RESS Flashcards
Define health equity:
Differences in the quality of health and healthcare across different populations.
Define health inequality:
Socio-economic differences in health outcome.
What is the epidemiological triad?
Time.
Person.
Place.
Describe the scientific method:
Observation.
Proposal of a hypothesis.
Testing of the hypothesis.
Rejected or not rejected.
Define a fact:
A statement confirmed to such a degree that disagreement would be perverse.
Define a scientific law:
Distillations of repeated observations or facts.
What is a hypothesis?
A testable statement that describes an observation.
What is the null hypothesis?
The hypothesis which you falsify. Assumes no effect.
What is the alternative hypothesis?
The opposite of the null hypothesis.
What are the three requirements of a health outcome?
Valid.
Reliable.
Responsive.
What do PROMs measure?
A patients health related quality of life.
What are the utility values given to a healthy and to a dead individual?
1: healthy.
0: dead.
How are QALYs calculated?
Utility value x survival data.
Which symbol is used to truncate words in literature searching?
*
Which symbol is used to signify a variable character in literature searching?
#
Which symbol is used to search for single character/no character in the middle of a word?
What is this commonly used for?
?
American/English variants.
In literature searching, what does separating your words with ADJ do?
Searches for both terms next to each other in the order typed.
In literature searching, what does separating your words with ADJ1 do?
Searches for both terms in either order.
In literature searching, what does separating your words with ADJn do?
Searches for both terms with up to (n-1) words between them.
Define incidence rate:
New cases arising in a period of time.
Number of new cases in a period/number at risk in population in a period.
How is the incidence calculation changed when the number at risk changes over time?
Number of new cases/Total person-time at risk.
What is prevalence?
How many people have the disease at a certain time.
Do epidemiologists prefer incidence or prevalence for etiological research?
Incidence.
Define case fatality rate:
Number of people who die from disease in period/Number of people with disease in period.
Define mortality rate:
Number of people who die from disease in period/Number of people who die in period.
How are data sets adjusted?
Calculating stratum specific rates and combining it with a weighted average.
What is categorical data?
Can only be divided into into distinct categories.
Distinguish between the two types of categorical data:
Nominal: no natural ordering (M vs F)
Ordinal: ordering present. (low, medium, high).
Differentiate between the two types of numerical data:
Discrete: only takes whole values.
Continuous: can take any value.
What is standard deviation a measure of?
Spread.
How is risk calculated?
Number of new cases/number at risk.
What is the risk ratio a measure of?
Relative risk.
How is risk ratio calculated?
(Exposed cases/All exposed). OVER (Unexposed cases/All unexposed).
How are the odds of an event calculated?
Probability of event/ probability event doesn’t occur.
Which type of study uses odds ratios?
Case-control study.
How is an odds ratio calculated?
(Exposed casesUnexposed controls). OVER (Unexposed casesExposed controls).
When are risk ratios used?
Cohort studies only.
When are the odds ratio and risk ratio similar?
Very rare diseases.
Which values of relative risk are associated with protective and harmful effects?
less than 1 is protective
1 is no effect
more than 1 is harmful
What does the highest point on a normal distribution curve represent?
Mean, mode and median.
What is the mean and s.d. of the standard normal distribution?
Mean: 1.
S.d: 0.
What percentage of values lie with 1 s.d. of the mean in the normal distribution?
68.2%.
What percentage of values lie with 2 s.d. of the mean in the normal distribution?
95.4%.
What percentage of values lie with 3 s.d. of the mean in the normal distribution?
99.7%.
What is standard error?
The spread of the sample means.
What is a confidence interval?
Range of values in which the true population mean is likely to lie.
When can the standard normal distribution be used to calculate confidence intervals?
When the s.d. of the population is known. OR.
When the sample is >200.
What is used to calculate confidence intervals when the sample size is below 200?
Student’s t distribution.
How are degrees of freedom calculated?
Sample size - 1.
Which confidence interval is most often used?
95%.
What is assumed when using a t statistic? (2).
Normal distribution.
Independant samples.
When is a statistically significant effect concluded?
Confidence interval is on one side of the mean.
P value is less than 0.05.
