Notes Flashcards
What do the lines ‘in the box’ show as part of a box plot?
The median and the two quartiles (1st and the 3rd)
What is the risk ratio?
The risk ratio is the probability of an event occurring in an exposed group compared to the probability occurring in a non-exposed group.
Which studies commonly use the odds ratio as a measure of association? [1]
Odds ratios are most frequently used in case-control studies, but may also be used in cross-sectional studies.
What is a type 2 error?
The failure to reject the null hypothesis when it is false
- false negative
What is the attributable risk?
The attributable risk is another measure of association, which is defined as the amount of proportion of disease incidence (disease risk) that can be attributed to the exposure.
What is a type 1 error?
The incorrect rejection of the null hypothesis when it is true
- false positive
What is the main disadvantage of cross-sectional studies?
Cross-sectional studies do not show causality or account for confounding factors which may also influence the data. Therefore, associations identified may be difficult to interpret.
How do you calculate relative risk?
The risk of poor outcomes in the intervention group divided by the risk of poor outcomes in the control group.
1
When do you use a Mann-Whitney U test? [1]
Non-parametric data, two groups
When do you use a Wilcoxon signed rank test? [1]
Non-parametric data, one group
When do you use a Kruskal-Wallis test? [1]
Non-parametric data, more than 2 groups
Non-parametric; more than two groups
Paired T-test
Kruskal Wallis Test
Mann Whitney U test
Wilcoxon test
Non-parametric; more than two groups
Paired T-test
Kruskal Wallis Test
Mann Whitney U test
Wilcoxon test
Non-parametric; two groups
Paired T-test
Kruskal Wallis Test
Mann Whitney U test
Wilcoxon test
Mann Whitney U test
Non-parametric; one group
Paired T-test
Kruskal Wallis Test
Mann Whitney U test
Wilcoxon test
Non-parametric; one group
Paired T-test
Kruskal Wallis Test
Mann Whitney U test
Wilcoxon test
Which tests do you use for parametric [3] and non-paremetric [3]
one group two groups multiple groups
When do you use Fisher’s exact test instead of Chi-Squared? [1]
When 2x2 table has date with data with < 5 in it
If you want to measure a binary outcome measured at two time points with a group of indivuals, which test do you use? [1]
McNemars test
Describe how you calculate risk ratio [1]. e.g. using
When would you use the following types of regression:
- Ordinal logistic
- Poission
- Tobit
- Cox’s proportional hazards
How do you calculate the Specificity of a result? [1]
True Negative / (True Negative + False Positive)
95+255 = 350 / (350 + 45)
How do you calculate sensitivity? [1]
True positive / (true positive + false negative )
How do you calculate the PPV of a result? [1]
True +ve / (TP + FP)
Negative predictive value?
TN / (TN + FN)
What is specificity? [1]
Specificity is the chance of someone without the disease testing negative.
e.g. A 68-year-old woman is being investigated for Helicobacter pylori infection. She is told that the specificity of the test is 89.8%.
- 89.8% of patients without H. pylori test negative
Work out by: TN / (TN + FP)
A question asks you to compare % and / or proportions - which test should you use? [1]
Chi-squared test - used to compare proportions or percentages e.g. compares the percentage of patients who improved following two different interventions
Patients were randomly allocated to either receive the new NRT or a placebo inhaler. The frequency of their cravings was assessed after a two-week period using the following scale: ‘Never’, ‘Rarely’, ‘Sometimes’, ‘Often’, ‘Always’.
Test used?
Mann-Whitney U test - compares ordinal, interval, or ratio scales of unpaired data
compares the percentage of patients who improved following two different interventions
Which test? [1]
Chi-squared test
How do you work out NNT? [+]
e.g.
A study looks at adding a new antiplatelet drug in addition to aspirin to patients who’ve had a stroke. One hundred and seventy patients are enrolled for the study with 120 receiving the new drug in addition to aspirin and the remainder receiving just aspirin. After 5 years 18 people who received the new drug had a further stroke compared to 10 people who just received aspirin. What is the number needed to treat?
- Work out absolute risk reduction: Control event rate - experimental event rate
- 1/absolute risk reduction
Control event rate = 10 / 50 = 0.2
Experimental event rate = 18 / 120 = 0.15
Absolute risk reduction = 0.2 - 0.15 = 0.05
Number needed to treat = 1 / 0.05 = 20
What does a Kaplein-Meier survival plot show? [1]
The Kaplan-Meier survival plot displays estimates of decreasing survival with time after an event
non-parametric & compares two sets of observations on a single sample, e.g. a ‘before’ and ‘after’ test on the same population following an intervention ?
Wilcoxon signed-rank test
How do you calculat variance of a sample? [1]
Variance = square of standard deviation
Variance is a measure of the spread of scores away from the mean
What is the relationship between standard deviation and variance?
Standard deviation = mean x variance
Standard deviation = mean - variance
Standard deviation and variance are the same
Standard deviation is the square of variance
Standard deviation is the square root of variance
What is the relationship between standard deviation and variance?
Standard deviation = mean x variance
Standard deviation = mean - variance
Standard deviation and variance are the same
Standard deviation is the square of variance
Standard deviation is the square root of variance
compares a group with a disease to a group without, looking at past exposure to a possible causal agent for the condition
Case control
compares two sets of observations on a single sample, e.g. a ‘before’ and ‘after’ test on the same population following an intervention
Wilcoxon signed-rank test
Type II error = ? [1]
Type II error - the null hypothesis is accepted when it is false
Likelihood ratio for a positive test result = []
Likelihood ratio for a positive test result = sensitivity / (1 - specificity)
What does the power of the study equate to? [1]
Power = 1 - the probability of a type II error
How do you calculate the liklihood ratio of a positive [1] and negative [1] event?
Liklihood ratio of positive event:
- Sensitivity / (1-specificity)
Liklihood ratio of negative event:
- (1-Sensitivity) / specificity
What type of calculation is used for a cohort study [1] and case-control study? [1]
Cohort: relative risk
Case control: odds ratio
How do you calculate standard error? [1]
~~~
STD / Sqr(number of participants)