Year 2 Brainscape Qs Flashcards
What is the formula for standard deviation? [1]
Square root of an average of all the individual points’ (distances from the mean)2
What measures of location and spread are used in a data contained in normal distribution? [2]
Mean and standard deviation
What measures of location and spread are used in the presence of outliers in the data? [2]
Median and interquartile range [2]
What change in standard deviation flattens the curve and what makes it taller and thinner? [2]
Flat: Increased SD
Tall and thin: Decreased SD
How many standard deviations from the mean
(+ and -) is 90%, 95% and 99% of population contained? [3]
90%:1.64 SD
95%: 1.96 SD
99%: 2.58 SD
How is standard error calculated and when is it used? [2]
SE = Standard deviation / sq root (sample size)
It is used as a standard deviation of the distribution of all possible sample means (used when dealing with sampling from the population)
How is confidence interval calculated? What does 95% CI = 21.4 to 22.6 mean BMI? [2]
95% CI = sample mean =/- 1.96 x standard error
It means that we are 95% sure that in the population the mean BMI can be as low as 21.4 and as high as 22.6.
What happens to CI if the sample size is increased? [1]
The confidence interval gets narrower
Which variable in the linear regression equation indicates outcome and which is the predictor? [2]
Y – outcome
X – predictor
What test is used to establish p-value while comparing two means? [1]
Two sample t-test
What are different types of clinical trials? [3]
Uncontrolled – everyone gets treatment
Controlled – treated group (‘new treatment’) vs untreated group (placebo or ‘standard treatment’)
Randomized controlled – allocation to groups is determined by chance
What are the types of blinding? [2]
Single blind: patients don’t know what treatment they’re getting
Double blind: both patients and observers (examining doctors) don’t know what treatment each patient is getting
What type of randomized controlled trail is used if the treatment is irreversible and which if it’s reversible? [2]
Irreversible: parallel group
Reversible: crossover
Name two disadvantages of crossover design? [2]
How can we compensate for one of them? [1]
Time consuming
Carry-over effects: washout period before second treatment
What is the difference between intention-to-treat and on-treatment analyses? [2]
Which one is better and why? [2]
TT: compare all subjects in treatment group regardless of whether they complied and acc took the drug
OT: compare subjects who acc took the treatment
ITT is better as it is a better representation of what would happen in real life
When is relative risk statistically significant? [1]
When 95% CI for that relative risk excludes 1
What would be the relative risk if the treatment didn’t have any effect?
RR = 1
What is the type of diagram used in meta-analyses? [1]
Forest plot
What diagram is used to assess publication bias of meta-analyses?
Funnel plot
Give overview of a design of cohort studies? [3]
What is the main question of cohort studies? [1]
Identify cohort
Measure exposure
Follow up over time to identify whether they have disease or not
Is the disease more common in exposed or unexposed?
What are the sources of bias in the cohort studies? [2]
Loss (of people) to follow-up
Measurement of exposure at only one point in time (it can change very rapidly); selection of cohort
How is relative risk measured in cohort studies? [1]
Incident of disease in exposed population / incidence of disease in unexposed population
What curve is used in survival analyses? [1]
What information does it provide? [1]
Kaplan-Meier curve; shows the proportion of people surviving over time
What is a hazard ratio? [1]
Risk of dying at any time point in one group compared to the other
How can the absolute excess risk be calculated? [1]
Risk in exposed – risk in unexposed
How is attributable proportion calculated? [1]
P(RR-1) / 1+p(RR-1) ; p= proportion exposed in population
What is the design of case-control studies? [3]
Identify those with disease [cases]
Measure exposure (from the past)
Identify those without disease [controls] and measure exposure (from the past)
What study design is useful for biomarkers studies where biomarkers are v expensive to measure? [1]
Nested case-control studies
What is the design of a nested case-control study design? [4]
Identify cohort and take & store blood sample
Follow up over time and detect if have disease.
All people with disease are cases and those without disease are selected for controls.
Measure exposure in cases and controls from the biomarkers.