What is the p value?
Probability of obtaining the results of a test given that the null hypothesis (no effect) is true.
What is the Pearson Correlation Coefficient?
Measure of correlation between two numeric variables.
What does the value of the Pearson correlation coefficient lie between?
What is the boundary for a good correlation?
1 and -1.
0.7.
When is the Spearman rank correlation coefficient used? (3).
Data not normally distributed.
One or both variables are ordinal.
Small sample size.
What are the assumptions when using linear regression? (2).
Approx linear relationship.
Residuals must be normally distributed.
When is a chi-squared test used?
Testing an association between two categorical variables.
What are the conditions for use of a Chi squared test? (2).
Expected values should be more than 1.
3/4 cells should have expected value more than 5.
Which correction is used for Chi squared tests with small sample sizes?
Yates correction.
What is primary prevention?
Preventing future occurrence by removing cause.
What is secondary prevention?
Prevention by screening/ detection/ treatment.
What is tertiary prevention?
Prevention by treating clinical cases.
What is sensitivity?
Calculation?
How well a test detects a condition.
True positives over all people with disease.
What is specificity?
Calculation?
How well a test correctly excludes those without a condition.
True negatives over all people without disease.
What is positive predictive value?
Calculation?
Probability someone has condition if they test positive.
True positives over all positive tests.
What is negative predictive value?
Probability someone doesn’t have a condition if they test negative.
True negatives over all negative tests.
What is the accuracy of a screening test?
Proportion of all the tests that have given the correct result.
What is a ‘failure’ in survival data?
Leaving the study.
What is ‘censoring’ in survival data?
Leaving the study before the event occurs.
Differentiate between left and right censoring in survival datasets:
Right: people did not reach failure before end of the study.
Left: not certain what happened before entering study (e.g. already have disease).
In survival data, what is the ‘survival function’?
Chance of survival until a certain time.
In survival data, what is the ‘hazard function’?
Chances of instantaneous failure at any one time.
What is a Log-rank test used for?
Comparing the survival functions between two groups.
How is the average and spread of normally distributed data reported?
Mean.
Standard deviation.
How is the average and spread of skewed data reported?
Median.
Interquartile range.
How do you check on the distribution of numerical data?
Histogram.
What does an adjusted R squared value tell you?
What % of the variability in the data can be explained by the model.
What does a narrow confidence interval tell you?
Good precision.
What is opportunity cost?
Value of benefit lost from the option(s) you didn’t choose.
Differentiate between technical and allocative efficiency:
Tech: meeting objective at least cost.
All: production that matches consumer demand.
What is marginal analysis?
Compares the benefits and costs of taking the next step.
Marginal cost and benefit
What is equity?
Fairness of the distribution of the cost and benefits.
What outcome is used to measure cost benefit?
Monetary value.
What type of outcome is used to measure cost effectiveness?
What does it address?
Problems?
Natural units.
Technical efficiency.
Interventions with different outcomes can’t be compared.
What outcome is used to measure cost utility?
QALYs
What are the four possible conclusions of a NICE technology appraisal?
Recommended.
Optimised (smaller subset).
Only in research.
Not recommended.
What is the NICE threshold for cost effectiveness?
£20,000-£30,000
Where two treatments are compared, if A costs less and is better than B, what is A said to be?
Dominant.
Differentiate between multi-way and probabilistic sensitivity analysis:
Multi-way: vary more than one variable at a time.
Probabilistic: vary all parameters simultaneously based on probability decisions.
What does the cost effectiveness acceptability curve represent?
The probability that the Incremental Cost Effectiveness Ratio is at or below any given Maximum Willingness to pay threshold.
What are the 3 secondary uses of electronic health records?
Developing improvement programme.
Clinical governance.
Epidemiological research.
What are the problems with big data in health?
Quality. Size. Data linkage (coding). Confidentiality + security. Consent.
What is a health related state? (Outcome or exposure)
Outcome
What is validity?
Wether data accurate measures what it is meant to measure.
What is reliability?
Wether it will give the same result on retesting.
What does the p value tell you?
Give the test that you did, how likely is it that you saw those results by chance?
