
Specificity: important in assessing the possibility of biases.
Consistency: several studies showing similar results. homogeneity statistically.
,• Confounders -✓✓ -non-causal association between exposure and outcome as a
result of a third variable.
-distortion of effect by other factors
-must be related to exposure AND outcome
-not an intermediate variable on causal pathway
• Controlling for confounders -✓✓ before data collection: random collection,
individual matching, frequency matching
After data collection: direct adjustment, indirect adjustment, mantel-haenszel,
regression techniques
• Quality Assurance vs. Quality Control -✓✓ QA: ensure quality before data
collection
QC: monitor and maintain quality during study
• reliability vs. validity -✓✓ R: precision, reproducibility
V: accuracy, absence of bias
• systematic error -✓✓ (lack of validity) if there's a difference between what is
actually being estimated and what is intended to be measured. Increasing sample
size doesn't help.
• Random Error -✓✓ (lack of precision) occurs, but increasing sample size helps.
• RCT studies -✓✓ Tests efficacy or effectiveness of healthcare services. random
allocation of participants to different treatments. Includes blinding, placebo. gold
standard for evidence.
,• Community Intervention/cluster RCT -✓✓ community-wide basis or groupwide
• Case-Crossover RCT design -✓✓ -cases serve as their own control
-exposure has transient effect
• Cross Sectional Studies -✓✓ SNAPSHOT! at a single point in time. tells the
prevalence and association. causation cannot be implied. a study that examines the
relationship between diseases and other variables as they exist in a defined
population at one particular time.
• Matching -✓✓ used to make cases and controls as similar as possible to avoid
confounding. ex: race, gender, age. +Maybe the only way to control confounding.
increases statistical power, straightforward. -requires use of special analytical
techniques, residual confounding can occur if you match continuous variables by
category.
• types of matching -✓✓ individual matching: case and control matched
individually
frequency matching: a group of controls
Minimum Euclidean Distance measure: match to closest person.
• Cohort Studies -✓✓ RISK RATIO, RELATIVE RISK, INCIDENCE RATE,
RATE RATIO
-rare exposures
-group of subjects who shared experiences during a particular time. Determines if
incidence is related to exposure.
, • Concurrent/longitudinal cohort studies -✓✓ starts now (with a baseline exam)
and goes into future. expensive and time intensive.
• non-concurrent/retrospective cohort studies -✓✓ assembled in past based on
existing records. faster and quicker, but records can be limited or biased. follow up
can be hard.
• Prevalence of disease -✓✓ measure of the burden of disease in a community
(new and existing cases). the number of events in a given population at a
designated time.
-obscures causal relationships
• point prevalence -✓✓ -proportion of pop that is diseased during a single point of
time.
-at a specific point in time
number of cases at a particular moment/
number in population at that moment
• period prevalence -✓✓ -proportion of pop that is diseased during a specified
duration of time.
-during a specific period of time
number of cases during a specified time period/
number in population at midpoint of period
• incidence of disease -✓✓ measure of risk (new cases)