What does odds ratio represent in rare disease? How do you calculate it? [1]
Relative risk; odds exposure in cases/odds exposure in control
How do you calculate absolute excess risk? [1]
(Odds ratio x risk of disease) – risk of disease
What does cross-sectional study measure? [1]
Measure existing disease and current exposure
Which study design has the biggest risk of bias? [1]
Cross-sectional surveys [1]
What study design has the bigger risk of confounding: cohort or case-control? [1]
Case-control studies
What study design would you use for rare disease? [1]
What study design would you use for rare exposure? [1]
Rare disease: case-control study
Rare exposure: cohort study
State the definition of incidence [1] and prevalence [1]
Incidence: Number of new cases of the disease occurring within a period of time
Prevalence: Number of cases of the disease that are present in a particular population during a period of time/particular date
What is the Declaration of Helsinki?
Ethical principles for medical research involving human subjects – promotes and safeguard health, wellbeing and rights of research. It’s the duty of physicians who are involved in medical research to protect life, health, dignity, integrity, right to self-determination, privacy and confidentiality of personal info of research subjects
List 5 types of cognitive bias [5]
Availability (disease seen most recently = most likely)
Representativeness (diagnosis more likely if more similar to typical cases)
Anchoring (perceived probability of diagnosis based on one trait);
Diagnosis momentum (once labelled it’s hard to go back); Fundamental attribution error (tendency to blame ppl for their illness and sometime; commission bias (action rather than inaction)
The reference range for 90% STD is
mean ± 1.69 SDs
mean ± 1.96 SDs
mean ± 1.64 SDs
mean ± 1.94 SDs
mean ± 2.58 SDs
The reference range for 90% STD is
mean ± 1.69 SDs
mean ± 1.96 SDs
mean ± 1.64 SDs
mean ± 1.94 SDs
mean ± 2.58 SDs
The reference range for 99% STD is
mean ± 1.69 SDs
mean ± 1.96 SDs
mean ± 1.64 SDs
mean ± 1.94 SDs
mean ± 2.58 SDs
The reference range for 90% STD is
mean ± 1.69 SDs
mean ± 1.96 SDs
mean ± 1.64 SDs
mean ± 1.94 SDs
mean ± 2.58 SDs
What value is changing between the different colours in this Gaussain Distribution?
Standard error
Standard deviation
Mean
Median
Inter quartile range
STD
What value is changing between the different colours in this Gaussain Distribution?
Standard error
Standard deviation
Mean
Median
Inter quartile range
Normal distribution aka? [1]
Gaussian distribution
What does 95% STD actually mean? [1]
95% of all values are within 1.96 standard deviations of a mean value
What does 95% STD actually mean? [1]
95% of all values are within 1.96 standard deviations of a mean value
Which value do you use to calculate 95% confidence interval of a sample mean?
Standard error
Standard deviation
Median
Inter quartile range
Which value do you use to calculate 95% confidence interval of a sample mean?
Standard error
95% CI = sample mean ± 1.96 × standard error
What does this describe?
a measure of the statistical accuracy of an estimate, equal to the standard deviation of the theoretical distribution of a large population of such estimates.
Standard error
Median
Mean
Confidence interval
Inter quartile range
Standard error
Which of the following do you use to calculate CIs?
SE
STD
SE - (for mean values)
Which of the following do you use to calculate reference ranges?
SE
STD
STD - (for individual values)
Which of the following are used to calculate confidence intervals?
Standard deviation and individual values
Standard deviation and mean values
Standard error and individual values
Standard error and mean values
Standard error and mean values
Which of the following are used to calculate reference ranges?
Standard deviation and individual values
Standard deviation and mean values
Standard error and individual values
Standard error and mean values
Which of the following are used to calculate reference ranges?
Standard deviation and individual values
When do you use Pearson vs Spearman correlation? [2]
Pearson correlation evaluates the linear relationship between two continuous variables.
Spearman correlation: Spearman correlation evaluates the monotonic relationship.
Which regression model would you use for the following?
Exponential growth
Exponential decay
Sigmoid
Polynomial – quadratic
Exponential growth
Which regression model would you use for the following?
