Critical Numbers Flashcards
What is a case control study?
identify individuals with a particular outcome
retrospectively look back to see if they had the risk factor in question
non randomised
observational
retrospective
advantages of case controls
good for rare outcomes
fast as uses past data so no need for long follow up
cheaper
disadvantages of case controls
difficult to prove causation
prone to biases
not ideal for rare exposures
what is a cross sectional study?
collect data from many individuals at a moment in time
non randomised
observational
advantages of cross sectional studies
can assess multiple exposures/ outcomes
relatively quick
cheap
disadvantages of cross sectional studies
not ideal for rare exposures/ outcomes
susceptible to bias
cannot prove causality
what is a RCT
randomly allocate participants to different interventions and follow up
experimental
prospective
advantages of RCTs
gold standard - randomisation reduces potential for confounding
can determine causality
can reduce bias via control and blinding
ABC of strengths - allocation at random, blinding, control
disadvantages of RCTs
randomisation can be unfeasible or unethical when evaluating harmful exposures
require expert management and oversight for high risk interventions
resource intensive and expensive
strict eligibility criteria may mean sample not representative
what is a cohort study
the individuals in the sample may or may not have the exposure in question
after a period of follow up, the number of people who develop an outcome are recorded
non randomised
observational
typically prospective
follow up over time
advantages of cohort studies
useful when random allocation not possible
can work on rare exposures
can examine multiple outcomes
cons of cohort studies
long follow up
not ideal for rare outcomes
can be expensive
what is an ecological study
the unit of observation is the group rather than the individual e.g electoral ward, country
pros of ecological studies
large scale comparisons
can quantify geographical or temporal trends
cons of ecological studies
ecological fallacy
cannot make inference at the individual level from data at the group level
what is a systematic review?
give one strength and one weakness
research article in which existing evidence on a topic is systematically identified, appraised and summarised according to predetermined criteria
transparent, systematic methods make the process replicable
publication bias
what is a meta-analysis?
statistical synthesis of the evidence
effect sizes from each individual study are combined to create a single overall effect size
shown on forest plot
what is variation between studies called?
heterogeneity
quantified using a Q or I^2 statistic
describe the hierarchy of evidence from top to bottom
systematic review/ meta analysis
RCT
cohort study
case control study
cross sectional study
case study/ expert opinion/ anecdote
what is a sample
a subset of individuals from a population (should be representative of the population of interest, but isn’t always!)
what are generisable results?
representative of the population of interest
when is a sample biased?
certain subgroups of the population are over/ underrepresented in the sample
what is bias?
imperfections in the research process cause findings to deviate from the truth
what is sampling bias?
sample does not represent population of interest
what is recall bias?
inaccurate recall of past events/ exposures/ behaviours
what is information bias?
incorrect measurement e.g miscalibrated machine
what is the Hawthorne effect?
participants change their behaviour when they know they are being watched
attrition bias
differential dropout from studies e.g sicker participants drop out so the outcome is only measured on healthier participants
what are confounders
variables that obscure the real effect of an exposure on an outcome
related to both exposure and outcome
what is the most important part of a study?
design!
what is an experimental study
researchers have intervened in some way
what is an observational study?
researchers have observed without intervening
what are the three divisions of observational studies?
retrospective, cross sectional and prospective
what is a retrospective study?
looking back in time
what is a cross sectional study?
single snapshot in time
what is a prospective study?
following up over time
what is simple random sampling?
each member of the population has an equal probability of being selected
what is an ecological study?
unit of observation is the group rather than the individual
give 5 types of sampling
random, systematic, quota, cluster, stratified
describe random sampling
using a random number number generator
describe systematic sampling
researchers select members of the population at a regular interval
describe quota sampling
non-probability sampling method that relies on the non-random selection of a predetermined number or proportion of units
describe cluster sampling
divide a population into clusters, such as districts or schools, and then randomly select some of these clusters as your sample
describe stratified sampling
researchers divide subjects into subgroups called strata based on characteristics that they share (e.g., race, gender, educational attainment). Once divided, each subgroup is randomly sampled using another probability sampling method.
give some examples of bias
sampling, recall, social desirability, information, volunteer, selection, lead time bias, length time bias
what is sampling bias?
