Week 5 a Flashcards
What are the 3 types of literature reviews?
- Narrative Review
- Systematic Review
- Meta-analysis
What is a Narrative Review?
- Selective review of the literature that broadly covers a specific topic
- Does not follow strict systematic methods to locate and synthesize articles
- a good source for background information
- Prone to bias and lower on the hierarchy evidence pyramid (ex. authors may be selective as to which articles are included and they will include articles that support their hypothesis and exclude those that don’t)
What is Systematic Review?
- Utilizes exacting search strategies to make certain that the maximum extend of relevant research has been considered
- Original articles are methodologically appraised and synthesized
- Searching, Appraising, Summarizing the existing info on a selected topic
- Most commonly address questions of effectiveness
- The entire process of collecting, reviewing, and presenting all available evidence
What is Meta-analysis?
- Quantitatively combines the results of a systematic literature review
- Capable of performing a statistical analysis of the pooled results of relevant studies
What is the goal of a Systematic Review?
Aim to find all studies addressing the reviews question using an object and transparent process
What is the Cochrane Collection?
Gold Standard for Systematic Reviews
What is the process of a systematic review?
- State the study objective
- Develop the protocol
- Develop a research strategy
- Conduct the research
- Retrieve Relevant Papers
- Screen and select papers that meet established criteria
- Evaluate methodological Quality of Selected Studies
- Analyze and Synthesize findings
- Determine if Statistical data are sufficient for further analysis
No = Report results of systematic Reveiw
Yes: Analyze Effect Size Estimates then report results of meta analysis
What is the PEDro Scale?
Asses the quality for interventional studies
What is Meta Analysis?
The statistical technique involved in extracting and combining data to produce a summary result (optional part of systematic review)
When should you use Meta Analysis?
- When more than one study has estimated an effect
- When there are no differences in the study characteristics that are likely to substantially affect outcome
- When the outcome has been measured in similar ways
What is a Forest Plot?
- A type of graph often used in meta analysis to illustrate the treatment effect sizes of the studies
- Each study is represented by a black square and that is an estimate of their effect size
- If the line includes the line of no effect the line is not statistically significant
When is Cohen’s d appropriate?
Continuous data
When is an Odds Ratio (OR) appropriate?
When the study’s outcome measure is dichotomous
ex) pain vs no pain
What is Odds Ratio?
A comparison of the odds of the outcome being present in the treatment group against the control group
What is Confidence Intervals (CI)?
- A range of values that we are confident contains the population value
What is Point Estimate?
- A single value that represents the best estimate of the population value
What are small samples more susceptible to?
Chance variation, thus they are given less weight than larger studies so they will have less influence on the final estimate
What is Weighting based on?
Quality
What happens when data from individual studies combine?
increase in sample size
What are the two types of diagnosis?
- Medical Diagnosis (Herniated disc L4-5 and Polymyosits)
2. Physical Therapy Diagnosis (Right sided radiculopathy centralizing w/ repeated extension)
When you see pattern recognition this is caused from?
- History Exam
2. Physical Exam
What is pretesting probability?
For any given patient there is a baseline probability of a certain condition pretesting
What is Post test probability?
Application of a clinical diagnostic test alters the baseline probability
What are the key thresholds for diagnostic test?
- Test threshold: the probability below which a diagnostic test will not be ordered or preformed because the possibility of the diagnosis is so remote
- Treatment Threshold = The probability above which is a diagnostic test will not be ordered or performed because the possibility of the diagnosis is so great that immediate treatment is indicated
What is PEcO?
- P = The patient or problem
- E = Exposure to the diagnostic test under consideration
- c = rarely a comparison of multiple diagnostic tests (can be w/ study of clinical predictions)
- o = Outcomes
What is the basic structure of Diagnostic Study?
- Series of patients (78 patients w/ shoulder pain)
- Index test (Neer’s test)
- Reference (gold) Standard ( Rotator cuff tendonitis via MRI)
- Compare the results of the index test w/ the reference standard (2x2 Table)
What are the key methodological Aspects: Gold Standard?