Exponential growth
Exponential decay
Sigmoid
Polynomial – quadratic
Polynomial – quadratic
Mann Whitney U test would be used if Gaussian distribution is not followed and for which type of test?
One sample T-test
Paired T-test
Two sample T-test
Equal variance T-test
Mann Whitney U test would be used if Gaussian distribution is not followed and instead of which type test
One sample T-test
Paired T-test
Two sample T-test
Equal variance T-test
Wilcoxon signed rank test would be used if Gaussian distribution is not followed and for which type of test?
One sample T-test
Paired T-test
Two sample T-test
Equal variance T-test
Wilcoxon signed rank test would be used if Gaussian distribution is not followed and for which type of test?
One sample T-test
Paired T-test
Two sample T-test
Equal variance T-test
Explain the difference in the following
One sample T-test
Paired T-test
Two sample T-test
A one-sample t-test is used to compare a single population to a standard value (for example, to determine whether the average lifespan of a specific town is different from the country average).
A paired t-test is used to compare a single population before and after some experimental intervention or at two different points in time (for example, measuring student performance on a test before and after being taught the material).
Which of the following uses same person as control
Parallel group randomised controlled trial
Cross over parallel group randomised controlled trial
Cluster randomised controlled trial
Cohort controlled trial
Which of the following uses same person as control
Parallel group randomised controlled trial
Cross over parallel group randomised controlled trial
Cluster randomised controlled trial
Cohort controlled trial
Which of the following is used when treatment is reversible
Parallel group randomised controlled trial
Cross over parallel group randomised controlled trial
Cluster randomised controlled trial
Cohort controlled trial
Cross over parallel group randomised controlled trial
Which of the following is used when treatment is irreversible
Parallel group randomised controlled trial
Cross over parallel group randomised controlled trial
Cluster randomised controlled trial
Cohort controlled trial
Parallel group randomised controlled trial
Investigating the efficacy of chemotherapy to treat cancer would best be studied using a
Parallel group randomised controlled trial
Cross over parallel group randomised controlled trial
Cluster randomised controlled trial
Cohort controlled trial
Parallel group randomised controlled trial
Which should be designed as a cluster randomised trial?
Trial of aspirin versus placebo to prevent pre-eclampsia
Trial of addition of flocculant disinfectant to drinking water in the prevention of diarrhoea
Trial of text message reminders versus standard care to improve drug adherence
Trial of addition of flocculant disinfectant to drinking water in the prevention of diarrhoea
Label A-C
A: parralel
B: cross over
C: factorial
Explain the following:
NNT of 9 for treating sinusitis with antibiotics
NNT of 9 for treating sinusitis with antibiotics
indicates that 9 patients need to be given anantibiotic to get one more patient better than would have improved without the antibiotics.
How can you tell if 95% CI difference for means [1] and for relative risk [1] are significant in a cohort study [2]
If 95% CI difference is means includes 0, result is not statisitically significant; if excludes 0 then is sig.
If 95% CI difference is relative risk includes 1, result is not statisitically significant; if excludes 1 then is sig.
Reverse causality is a problem in
cohort trial
case-control trial
randomised control trial
cross sectional trial
case-control trial
I want to investigate a risk factor for a rare disease. Which should you use?
cohort trial
case-control trial
randomised control trial
cross sectional trial
case-control trial
I want to investigate if a rare exposure is associated with a disease ? Which should I use and why?
cohort trial
case-control trial
randomised control trial
cross sectional trial
cohort trial: are particularly advantageous for examining rare exposures because subjects are selected by their exposure status
Odds ratios are used instead of relative risk in which type of trial
cohort trial
case-control trial
randomised control trial
cross sectional trial
case-control trial
Does exposure to blue asbestos cause lung cancer - What study should you do?
Case-control
Cohort
Clinical trial
Cross-sectional survey
Cohort: rare exposure
Is prevalence of back pain more common in men or women ?
Case-control
Cohort
Clinical trial
Cross-sectional survey
Cross-sectional survey
Place the following in order of most to least likely to have bias
Case-control
Cohort
Clinical trial
Cross-sectional survey
In a randomised controlled trial of preventive PCI versus infarct artery only PCI in patients with ST elevated myocardial infarction, 10% of patients randomised to receive preventive PCI died, and 30% of patients randomised to receive infarct artery only PCI died.