occurs when some members of a population are systematically more likely to be selected in a sample than others
what is recall bias?
systematic error that occurs when participants do not remember previous events or experiences accurately or omit details
what is social desirability bias?
respondents conceal their true opinion on a subject in order to make themselves look good to others.
what is information bias?
key study variables are inaccurately measured
what is volunteer bias?
arises in any research study in which participants choose if they want to be part of the sample
what is selection bias?
distortion in a measure of association (such as a risk ratio) due to a sample selection that does not accurately reflect the target population
which two types of biases are associated with screening?
lead time and length time biases
what is lead time bias?
occurs when a disease is detected by a screening or surveillance test at an earlier time point than it would have been if it had been diagnosed by its clinical appearance
what is length time bias?
overestimation of survival duration due to the relative excess of cases detected that are asymptomatically slowly progressing, while fast progressing cases are detected after giving symptoms
what are confounding factors?
related to outcome and exposure
what is critical appraisal?
the process of systematically examining research to judge its trustworthiness, and its value and relevance in a particular context
the process of assessing and interpreting evidence by systematically considering its validity, results and relevance to your own context
what is evidence based medicine?
the conscientious explicit and judicious use of current best evidence in making decisions about the care of individual patients
what is reliability?
the overall consistency of a measure
a measure is said to have a high reliability if it produces similar results under consistent conditions
what is validity?
extent to which a concept is accurately measured
what is internal validity?
accurately measuring those within the trial
what is external validity?
generalisable outside the trial
what are the 4 components/ questions of CASP?
- is the basic study design valid for a randomised controlled trial?
- was the study methodologically sound?
- what are the results?
- will the results help locally?
key considerations of RCTs
randomisation
- were key confounders included to reduce bias
- was the allocation sequence concealed
attrition
- exclusion bias
- drop out or lost to follow up
confidence in results
- are the results clinically important
- do the effect estimates include confidence intervals
what is PICO and what does it stand for?
a way of generating research questions
P - population
I - intervention
C - comparator
O - outcome
what is a variable
anything that varies within a dataset
what are the types of variables?
categorical, numerical
what are the types of categorical variables?
binary - only two categories e.g yes/ no
ordinal - categories with natural order e.g social class
nominal - categories with no natural order e.g hair colour
what are the types of numeric variables?
discrete - observations can only take certain numerical values
continuous - observations can take any value within a range (the only restriction is the precision of the measurement tool)
what is the mean?
add all the numbers in a data set and divide by the number of values
what is the median?
found by ordering the data points and selecting the middle number
what is the mode?
most frequent number
which measures of central location are the same in a normal distribution?
mean, median and mode
what is the standard deviation?
describes the variation of observations in our sample around the mean
also need to know the equation
used if we want to create a normal or reference range
(spread of data, describes spread around mean)
how do we calculate standard deviation?
calculate difference between each observation and the mean
square them to make them positive
sum then
divide by the number of observations minus one
take the square root to reverse the earlier squaring
which quartiles give the central 50% range
interquartile range 25th to 75th centile
how do we find the interquartile range?
halve data
difference between median of first half and median of second half
what is the standard error?
quantifies the precision of an estimate of the mean value
standard deviation of the sampling distribution
also need to know equation
used when we want to create a confidence interval around a point estimate
(spread of means, estimates real mean)
what affects the standard error?
how variable the data is
sample size
how much data is within one, two and three standard deviations of the mean?
68%, 95%, 99.7%
what is skew?
data is not normally distributed
what would a negatively skewed bell curve look like?
what are the relative mean, median and mode?
peak is further to the RIGHT (seems unusual, but remember that the tail is on the negative side of the graph)
mode>median>mean
what would a positively skewed bell curve look like?
what are the relative mean, median and mode?
peak is further to the LEFT (seems unusual, but remember that the tail is on the positive side of the graph)
mean>median>mode
what is kurtosis?
vertical skew
for skewed distributions, do we use the mean or median?
median as it has half of the data points on each side
what is the IQR?
measure of spread
used in conjunction with median
expressed as UQ - LQ
used when data isn’t normally distributed
what are confidence intervals?
