Lecture 1 Flashcards
Evidence-based medicine is
Use of mathematical estimates of risk of benefit & harm, derived from high-quality research on population samples, to inform clinical decision-making in the diagnosis, investigation, or management of individual patients
EBM practice emphasizes
best evidence, clinical expertise, patient values & circumstances
Basic steps
- Ask answerable questions
- Search for best evidence
- Critical appraisal for validity & relevance
- Integrate evidence, clinical expertise, & patient values/preferences & apply
- Evaluate results
Answerable questions consist of:
PICO Patient or problem Intervention Comparison intervention Outcomes
Best medicine should be
patient-centered
What is the best source for evidence
Electronic online sources
Current, recently updated, & evidence-based
Most reliable studies are
Systematic Reviews (or meta-analyses) of RCTs (followed by Randomized, controlled trials)
Least reliable studies are
Case reports
Hierarchy of study reliability:
- Systematic Reviews (or meta-analyses) of RCTs
- Randomized, controlled trials
- Prospective studies
- Retrospective studies
- Cross-sectional surveys
- Case series
- Case reports
Are prospective or retrospective studies more reliable?
Prospective studies
Systematic reviews present
evaluations of groups of studies relating to diagnosis & screening, therapy, prognosis, & Harm/risk/etiology
Descriptive statistics
Involves collecting data, presenting data, & characterizing data in order to describe data
Inferential statistics
Involves estimation & hypothesis testing in order to make decisions about population characteristics
Experimental units (elements)
Object upon which we collect data
Sometimes called units of analyses
Population
All items of interest
Statistic coming from whole population is called a parameter
Sample
Subset of units of a population
Statistics come from samples
Variable
Characteristic of an individual experimental unit
Symbol of event, act, trait, or attribute that can be measured & we assign some values
Categorical Variable
Some numeric or character codes that represent either the presence or absence of something that is of interest OR
the relative weight or rank of the thing that is of interest
Quantitative variable
Variable that holds the numerical result of some measurement usually taken using some standard unit
Process
Series of actions or operations that transforms inputs to outputs; produces output over time
Nominal Scale
Simplest level of measurement - categories without order
i.e. hair color
Ordinal Scale
Nominal variables with an inherent order among categories
i.e. disease stages (no set difference, equidistant between values)
Interval Scale
Measurable difference or interval or distance between observations
I.e. height, age
Ratio
Same as interval but with an absolute reference point (such as 0)
Qualitative Data is described using a
Bar Graph or Pie Chart (derived from a summary table or frequency table)
Quantitative Data is described using
Dot Plot, Stem & Leaf display, Frequency distribution (histogram)
Bar Graph
Has vertical bars for qualitative variables
Equal bar widths, height shows frequency
Class
One of the categories into which qualitative data can be classified
Class relative frequency
Class frequency (number of observations) divided by total numbers of observations in data set
Histogram
Used for quantitative variables (numbers not categories) arranged into intervals (same width)
Bars touch
Central tendency
tendency of data to cluster, or center about certain numerical values - location
Measured by mean, median, & mode
Variability
Dispersion, spread of data
measured by range
Variance & Standard deviation
Measure dispersion & show variation about mean
Consider how data are distributed
Standard deviation
s
Dispersion about sample or population mean
Variance
s^2
squared dispersion about sample mean
Left-Skewed
Means that the majority of the data points are more than the mean
(mean is skewed to left; median is greater than mean)
Right-Skewed
Majority of data points are less than the mean
mean is skewed to the right; median is less than mean
Symmetric
Mean = median
Normal distribution
68% of measurements lie in interval
sample mean - s to sample mean + s
One standard deviation
z-score -1 to 1
95% of measurements lie in interval
sample mean - 2s to sample mean + 2s
2 standard deviations
z-score -2 to 2
99.7% of measurements lie in interval
sample mean - 3s to sample mean + 3s
3 standard deviations
z-score -3 to 3
Calculate probability
How often an outcome of interest occurs/ sample space (number of total possible outcomes)
Event
Occurrence: event that 30 year old lives to be 70
OR
Collection of one or more outcomes of an experiments
Intersection
Both event A & event B occur
Union
Either event A or event B or both event A & B occur
Complement
Everything except event A
Special rule of addition
Probability that 2 mutually exclusive events will occur is equal to the sum of the probabilities of individual events
P(A or B) = P(A) + P(B)
2 events that can not occur simultaneously are
Mutually exclusive or disjoint
General rule of addition
If 2 events can occur at same time, probability of both occurring is equal to the sum of the probabilities of the individual events minus the probability of the intersection (of the 2 events)
P(A or B) = P(A) + P(B) - P(A and B)
Probability of event B occurring given that event A has already occurred (B given A) if events are not independent is
P(A given B) = P(A and B)/P(B)
OR
P(A and B) = P(A given B) x P (B)
Special rule of multiplication
If 2 events are unrelated, or independent
P(A given B) = P(A) and P(B given A) = P(B), then
P (A and B) = P(A) x P(B)
Unbiased estimate of sampling distribution
If sample statistic has a mean equal to the population parameter
Biased estimate of sampling distribution
If mean of sample statistic is not equal to the population parameter