Calculate the number needed to treat (NNT) with preventive PCI to prevent one death. Show your calculations. [1]
Absolute difference in risk = 30%-10%=20%
NNT = 100/20= 5
If a variable follows a Gaussian distribution then how many standard deviations either side of the mean would we expect 95% of values to lie? [1]
1.96
Tick(+-1.96, 1.96, 1.96 sd, 1.96 standard deviations, ± 1.96)
A randomised placebo controlled trial of a new treatment showed that the risk of death in the new treatment group was 5% and the risk of death in the placebo group was 10%.
What is the number needed to treat with the new treatment to prevent one death?
20
Tick (20 people, 20, 20%)
What do A & B represent on this forest plot? [2]
A: Result of study
B: 95% CI of result
How do you determine the size of a study from a forest plot? [1]
This black box also gives a representation of the size of the study. The bigger the box, the more participants in the study.
What does the diamond represent on the forest plot? [1]
The diamond represents the point estimate and confidence intervals when you combine and average all the individual studies together.
Describe the differences in horizontal and box size in forest plots in samll versus big studies [2]
The bigger the study, the smaller the horizontal line and bigger the black box representing the point estimate.
The smaller the study, the wider the horizontal line and smaller the black box representing the point estimate.
The forest plot shows the results of randomised trials of aspirin versus placebo in the prevention of preeclampsia.
Which was the biggest trial?
Ebrashy: has the biggest box
.Measures of inequality are important to track patterns and trends in ill-health in order to prioritise service provision.
Please show your understanding of these issues by labelling the following graph.
Use the following labels:
Healthy life expectnacy, multi morbidities percentile, race, gender, healthcare access, neighbourhood deprivation percentiges
Label A&C of this funnel plot [2]
A: 95% CI
B: Study precision
C: Overall effect
D: Study result
Describe the results of this funnel plot [1]
Asymmetrical: indicates bias
What study types is this graph often used in?
A. Cohort study
B. Case control study
C. Clinical trial
D. Cross-sectional survey
A. Cohort study
B. Case control study
C. Clinical trial
D. Cross-sectional survey
Kaplan-Meier curves
Plot proportion of people surviving over time
Compared using a log rank test
Used in cohort studies and randomized trials.
What happens to the confidence interval as sample size increases?
A. Doubles
B. Stays the same
C. Gets bigger
D. Gets narrower
D. Gets narrower
If the sample size increases, the 95% confidence range stays the same;
however the 95% confidence interval gets narrow since the standard error decreases if the sample size is larger
There are different types of analysis used within statistics. What is a t-test used to compare?
A – Medians of a continuous variable between 2 groups
B – Proportions of a binary variable between 2 groups
C – Proportions of binary variable within one group, measured at 2 time points
D – Standard deviations of a continuous variable between 2 groups
E – Means of a continuous variable between 2 groups
There are different types of analysis used within statistics. What is a t-test used to compare?
A – Medians of a continuous variable between 2 groups
B – Proportions of a binary variable between 2 groups
C – Proportions of binary variable within one group, measured at 2 time points
D – Standard deviations of a continuous variable between 2 groups
E – Means of a continuous variable between 2 groups
Which of the following best describes ‘measure of a statistical accuracy of an estimate’
Standard deviation
Confidence interval
Standard error
Median
Standard error
This trial demonstrates a
case-control study
cohort study
factorial study
cross sectional study
case-control study
This trial demonstrates a
case-control study
cohort study
factorial study
cross sectional study
cohort study
Why are case control studies useful for general population? [1]
Because RR & OR are generally robust to different populations - to everyone who might or might not smoke.
