Exam 1 Flashcards
Q3: Are studies that would answer my clinical question likely to be included?
Review the inclusion and exclusion criteria to determine
Are the types of interventions investigated relevant to my clinical question and clinical practice?
Assess if the included interventions are
- Relevant to the clinical question
- Relevant to the clinical practice
Are the outcome measures relevant to my clinical question
- Methods section: outcome measures eligible to be included
- Outcome measures assessed
Is the study population (sample population) sufficiently similar to my patient to justify the expectation that my patient would respond similarly to the population?
- Summary of the demographics of the participants
- Close enough is fine
A systematic review should provide
a detailed list of the terms and search strategies
Was a comprehensive number of appropriate databases used?
- Attempt to find all of the available literature on a topic
- usually 3-7
Language bias results when
Important study results are excluded
Unpublished data can be obtained by?
Personally contacting the researcher who conducted the study
How recent was the Search
Time passed between the search window and the time that you are reviewing it
What search are likely to have a reduced risk for bias?
RCTs
Broader searches can provide?
a more expansive perspective on current evidence for that topic
Quality of the studies in the systematic review
Risk of Bias tool
- To Evaluate & judge
2 or more independent reviewers used a standardized study appraisal tool, if needed, a third review is consulted to resolve the discrepancy
Meta-analysis is useful because
it allows data from several studies to be combined to give a more precise estimate of the difference between two interventions
To produce valid results, combined studies must be similar in 3 aspects
- Study population (Patient characteristics)
- Intervention
- Outcome measures
Assessment of clinical heterogeneity involves the authors to conduct an analysis of which variables
- Study Population
- Intervention
- Outcome measures
When a meta-analysis is conducted
should be paired with a test of statistical heterogeneity
Name of statistical heterogeneity
Between-study variance
Cochran’s Q
Index of variability
If a meta-analysis was not conducted
should provide a narrative synthesis
If the outcome of interest is dichotomous
relative risk and odds ratio can be used
Relative Risk
- Greater than 1: increased risk
- Less than 1: decreased risk
- 1: identical risks
Odds Ratio
- 1: odds of exposure are similar
- Greater than 1: Increased likehood of exposure (positive association)
- Less than 1: Decreased likehood of exposure (negative association)
Outcome of interest is continuous (range of motion, strength)
mean difference or effect size between treatment and control groups is most commonly reported
If a meta-analysis is included
forest plot
Cohort studies
- a group of participants
- likely to develop
- followed into the future (prospectively)
Retrospective for Cohort studies
Previous history/ record
Case control studies
- Conducted after an outcome of interest has occurred
- Case group vs. controlled group
A sample that sufficiently represents the study event must
be identified at a common point in the progression of the patients’ condition
If the number of factors increase
The sample size of a study must also increase
In a longitudinal cohort study
the sample participants must not already have the study outcome
Time points should have relevance to?
The conditions and course of care
ICC
continuous
Spearman’s rho
ordinal
Kappa
Nominal
The objectivity is best obtained if
the measurers do not know the study purpose or the group status of the particpants
The greater the percentage of participants who have completed the study
the better
Linear correlation
- Direction of trend
- Quality of straight line fit
Regression
Prediction
The close the value to 1, the more robust the prediction
Linear Regression
- One variable is used to predict the level of another variable
- Assuming that the two variables have a linear relationship
- Outcome of interest must be continuous
Multiple regression
Predictions with multiple contributing variables
Outcome of interest is continuous
Logistic regression
- outcome of interest is assessed with a categorical variable
- one vs. another
Risk ratio typically used with
cohort design study
Odds ratio typically used with
case control studies
What are the two most common study appraisal tools in PT-related systematic reviews?
- Physiotherapy Evidence Database (PEDro) Scale
- Cochrane Collaboration Risk of Bias Tool for Randomized Controlled Trials
What is the most challenging elements of conducting a systematic review?
Determining if the data from multiple studies can be combined in a meta-analysis
What is “vote-counting”
The authors count the number of studies favoring one intervention compared to the number of studies favoring an alternative intervention
Forest plot
used to illustrate individual studies and pooled results from the meta-analysis
