Statistics Flashcards
How do you calculate the positive likelihood ratio?
Positive likelihood ratio = sensitivity / (1 - specificity)
Describe what percentage of people like within 1, 2, and 3 standard deviations from the mean
- 1 standard deviation account for 68% of the values
- 2 standard deviations account for 95% of the values
- 3 standard deviations account for 99.7% of the values
Describe the positive predictive value and how to calculate it
- If the test is positive, what is the chance the patient has the disease?
- PPV = true +ve/true+ve and false +ves
Describe the negative predictive value and how to calculate it
- If the test is negative, what is the chance the patient doesn’t have the disease?
- NPV = true -ve/true-ve and false -ves
Describe sensitivity and how to calculate it
- True positive rate
- Sensitivity = true +ve/true +ve and false -ves
- A negative result in a test with a high sensitivity is useful for ruling out a disease
Describe specificity and how to calculate it
- True negative rate
- Specificity = true -ve/true -ve and false +ves
- A positive result in a test with a high specificity is useful for ruling in a disease
What is the value of a likelihood ratio?
- Assess the value of performing a diagnostic test
- Uses sensitivity and specificity of the test to determine whether a test result usefully changes the probability that a condition exists
- If you have a pre-test probability and a likelihood ratio, you can work out the post-test probability (+ve LR increases this, -ve LR decreases this)
Describe the positive likelihood ratio and how to calculate it
- Likelihood ratio of a positive test is the probability of a true positive to a false positive
- +ve LR = sensitivity/100% - specificity
Describe the negative likelihood ratio and how to calculate it
- Likelihood ratio of a negative test is the probability of a false negative to a true negative
- -ve LR = 100% - specificity/ sensitivity
What is the difference between incidence and prevalence?
- Incidence is a rate e.g. cases per 5 years
- Prevalence is a cross section in time e.g. point prevalence studies
What is the most important property of a screening test?
Sensitivity
What are the criteria for a screening test (Wilson’s criteria)?
- The condition sought should be an important health problem.
- There should be an accepted treatment for patients with recognized disease.
- Facilities for diagnosis and treatment should be available.
- There should be a recognizable latent or early symptomatic stage.
- There should be a suitable test or examination.
- The test should be acceptable to the population.
- The natural history of the condition, including development from latent to declared disease, should be adequately understood.
- There should be an agreed policy on whom to treat as patients.
- The cost of case-finding (including diagnosis and treatment of patients diagnosed) should be economically balanced in relation to possible expenditure on medical care as a whole.
- Case-finding should be a continuing process and not a “once and for all” project.
What is the relative risk and how do you calculate it?
- The probability of an event occurring in the exposed group vs. the non exposed group (event occurring/ total events)
- RR = a/ (a+b)
————
c/ (c+d)
What is the odds ratio and how do you calculate it?
- The event occurring: event not occurring
- e.g. OR of a baby girl is 1:1, probability (RR) is 50%
- Used when population risk is unknown in case-control studies. With low prevalence, OR~RR
- OR = ad/cb
What is the absolute risk and how do you calculate it?
- The real difference in absolute terms between the two exposure or treatment groups
- Same as the risk difference
- e.g. if death occurs in 20% of placebo and 10% of drug, then RRR = 50%, but ARR = 10%
What is the number needed to treat and how do you calculate it?
- The number of people needed to be treated to avoid one adverse outcome
- NNT= 1/ (incidence of unexposed – incidence of exposed)
- NNT = 100%/ARR = 1/ (probability in exposed - probability in non-exposed)
What is the number needed to harm and how do you calculate it?
- NNH = 100%/ARI (absolute risk increase)
What is a type 1 error?
= alpha error = false positive = p value
- Rejecting a null hypothesis when it is actually true
- Usually 0.05
What is a type 2 error?
= beta error = false negative
- failing to reject the null hypothesis when it should have been rejected
What increases the risk of a type 2 error (rejecting a good treatment)?
= beta error = false negative
= failing to reject the null hypothesis when it should have been rejected
= risk when the study is underpowered
What is the power of a study and what is it determined by?
- The power is the probability that the test will reject a false null hypothesis
- Power = 1 - beta error (false negatives) = sensitivity
- Usually 0.8 or 0.9
- Determined by: population size, size of effect, variance within populations
Describe the phases of drug design
Stage 0 - first in human studies, looks at pharmacokinetics and pharmacodynamics in humans, single sub-therapeutic dose given
Stage 1 - safety, healthy volunteers, determines safe dosing ranges, identify adverse events
Stage 2 - safety and efficacy (placebo), dosing requirements. Some combine stage 1 + 2
Stage 3 - confirmation of safety and efficacy, compare to active treatment (assess gold standard), RCT, requires approval from FDA
Stage 4 - continual pharmacovigilance (post marketing surveillance), new uses/populations, children, larger populations but less controlled, longer follow-up, interactions with other meds
What are types of observational studies?
- Cohort - divide on exposure yes or no, prospective or retrospective
- Case-control - divide on disease yes or no
- Cross-sectional e.g. community survey
- Longitudinal - follow over time
- Ecological - 1+ variable measured at a group level
What is the difference between internal and external validity?
- Internal - study well designed and minimises possibility of systematic bias (randomisation, blinding, ITT, published protocol, allocation concealment)
- External - applicable to a different unit/environment (multi-center, wide inclusion, limited exclusion criteria)