One-half of a person’s ability to recover from illness is determined by [] factors? [1]
One-half of a person’s ability to recover from illness is determined by socioeconomic factors
A study was based on volunteers. Of the three main types of systematic errors
in public health research, what type of bias is most typically associated with the use
of volunteers?
information bias
selection bias
confounding bias
random error
A study was based on volunteers. Of the three main types of systematic errors
in public health research, what type of bias is most typically associated with the use
of volunteers?
information bias
selection bias
confounding bias
random error
Confidence intervals are a good way to deal with
information bias
selection bias
confounding bias
random error
random error
P-values are a good way to deal with
information bias
selection bias
confounding bias
random error
random error
Objective
To assess the impact of breast feeding on the risk of being overweight in children at the time of entry to school
Methods
A group of overweight children were identified when they started school and the same number of normal weight children at the same school were identified.
Their mothers were asked about breastfeeding
Which of the following best describes this trial?
case-control study
cohort study
factorial study
cross sectional study
case-control study
Objective
To assess the impact of breast feeding on the risk of being overweight in children at the time of entry to school.
Methods
Mothers asked about breast feeding shortly after delivery and at 6 months.
At 5 years children were assessed physically.
Which of the following best describes this trial?
case-control study
cohort study
factorial study
cross sectional study
cohort study
Interpret this result:
After adjusting for age, sex and tumour stage the Hazard Ratio for exercisers vs non-exercisers = 0.73
People who exercised prior to developing colorectal cancer had a 27% lower risk of mortality than non-exercisers at any time during follow-up
If want to investigate a study quickly?
case-control study
cohort study
factorial study
cross sectional study
case-control study
Recall bias is a problem in these studies
case-control study
cohort study
factorial study
cross sectional study
case-control study
Which of the following can you not calculate risk in?
case-control study
cohort study
factorial study
cross sectional study
Which of the following can you not calculate risk in?
case-control study
Can’t use relative risk as do not know risk of the disease (as you have started with cases with the disease)
Which of the following woudl you calucate an odds ratio and not a relative risk in?
case-control study
cohort study
factorial study
cross sectional study
Which of the following woudl you calucate an odds ratio and not a relative risk in?
case-control study
Can’t use relative risk as do not know risk of the disease (as you have started with cases with the disease)
A non-randomised study looking at the association between preventive percutaneous coronary intervention (PCI) versus infarct artery only PCI in patients with ST elevated myocardial infarction was carried out. It found that in 280 patients receiving preventive PCI, 80 died after 5 years of follow-up. Of 120 patients receiving infarct artery only PCI 20 died after 5 years of follow-up.
Calculate the odds ratio of death in the preventive PCI group compared with the infarct artery only PCI group. Show your calculations.
PPCI (case): total =280; died = 80
IPCI (control): total = 120; died = 20
Odds of preventive PCI in patients that died = 80/20 = 4
Odds of preventive PCI in those that did not die = 200/100 = 2
Odds ratio = 4/2 = 2
State the reason for association
Randomised placebo controlled trial of candsartan in patients with chronic heart failure found it increased risk; RR = 1.6 (95% CI 1.1-2.2)
Recall bias
Reverse Causality
Chance
Confounding
Selection bias
chance
State the reason for association
A case control study of stomach cancer patients found that they had a higher antacid usage in the 6 months prior to diagnosis than did the controls.
Recall bias
Reverse Causality
Chance
Confounding
Selection bias
Reverse Causality
People with undiagnosed stomach cancer tend to have the same symptoms for which antacids are taken
State the reason for association
In a cohort study heavy drinkers were found to be more likely to develop lung cancer. Does drinking cause lung cancer ?
Recall bias
Reverse Causality
Chance
Confounding
Selection bias
Confounding
Heavy smokers are more likely to be heavy drinkers.
Smoking causes lung cancer
State the reason for association
A case control study of patient’s with Chron’s disease showed that they were much more likely to have reported ever eating Corn flakes.
Recall bias
Reverse Causality
Chance
Confounding
Selection bias
Recall bias
Those with Chron’s disease more likely to remember everything they ate. Most people have eaten corn flakes at some time.
State the reason for association
Non-randomised studies have shown an increased risk of death in preventive (multivessel) percutaneous coronary intervention versus infarct artery only PCI in patients with STEMI
Recall bias
Reverse Causality
Chance
Confounding
Selection bias
Selection bias
Patients who had preventive PCI more likely to be sicker
Randomised controlled trials have shown the benefit of preventive PCI
