5. intro to Evidence Based Medicine Flashcards
what is Evidence-Based Medicine
the integration of BEST RESEARCH EVIDENCE with CLINICAL EXPERTISE and PATIENT VALUES
(dave sackett)
how can EBM help you
- make better clinical decisions
- share decision making with patients
- provide better diagnostic reasoning
- understand benefits vs harms
- practice more safely
4 STEPS of EBM
- Formulate an Answerable QUESTION
- Track down the BEST EVIDENCE
- critically APPRAISE the EVIDENCE (find out how good it is)
- APPLY the EVIDENCE (integrate results with clinical expertise and patient values)
what PRINCIPLE do we use when FORMULATING AN ANSWERABLE QUESTION and TRACKING DOWN EVIDENCE
PICO
- POPULATION/PATIENTS
- INTERVENTION (exposure, test)
- COMPARISON/CONTROL
- OUTCOME
what are the most common type of CLINICAL QUESTION about
how to TREAT a disease
treatments and therapies are called INTERVENTIONS
(but not all questions are about interventions)
DIFFERENT TYPES of Clinical QUESTIONS require..
DIFFERENT TYPES of RESEARCH EVIDENCE
a CROSS-SECTIONAL / PREVALENCE study is a study in which information is collected in a planned way in a defined population..
AT ONE POINT IN TIME
what is a RANDOMISED CONTROLLED TRIAL (RCT)
an EXPERIMENTAL STUDY where participants are randomised either to receive the NEW INTERVENTION being tested OR receive a CONTROL TREATMENT (standard treatment or PLACEBO)
APPRAISAL of any study must consider:
- VALIDITY
can the results be trusted? - RESULTS
what are the results?
how are they (or can they be) expressed?
what do they mean? - RELEVANCE (step 4 of EBM)
do these results APPLY to the local context
2 TYPES of VALIDITY
INTERNAL VALIDITY
- are the results of the study CORRECT for the people in the SAMPLE
EXTERNAL VALIDITY
- if so, does the sample REPRESENT FAIRLY in the kind of PATIENTS in MY PRACTICE
INTERNAL VALIDITY is determined by..
how well the DESIGN, DATA COLLECTION and ANALYSES are carried out
THREATENED by BIASES
EXTERNAL VALIDITY assumes…
and is a matter of..
ASSUMES INTERNAL VALIDITY
is a matter of PERSONAL JUDGMENT
a SAMPLE is used to..
MAKE INFERENCE about the POPULATION
What is a RANDOM (sampling) ERROR
the answer may differ from the truth when taking several different samples from the population
by CHANCE the people used are different from average so result will be different
the best way to MINIMISE RANDOM SAMPLING ERROR is to
INCREASE SAMPLE SIZE
what are 2 METHODS of ASSESSING the ROLE of CHANCE
P-VALUES (HYPOTHESIS TESTING)
CONFIDENCE INTERVALS (ESTIMATION)
(more informative)
what is a HYPOTHESIS
a statement about a POPULATION PARAMETER (such as a population proportion or population mean) or some other characteristic of a population
HYPOTHESIS
is it EASIER to PROVE or DISPROVE a statement
DISPROVE is easier
2 competing Hypotheses
what is a NULL HYPOTHESIS (H0) and ALTERNATIVE HYPOTHESIS (H1)
NULL
- there is NO ASSOCIATION (/no diffference) between X and Y
PRESUMED to be TRUE unless the data gives strong EVIDENCE AGAINST it
ALTERNATIVE
- there IS an ASSOCIATION/difference between X and Y
STEPS in HYPOTHESIS TESTING
- state the NULL & ALTERNATIVE HYPOTHESES
- Decide what STATISTICAL TEST is appropriate
- Use the test to CALCULATE the P-VALUE
- WEIGH the EVIDENCE AGAINST the NULL
how to address if a result is a CHANCE finding (by chance)
calculate a TEST STATISTIC and its corresponding P-VALUE
what is a P-VALUE
a PROBABILITY (lies from 0 to 1)
- comes from a STATISTICAL TEST testing the null hypothesis
(start by assuming the null is TRUE)
measures the strength of the evidence AGAINST the null
what is used as the CUT OFF for a P-VALUE to REJECT the NULL hypothesis
0.05 / 5%
if P LESS than or EQUAL TO 0.05 then REJECT NULL HYPOTHESIS
if P GREATER than 0.05, DO NOT REJECT
the SMALLER the P-VALUE..
the STRONGER the EVIDENCE AGAINST the NULL HYPOTHESIS
2 possible of a Hypothesis Test
- REJECT NULL HYPOTHESIS
- DO NOT REJECT NULL HYPOTHESIS (can not say ‘accept’)
what is a TYPE I ERROR in a Hypothesis Test result
FALSE NEGATIVE
INCORRECTLY REJECTING the NULL
H0 is actually not rejected
(innocent)
what is a TYPE II ERROR in a Hypothesis Test result
FALSE POSTIVE
INCORRECTING FAILING TO REJECT the NULL
H0 should
what is a STATISTICAL POWER and how is it calculated
CORRECTLY REJECTING NULL (H0)
the PROBABILITY that a test will find a statistically signicant difference when a difference really exists
(probability of correctly finding that H0 is FALSE and a difference does actually exist)
POWER = 1 - TYPE 2 ERROR
if a statistical test result is ‘Not statistically significant’ what does it mean
there is INSUFFICIENT EVIDENCE for a DIFFERENCE or effect
(does NOT mean ‘no difference/effect’)
what happens to P-VALUE as the SAMPLE SIZE INCREASES
DECREASES
what is a better way of expressing UNCERTAINTY
CONFIDENCE INTERVALS
what are CONFIDENCE INTERVALS (CI)
a way of showing the UNCERTAINTY SURROUNDING a POINT ESTIMATE (sample statistic = point estimate of a parameter) by giving an INTERVAL (range) of LIKELY VALUES where the TRUE POPULATION VALUE IS LIKELY TO LIE
WIDER CI = MORE UNCERTAINTY
indicates the estimate is imprecise
by convention we usually quote …% CONFIDENCE INTERVALS
95%
What does a 95% CONFIDENCE INTERVAL mean
if we were to repeat the study many timeswith different samples, 95% of the 95% CIs would contain the true value
/
we are 95% CONFIDENT that the true value in the population lies within this range
what CLINICAL IMPLICATIONS can be derived from the CI
the UPPER and LOWER LIMITS provide a way of assessing whether the results are CLINICALLY IMPORTANT
what can you also check within your CI
if a HYPOTHESISED VALUE (ie Null value) for the population parameter falls within the confidence interval.
if so, our results are consistent with the hypothesised value
if not, it is unlikely that the parameter has this value.