range of values our population mean is likely to lie in
describes variability around data
how do we calculate the 95% confidence intervals?
mean +- (1.96 x SE)
requires you to know the equation for standard error!
how to construct a box and whisker chart
median is middle line
next two lines outwards are LQ and UQ
boundaries are LQ - (1.5 x IQR) and UQ + (1.5 x IQR)
outliers represented by crosses
when are scatterplots used?
display 2 continuous variables
used to assess correlation and regression
what is risk and what range of values can it take?
number with an outcome divided by the total number
synonym for probability
0-1
what is the absolute risk difference?
the difference between 2 risks
what does the absolute risk difference give us?
the number needed to treat/ harm
this is the number of patients you need to treat to prevent one additional bad outcome
formula for number needed to treat/ harm
1/ARD
what is relative risk (risk ratio) and how is it calculated?
the ratio between two risks
divide one risk by the other
what does a risk ratio of less than one indicate?
lower risk
what does a risk ratio of more than one indicate?
higher risk
what are odds?
the number with an exposure or outcome divided by the number without
what is the odds ratio?
odds in one group divided by the odds in the other
what does it mean if there is an odds ratio of 1?
there is no between-group difference
why use odds over risk?
odds are symmetrical
the odds ratio for outcome Y is the inverse of the odds ratio for outcome not Y
risk ratios lack this symmetry
do case control studies estimate odds or risk?
odds
what is sensitivity and how is it calculated?
ability of a test to detect true positives
ability of a test to correctly identify individuals with the disease
number of true positives successfully identified / actual number of positives
what is specificity and how is it calculated?
ability of a test to successfully exclude negatives
ability of a test to correctly identify those who do not have the disease
number of true negatives successfully identified / actual number of negatives
what is positive predictive value and how is it calculated?
the proportion of people with a positive test who actually have the disease
number of true positives successfully identified/ number of positives identified in total
what is negative predictive value and how is it calculated?
the proportion of people with a negative test who are correctly excluded by the screening test
number of true negatives/ number of negatives identified in total
what is test accuracy and how is it calculated?
measures the ability of a test to detect a condition when it is present and detect the absence of a condition when it is absent
number of successfully identified negatives or positives/ total number of people tested
what is prevalence and how is it calculated?
proportion of people in a population who have a particular disease or attribute at a specified point in time
number of people who have the disease (including false negatives)/ total number tested
what are the two hypotheses?
null - H0 and alternative - H1
what is the significance level?
determines whether a result is statistically significant
also the probability we incorrectly reject the null hypothesis
what is the usual value of the significance level?
0.05
what does a p value indicate?
probability of obtaining result or a result more extreme if the null hypothesis is true
(probability result is due to chance)
when do we reject the null hypothesis?
when the p value is lower than the significance level
what do the confidence intervals indicate?
plausible range for a variable
precision of an estimation
statistical and clinical significance (although these are not the same)
what is the relationship between the significance of a result, the null value (0) and the confidence intervals?
a result is significant at the 5% level if the 95% confidence interval does not include the null value
a result is significant at the 1% level if the 99% confidence interval does not include the null value
and so on
for skewed distributions, do we use the mean or median?
median as it has half of the data points on each side
what percentage of data points lie within
a) one standard deviation of the mean
b) two standard deviations of the mean
c) three standards deviations of the mean
a) 68%
b) 95%
c) 99.7%
what is correlation?
measure of the linear relationship between variables
what letter represents the correlation coefficient?
r
what range of values can the correlation coefficient take?
-1 to 1
what is the difference between descriptive and inferential statistics?
descriptive statistics relate to the sample
inferential statistics relate to the population
what is the central limit theorem?
when taking repeat samples of a population and calculating the mean, the sample means will be normally distributed around the true population mean
what is the difference between standard deviation and standard error?
standard deviation describes the variation of observations in our sample
standard error quantifies the precision of an estimate of a population parameter
when do we use the standard deviation and when do we use the standard error?
standard deviation - create a normal or reference range
standard error - create a confidence interval around a point estimate
what is the difference between correlation and regression?
correlation quantifies the linear relationship between two numeric variables
order does not matter
regression allows one variable to be predicted from another
quantify associations between exposures and outcomes
order matters (e.g predict y from x)
can handle multiple predictors
what equation does regression take its form in?
