Biostatistics Flashcards
What does Prevalence, Incidence and Attack Rate tell us?
What is the frequency of disease in a population?
What does Sensitivity and specificity tell us?
How well does a test differentiate sick from healthy people?
What does predictive value tell us?
Of those in a population who test as sick of healthy, how many are truly sick or healthy?
What does Risk Reduction/Increase and Number-needed-to-treat/harm tell us?
What is the impact of a medicine/treatment?
What does point prevalence help us understand?
Disease burden or extent of a health problem.
What is prevalence?
[Number with a disease at a specific point in time]/[Number at risk of illness during that time period]
What is period prevalence?
Prevalence during a period of time
What is lifetime prevalence?
Prevalence over the course of a lifetime
What does incidence help us understand?
The risk of a specific health event
What is incidence?
[Number of NEW people with DZ during a time period]/[Number at risk of illness during that time period]
What is the main measure of acute diseases?
Incidence
What helps determine causation?
Incidence
What is cumulative incidence?
Total number reported over time
What is Attack Rate?
Refers to outbreaks - similar to prevalence over a very short period of time
When is Attack Rate used?
When the nature of disease is acute and population observed for short period of time (ex. outbreaks, specific exposures)
How do you calculate Attack Rate?
[Number new cases]/[Number exposed]
How do you calculate Secondary Attack Rate?
[Number new cases]/[Number exposed - primary cases]
What does Secondary Attack Rate measure?
Person to person spread of disease after initial exposure
What is Secondary Attack Rate similar to over a very short period of time?
Incidence
What affects prevalence and incidence?
- Duration of illness (longer –> higher prevalence)
- Number of new cases (more new cases –> higher prevalence) - incidence high
- Migration - In (ill –> higher prevalence); Out (well –> higher prevalence)
- ->Recovery and death –> lower prevalence
- Prevention –> lower incidence
- Changes in diagnostic criteria or reporting
What is the relationship between prevalence and incidence if the disease is long term (ex. diabetes)?
Prevalence > Incidence
What is the relationship between prevalence and incidence if the illness is acute (ex. flu)?
Prevalence ~ Incidence
What is sensitivity?
The probability that a diseased person will be identified correctly by a diagnostic/screening test
What is another name for sensitivity?
True-positiive probability or true-positive rate
What is the equation for sensitivity?
True Positives/ Total # of ill people
What should you remember with Sensitivity?
SNOUT - High sensitivity rules disease out
What is the total number of ill people?
True positives + False negatives
What is Specificity?
Probability that a well (non-diseased) person will be identified correctly by a diagnostic/screening test
What is another name for specificity?
True-negative probability
What is the equation for specificity?
True negatives/total # of well people
What should you remember with Specificity?
SPIN - High specificity disease rules in
What is the total # of well people?
TN + FP
What does a high sensitivity test err on the side of?
Over-diagnosing
What does a high specificity test err on the side of?
Under-diagnosing
What should you remember with high sensitivity tests?
- Identify most or all possible disease cases; may identify some healthy people as sick
- Most useful when under-diagnosing may lead to severe consequences (ex. fast developing cancers)
What should you remember with high specificity tests?
- Identify most or all well people; may miss some of the sick people
- Most useful when over-diagnosing leads to dangerous, painful or unnecessary treatment
What is a predictive value?
Probability that a test will give the correct diagnosis
What does predictive value depend on?
- Test sensitivity and specificity; prevalence of the DZ in the population being tested
- Predictive values will vary from population to population and study to study
What is Positive Predictive Value?
Probability that a person who tests positive for a disease truly has it (is really sick)
What is the equation for PPV?
PPV = TP/(TP + FP) –> Top row of a 2x2 table
What is the equation for NPV?
NPV = NP/(NP + FN) –> Bottom row of a 2x2 table
What is Negative Predictive Value?
Probability that a person who tests negative for a disease truly is well
How does High prevalence relate to predictive value?
- Higher disease prevalence –> Higher PPV (greater chance that positive test result reflects true illness)
- -> Lower NPV (lower change that negative test reflects disease-free status)
How does Low prevalence relate to predictive value?
- Lower disease prevalence –> Lower PPV (lower chance that positive test result reflects true illness)
- -> Higher NPV (greater chance that negative test result reflects disease-free status)
When is Risk Reduction and Number-Needed-To-Treat relevant?
When comparing effects in randomized controlled trials.
Why are we interested in Risk Reduction and Number-Needed-To-Treat?
Interested in understanding risk of treatment vs. no treatment
What are we asking in Risk Reduction and Number-Needed-To-Treat studies?
What is the frequency of bad outcomes in group being treated compared to the group not being treated?
Randomized Controlled Trials (RCT):
- Have at least one treatment group and one control group
- People in both groups may have positively (placebo effect) or negatively (harmful effects)
- How do we compare different group response rates?
What is Control Event Rate (CER)?
Proportion of control group participants who have a bad outcome after “treatment” (ex. placebo or no rx)
What is the CER if 10 of 30 control group participants become sicker?
CER = 10/30 = 33% have adverse outcomes
What is Experimental Event Rate (EER)?
Proportion of treatment group participants who have a bad outcome after treatment (ex. new drug)
What is the EER if 4 of 30 treatment groups become sicker?
EER = 4/30 = 13% had adverse outcomes
What is Absolute Risk?
“risk difference” = difference in risk of developing a DZ or undesired outcome after treatment
How do you calculate Absolute Risk?
CER-EER
What is an Absolute Risk Reduction (ARR)?
When CER > EER - higher rate of adverse outcomes in control group –> sometimes referred to as “attributable risk”
What is an Absolute Risk Increase (ARI)?
When EER > CER - higher rate of adverse outcomes in treatment group
What is Relative Risk?
“risk ratio” = proportion of treatment group risk to control group risk
How do you calculate Relative Risk?
EER/CER
How does risk of bad outcome change in the treatment group with RR?
Risk Increases when RR > 1
Risk Decreases when RR
What is Relative Risk reduction/increase?
Difference in 2 event rates, as a proportion of the event rate in the control group
What is the equation for Relative Risk Reduction/Increase?
1-RR or AR/CER
What is the equation for Relative Risk Reduction?
CER > EER
What is the equation for Relative Risk Increase?
EER > CER
What is Number Needed To Treat (NNT)?
Number of patients who need to be treated to get 1 additional patient a favorable outcome
What is the equation for NNT?
NNT = 1/ARR
Explain what NNT = 5 means?
For every 5 people treated, 1 more person would respond to the drug
What is Number Needed to Harm (NNH)?
Number of patients who, if they were treated, would result in 1 additional patient being harmed
How does NNH relate to ARI?
NNH = 1/ARI
Explain wheat NNH = 3 means?
If 3 people were treated, 1 more person would not respond compared with the control group.