- verifies the test (measure) does what its supposed to do
What are some commonly used gold standards?
- Radiographs
- Surgical exploration
- A test/measure with previous demonstrated consistency and usefulness
Gold Standard =’s?
Reference Standard
What are the key methodological aspects of blinding?
- were the individuals interpreting each tests results unaware of the other tests (measures) results
- Address the issue of measurement bias
What are the key methodological Aspects of spectrum of Patients?
- Were the subjects with all levels/stages of the disorder included in the study
- Clinicians must face diagnostic uncertainty
What is Spectrum Bias?
Lack of sufficient heterogeneity of subjects (All patients are at an extreme of a spectrum)
What are the key aspects of both tests?
- Did all subjects undergo the test (measure) of interest and the gold standard test (measure)
- Otherwise, study will distort properties of the proposed diagnostic test (Work up Bias or Verification Bias)
What are the 3 steps for appraising diagnostic tests?
- Are the results Valid?
- What are the results?
- Will the results help care for my patient
What is QUADAS Tool?
14 item survey used to assess quality of diagnostic test studies
What is a Cross Sectional Study?
Comparing a index test result and gold standard (dictomus)
What is Sensitivity?
- Portion of people with the disease who have a positive result
- left side of the table
What is Specificity?
- Proportion of people without the disease who have a negative test result
What is helpful for ruling out the condition
100% Sensitive
What helps for ruling in the condition
100% Specificity
What is SpPins?
- A test with high specificity
- This is positive
- Helps rule in condition
What is SnNout?
- A test with high sensitivity
- This test is negative
- Helps rule a condition out
What is Likehood Ratio (LR)?
- sensitivity info combined with specificity info
- +LR = Sensitivity/ (1-specificity)
- -LR (1-sensitivity/specificity
If a diagnostic test is positive use….
+LR
If a diagnostic test is negative use….
-LR
What yields a post test probability?
Pre test probability of having the condition combined with LR
In Clinical Prediction Rules…
its not really accurate to use LRs in series due to shared variance
What are synonyms for Clinical Prediction Rules?
- Clinical Prediction Guides
- Clinical Decision rules
- Test item cluster
What can CPR be used for?
- Diagnosis (Specificity, +LR)
- Screening Sensitivity, -LR)
- Factors that predict response to treatment
What is Canadian C-Spine Rule?
- Biased to maximize sensitivity adn -LR
What is Continuous Diagnostic Test Variables?
- Receiver operator characteristic (ROC) curves
- Plots probability of true positives (sensitivity) against probability of false positives (1-specificity) for all possible cut scores
What is Area under curve (AUC)?
overall measure of sensitivity and specificity
What is Dichotomizing Continuous Variables?
- To end up with a dichotomous CPR (yes or no) and calculate sensitivity, specificity, LRs, all variables need to be positive/negative
What are some issues with CPRs/TIC?
- Sample size
- Predictive of response to tretment or just good prognostic indicator
What is the development of a clinical Decision Rule?
- Derivation
- Validation
- Impact Analysis
What is Derivation?
- Someone comes up with an idea
- Is a Clinical Decision instrument possible
- Come up with a list of candidate variables
- Statistical Analysis
What is Validation?
- An unvalidated CDI should not be used
- confirm if the CDI actually works
Why do validation of CDI?
- The results could of happened by chance and we need to confirm the results
How do you validate a CCDI?
- Use the CDI and apply it to a population
- Internal (same)/External (different population)
- Narrow/Diverse Studies
What is Impact Analysis?
- make sure the CDI that was derived, an dvalidated actually works in the real world
- Make sure the CDI does what its suppose too
- Make sure its useable
How do we test the Impac Analysis?
- Take clinicians and randomize them (usual care and CDI) then look at benefit harm cost metric
- Pre/Post studies: take an institution and analyze how well they were preformed before/after CDI (Not randomized) then benefit harm cost