Stats - Evidence based medicine Flashcards
Parametric (normally distributed) tests
T Test - paired or unpaired data
Pearson’s coefficient - correlation data
Data types
Nominal
Ordinal
Discrete
Continuous
Binomial
Interval
Nominal - Observed values into set categories which have no particular order or hierarchy. You can not order or measure nominal data (for example birthplace, hair colour)
Ordinal - Observed values can be put into set categories which themselves can be ordered (for example NYHA classification of heart failure symptoms)
Discrete - Observed values are confined to a certain values, usually a finite number of whole numbers (for example the number of asthma exacerbations in a year, whole number classification)
Continuous - Data can take any value with certain range (for example weight)
Binomial - Data may take one of two values (for example gender)
Interval - A measurement where the difference between two values is meaningful, such that equal differences between values correspond to real differences between the quantities that the scale measures (for example temperature)
Non parametric (skewed) tests
Mann-Whitney U - compares ordinal, interval, or ratio scales of unpaired data
eg 5 point scale, NYHA breathlessness scores
Wilcoxon signed rank - compares two sets of observations on a single sample, e.g. a ‘before’ and ‘after’ test on the same population following an intervention
Chi squared - used to compare proportions or percentages e.g. compares the percentage of patients who improved following two different interventions
Spearman, Kendall rank - correlation
Economic evaluation
cost-effectiveness analysis (CEA)
cost-benefit analysis (CBA)
cost-utility analysis (CUA)
cost-minimisation analysis (CMA)
CEA - compares a number of interventions by relating costs to a single clinical measure of effectiveness (e.g. symptom reduction, improvement in activities of daily living).
- eg how much £’s spent per lives saved, how much £’s spent per each depression free day
CBA - Monetary value of the benefits of intervention
- eg. life years saved, side effects, symptom relief are all given a monetary value
CUA - form of CEA, broader, able to compare between different conditions more easily
- eg. Quality-Adjusted-Life-Years (QALYs)
CMA - Compare two interventions. The aim is to decide the least costly way of achieving the same outcome.
Variables
Independent
Dependent
Controlled
Independent - the variable that the experimenter purposely changes over the course of the investigation.
- eg amount of alcohol drunk
Dependant - the variable that is observed and changes in response to the independent variable.
- eg counts of liver cirrhosis
Controlled - variables that are not changed
Graphical representations of statistical data
Box and whisker
Funnel
Histogram
Forest
Scatter
Kaplan-Meier
Box-and-whisker - Graphical representation of the sample minimum, lower quartile, median, upper quartile and sample maximum
Funnel plot - Used to demonstrate the existence of publication bias in meta-analyses
Histogram - A graphical display of continuous data where the values have been categorised into a number of categories
Forest plot - Forest plots are usually found in meta-analyses and provide a graphical representation of the strength of evidence of the constituent trials
Scatter plot - Graphical representation using Cartesian coordinates to display values for two variables for a set of data
Kaplan-Meier survival plot - A plot of the Kaplan-Meier estimate of the survival function showing decreasing survival with time
Significance tests
Null hypothesis
Alternative hypothesis
p value
Type 1 error
Type 2 error
Null hypothesis - two treatments are effectively equal
Alternative hypothesis - two treatments are effectively different
P value - probability of obtaining a results by chance
- the lower the p value, the lower the probability of obtaining the result by chance
- P <0.05 = statistically significant = good
Type 1 error - null hypothesis is rejected when it is true (false positive)
Type 2 error - null hypothesis is accepted when it is false (false negative)
Power definition
Power - probability of (correctly)rejecting the null hypothesis when it is false
Power = 1 - probability of type 2 error
- power can be increased by increasing sample size
Correlation
Correlation is used to test the association between two variables
eg whether salary and IQ are related
In parametric (normally distributed) data = use Pearson’s coefficient (r)
r = 1 = positive correlation
r = 0 = no correlation
r = -1 = negative correlation
In non-parametric (skewed) data = use Spearman’s coefficient (p or rs)
Linear regression may be used to predict how much one variable changes when a second variable is changed.
Levels of evidence system
1 Meta analysis/ Systematic review
2 Randomised controlled trial
3 Controlled trial without randomisation
4 Case control/ Cohort study
5 Expert opinion
Confounding
Confounding refers to an un-noticed variable that effects the variables in question leading to inaccurate results
eg ice cream leads to skin ca - this is incorrect. the confounding variable would be the sun.
Confounding is controlled during Design phase through ranDomisation
Confounding is controlled during analySiS phase through Stratification
Clinical Trial Phases
Phase 0 - Exploratory studies
- Very small number of participants and aim to assess how a drug behaves in the human body.
Phase I - Safety assessment
- Determines side-effects. Conducted on healthy volunteers
Phase II - Assess efficacy
- Involves small number of patients affected by particular disease
- IIa - assesses optimal dosing
- IIb - assesses efficacy
Phase III - Assess effectiveness
- Involve larger group, randomised controlled trial, comparing new treatment with established treatments
Phase IV - Postmarketing surveillance
- Monitors for long-term effectiveness and side-effects
Hawthorne Effect
describes a group changing it’s behaviour due to the knowledge that it is being studied
Meta-analysis of RCT + Odds ratio = what test?
Forrest plot - to show publication bias
?tip of pattern recognition rather than understanding the facts
Lead time bias is…
occurs when two tests for a disease are compared, the new test diagnoses the disease earlier, but there is no effect on the outcome of the disease
Late look bias is…
is a type of selection bias where gathering of information occurs at an inappropriate time. For example, studying a fatal disease many years after patients suffering from it have died.
