Biostats Flashcards
Relative risk equation?
RR = risk of an outcome in an exposed group/ risk of the same outcome in an non exposed group
What is a cross sectional analysis and what is it used for?
Data gathers at one point in time
Can determine prevalence of an outcome in a population
Identify specific risk factors and help generate hypothesis for further research
False negative equation?
FN = 1- sensitivity x number of pts who truly have the disease
True positive equation?
TP= Sensitivity x number of patients who truly have the disease
Meta analysis
Groups results of several trials to increase statistical power and provide overall pooled effect estimates
How to determine if CI is statistically significant
For odds ratio and relative risk the null value is 1 ( as long as CIS do not cross the null value it is statistically significant)
For difference the null value is 0
Two sample test
Statistical method to compare the means of 2 groups of subjects
P value in a study with a large number of pts
A smaller p value is used to decrease the risk of of false positive results
Precision of test
Precision ( reliability) is the ability of the test to reproduce identical results with repeated measurements
Health promotion
Is an e.g of 1o prevention
Enabling people to control their health and determinants
Observer bias
Occurs with the investigator’s evaluation is affected by preconceived expectations or prior knowledge leading to overestimation
Independent event probability
If you have multiple independent events you multiply the probability
Prospective cohort study
Patients are selected on their exposure status and followed up over time for development of disease
Risk equation
Risk of developing the disease with exposure/ all subjects who were exposed
Main purpose of blinding in a study is to prevent
Observer bias by either pt or researcher
Area under the curve
The more accurate a test is ( high sensitivity and specificity) the closer AUC value is to 1
The average of a dataset of values
Equivalent to mean
Total number of episodes/total number of pts
Allergic contact dermatitis features
Type IV hypersensitivity reaction
Intercellular epidermal oedema, with infiltration of lymphocytes and eosinophils
Reducing significance set alpha in a study will lead to
Any significant findings will be reported with greater confidence
Decreased significance level = increased confidence level
Cross sectional study
Exposure and outcome measured simultaneously at a particular point in time
Prevalence study
Positively skewed curve relationship between mean, median and mode
Mode is closer to the middle and mean closer to tail ( RT tail)
Median in the middle
Mode <median < mean
Attack rate equation
Number of people who contact an illness/ number of people who are at risk of contracting that illness
Prevalence can be increased by
Incidence or prolonged disease duration
Prevalence = incidence x disease duration
Normal distribution mean and median
In a normal distribution curve mean and median are equal therefore 50% of observations will lie below the mean and 50% above
Lead time bias
When a test diagnoses a condition earlier than conventional causing an apparent increase in survival time but no improvement on actual time to mortality
Negative predictive value
Probability that an individual doesn’t have the disease given a negative result
NPV= TN/ TN+FN
Misclassification bias
Incorrect categorization of subject re: their outcome status or both
Hawthorne effect
Tendency of the study subjects to change their behaviour as a result of awareness that they are being studied
Greatest preventive intervention in almost all patients
Smoking cessation
T test ( paired t test )
Compares the mean of 2 groups
Quantitative test
Precision and accuracy
High precision it consistently provides same or similar value
High accuracy test provides measurements see or similar to actual value
Absolute risk reduction
Event rate in control group - event rate in the treatment group
Event rate in treatment= # of event in treatment arm / # of subjects
Absolute risk increase
Difference in the rate (risk ) of the adverse event between treatment and control group
ARI= Rate AE treatment - Rate AE control
Number needed to harm
NNH= 1/ ARI