y = a + bx (straight line graph)
what do each of the variables of the regression line mean?
y
a
b
x
y - variable being predicted (dependent variable)
a - y intercept of the regression line (the constant)
b - regression coefficient (gradient of the regression line)
x - the predictor (independent variable)
how can we tell if a factor is a significant predictor of another?
p value is less than significance level
what is the difference between uninvariable and multivariable regression?
univariable regression looks at one predictors
multivariable regression looks at multiple predictors
what are the advantages of multivariable regression?
there are often multiple predictors/ explanatory factors of a given outcome
can adjust/ control for confounders
make prediction based on a combination of risk factors
what are the 9 Bradford Hill criteria
strength of association - the stronger an association, the more likely it is to be causal
consistency - association shown across different studies in different locations, populations, using different methods
specificity - specific exposure-outcome relationship e.g asbestos and asbestosis
temporality - exposure must proceed outcome
biological gradient - dose response i.e increase in exposure = increase in outcome
plausibility - biological mechanism that would explain outcome development
coherence - compatible with existing theories
experiment - outcome altered with experimentation e.g reversible
analogy - similar cause-effect relationships established
what is epistemology?
involves knowledge claims and what we can assert about the world and limits of what can be known
what are the two main epistemological positions?
positivism and interpretivism
what is positivism?
about EXPLANATION - need for statical generalisation
key philosophy which underpins research in the natural and physical sciences
based on concepts such as
- objectivity
- scientific method
- empiricism
what is interpretivism?
about EXPLORING AND UNDERSTANDING - need for depth and context
not a single philosophical approach, but linked to several
hermeneutics - interpreting unique human activity
phenomenology - how individuals experience the world
the assumption is that social reality can only be understood through social constructions such as language, consciousness and shared meanings and understandings
does not predefine variables, but explores human sense making in naturalistic settings
what is the difference between methodology and method
method is a specific technique for data collection
methodology is the study of methods and refers to the strategy or approach to research
compare the two methodologies
quantitative
- deductive (theory testing)
- large random samples
- results as numbers and statistics
- emphasis on following original research plan
qualitative
- inductive (develop theory)
- small purposeful samples
- results as words and concepts
- flexibility of approach
give formats of qualitative data
interview transcripts
photographs
blogs
social media
what is the purpose of qualitative analysis?
provided interpretation of seemingly inexplicable activities e.g drug taking
very good at getting contradiction around an issue
provide general statements about relationships among categories of data
give examples of analysis of qualitative data
narrative, IPA, grounded theory
characteristics of qualitative research
natural context
non-manipulative
subjectivity/ reflexivity
are systematic reviews/ meta analyses primary or secondary evidence?
secondary
what are scoping/ narrative reviews?
summarise available research on a given topic
question may be broader
do not necessarily follow such strict, standardised, transparent methodology
less rigorous, more subjective and more prone to selection bias
what is a systematic review?
a research article in which existing evidence on a topic is systematically identified, appraised and summarised according to predetermines criteria
synthesises the available evidence on a given topic to answer a research question
why are systematic reviews top of the hierarchy of research evidence?
transparent, systematic methods make the process replicable
bias is addressed by assessing each individual study for bias
provide reliable estimates of intervention effects
steps of a systematic review
specify research question
develop search strategies and inclusion/ exclusion criteria
identify relevant studies
assess quality and risk of bias
extract data from each study
pool and interpret results
answer research question
what is a meta-analysis?
‘the analysis of analyses’
statistical method for combining evidence from different sources
often used in systematic reviews
what type of graph shows a meta analysis?
forest plot
describe a forest plot
one row per study
the point estimate is shown as a square with size proportional to the size of the study
horizontal lines are confidence intervals
shows the measure of effect (for example, odds ratio)
solid vertical line indicates the lie of no effect (null). if an odds ratio has been used, the line of no effect is at the null value of 1
summary measure included - diamond shows the pooled estimate from the meta-analysis
what are fixed effect and random effects?
different approach to meta analysis
what is sensitivity analysis?
analysis to test the robustness of the findings of primary analysis - looks at the effect of assumptions or variations in approach