EBP Quiz Flashcards
Evidence based practice
Integration of best available, evidence, expertise, and patient values when evaluating in treating
History of manual muscle tests
In the 1930s, from the polio epidemic
History of physical therapy
World War I, Mary McMillan. A PTA started in 1921.
Archie Cochrane
Created grouping of articles and summarizing them. One of the founding fathers of evidence base practice.
David sackett
One of the founding fathers of evidence based practice. Developed levels of evidence and the pyramids of articles.
Process of evidence based practice
Patient dilemma, asking questions, acquiring evidence, assessing evidence, and apply and reassess
PICO
Patient/problem, intervention, comparison intervention, outcome
FINER acronym
Feasible, interesting, novel, ethical, relevant
Database searching at TWU
Go to library, click on databases, start with PubMed, and CINAHL
Medline database
National library of medicine
Pubmed database
Provides free access to Medline. Can search PICO literature
Primary sources of evidence
Original research
Secondary sources of evidence
Predigested research
Randomly controlled trial
Highest quality. Subjects are randomized and typically blinded
Cohort study
Predetermined groups with specific disease
Case-control studies
Groups have disease, look back in time to find possible causes
Systematic review versus meta-analysis
Systematic review: asks specific clinical questions and performs a comprehensive search that illuminates the poorly done studies
Meta-analysis: systematic review that includes data
CEBM information resources pyramid levels
1a, 1b, 1c. 2a, 2b, 2c. 3a, 3b. 4, 5.
CEBM information resources pyramid grades
A-F
1A
Systematic reviews/meta-analysis.
1B
Randomized controlled trials with a narrow confidence interval
1C
All or none. All patients died before research was available, now some survive. Some patients died before research was available, now, none die.
Confidence interval
Shows the direction research favors and precision of estimated value. Do not want to cross zero. less than zero or greater than zero is preferred.
Narrow confidence, interval: no overlap in the confidence intervals. Wide confidence intervals: lots of overlap in the confidence intervals.
Standardized mean difference
Measure of effect size. The main difference divided by the mean standard deviation of the difference.
2A
Systematic review of cohort studies. Identification of two groups, one that receive exposure, and one that did not.
2B
Individual cohort study. Or low quality randomized controlled trial.
2C
Outcomes research. Focused on the value for the patient (effectiveness). Does the intervention make a difference in the eyes of the patient?
3A
Systematic review of case-control studies. Identifying patients who have the outcome of interest and then looking backwards in time to find the causes.
3B
Individual case control study. I didn’t find patients who have the outcome of interest, and then looking backwards to find the cause.
4
Case study or series. Detailed analysis of a well defined case.
5
Expert opinion or bench study (research done, not on humans)
Grade A evidence
Strong evidence with at least one level one study
Grade B evidence
Moderate evidence with a single high-quality randomized controlled trial
Grade c evidence
Weak evidence, single level 2 study
Grade D evidence
Conflicting evidence, higher quality studies on the topic disagree
Grade E evidence
Theoretical/foundational evidence, evidence from animal or cadaver studies, or bench search
Grade F evidence
Expert opinion, based on clinical evidence alone
PEDro
Tool to measure quality of randomized controlled trials, ranging from 0 to 10 with greater than six being considered high quality
Article components
Title, abstract, key words, introduction, Methods, results, discussion, conclusion
Authorship order
First author: designs and conceives to study and writes the first draft
Last author/senior author: usually experienced investigator, who partners with the first
Second author: major contributor
Middle authors: everyone else who qualifies for authorship
Beall’s list
Oh, wait to search for predatory journals, 15,000 predatory journals identified
Hypothesis
Null or alternative. Null not used as often.
Methods section
4 subsections: participants, instrumentation, procedures, data analysis
Methods participants
Clear, inclusion and exclusion criteria, should not be redundant. Should have a reasonable amount of participants, as well as a consort flow diagram to show the allocation of participants.
Descriptive statistics: standard deviation
Square root of sum of squares/number of participants -1
Descriptive statistics: variance
Standard deviation squared
T-test
Differences between two means. Can be matched or non-matched.
ANOVA
Differences between multiple means. Compares affect of intervention, effect of time on intervention, and interactions between time and affect in the groups
R-value
Measures the degree and direction of a relationship between two variables. Do you want it to be closer to 1. Negative is a negative relationship, and positive indicates a positive relationship.
0.05
Alpha level
Set an advanced as a goal for the significance level typically at 0.05, or a 5% risk of committing a type one error
P value
Actual probability that the results occurred are just an error, do not want it to be greater than 0.05.
Reliability
Consistency of the results. you want a scale to be reliable
Validity
How accurate are the results? you want a thermostat to be valid.
Types of reliability
Intertester: can different subjects, get the same score
Intra-tester: can the same subject get the same score
Test retest: can multiple measurements by an instrument be consistent and canned multiple measurements by a patient be consistent
Agreement
With two or more ratings do the scores remain reliable
Types of validity
Face: does an instrument measure what it is supposed to
Contant: does the instrument determine the entire domain it is to be measured (does the IQ questionnaire cover all areas of intelligence)
Criterion based: the degree of which the outcomes of one test correlate with the gold standard
Concurrent: degree of which the outcome of one test correlates with the outcome of another test
Predictive: can an instrument be used to predict future outcomes
Construct: degree of which theoretical construct is measured