Exam 1: Hoff Flashcards
(89 cards)
1
Q
Modern Definition of EBM
A
USe of mathematical estimates of risk of benefit and harm, deried from high-quality research on population samples, to inform clinical decision-making in the diagnosis, investigation, or management of individual patients
2
Q
EBM emphasizes… (4 things)
A
Best Evidence
Clinical Expertise
Patient Values
Patient Circumstances
3
Q
Best Evidence
A
Valid and Clinically Relevant
4
Q
Clinical Expertise
A
Clinical skills/experience to understand patient’s situation, diagnosis, risks, and possible benefits from interventions
5
Q
Patient Values
A
Preferences, concerns and fears, expectations
6
Q
Patient Circumstances
A
Individual status and clinical setting
7
Q
What are the 5 basic steps in EBM?
A
- Ask answerable questions
- Search for best evidence
- Critical appraisal for validity and relevance
- Integrate evidence, clinical expetise, and patient values/preferences and apply
- Evaluate results
8
Q
Clinicians Incorporate Evidence by: Doing Mode (Define)
A
Carry out at least the first 4 steps (Ask AQ, Best Ev, Appraisal, Integrate/Apply, Evaluate)
9
Q
Using Mode (Define)
A
Restricts searches to sources have already been critically appraised (only step 1 and 2)
***Generally for uncommonly seen problems***
10
Q
Replicating Mode
A
Following the opinions of recognized experts regarding very rare entities
11
Q
What does PICO stand for?
A
Patient or Problem
Intervention
Comparison Intervention
Outcomes
12
Q
Background Questions
A
Seek General Knowledge (Students)
Two Components:
- Who, what, where, when, why, how?
- Plus the disorder itself
13
Q
Foregroud Questions (Majority)
A
***Clinicians***
Seek specific knowledge for patient management
-Comprise three or four elements (PICO)
14
Q
What source is the best when searching for evidence?
A
Electronic (online) sources
-More likely to be current, updated, and (now in many cases) evidence-based
15
Q
What constitutes best evidence?
A
Current, valid, and clinically relevant
16
Q
Hierarchy of study reliability:
A
Systematic reviews (or meta-analyses) of RCTs
Randomized, controlled trials (RCTs)
Prospective Studies
Retrospective Studies
Cross-sectional Surveys
Case Series
Case Reports
***SRPRCCC***
17
Q
What is the best site for systematic reviews?
A
The Cochrane Collaboration
18
Q
What type of study design is not randomized?
A
Case-control studies
19
Q
Which type of study will bring bias, as certain things are being looked for?
A
Cross-sectional surveys
20
Q
What is IMRAD and what is it used for?
A
Abstract
Introduction (why the research was done)
Methods (how the research was structured)
Results (what was found)
How Analyzed
Discussion (what the researchers think the results mean)
IMRAD is the standard structure of research papers
21
Q
What 3 preliminary questions should be asked?
A
- Why was the study in question done; what hypothesis was tested?
- What type of study was done? (e.g. primary or secondary)
- Was the study design appropriate to the broad topic of research addressed?
22
Q
What are examples of primary studies? Secondary studies?
A
Primary: Experiments, Observations, Clinical Trials (Therapies, Diagnostics, etc.), Surveys, Questionnaires
Secondary: Reviews (systematic or non-systematic), Economic analyses, Decision analyses
23
Q
Intervention (Therapy)
- Best Study Design
- Description
A
- RCT
- Subjects randomly allocated to treatment or control and outcomes assessed
24
Q
Harm/Risk/Etiology
- RCT
- Description
A
- RCT
- Similar to intervention questions, but RCTs to assess harmful outcomes are not ethical
25
**Harm/Risk/Etiology**
1. Cohort Study
2. Description
Outcome(s) compared for matched groups with and without exposure or risk factor (**prospective**)
26
**Harm/Risk/Etiology**
1. Case-Control Study
2. Description
Subjects with and without outcome of interest are compared for **previous** exposure (**retrospective**)
\*\*\*more **bias**, but occasionally done to look at harm or risk\*\*\*
27
**Diagnosis**
1. Best Study Design
2. Description
1. Diagnostic Validation Study
2. Independent, blinded comparison with reference ("**gold**") standard
28
**Prognosis/Prediction**
1. Best Study Design
2. Description
1. Cohort Study
2. Long-term followup of representative cohort to determine outcome
29
Randomized Control Trials
1. Concealed allocation
1. Randomization generally should be **_concealed_**
30
Randomized Controlled Trials
1. Single- vs. Double-Blind
Single (only subjects)
Double (subjects and investigators)
31
RCTs
1. Advantages
Allow rigorous evaluation of a **single variable**
Designed **prospectively**
Seek to confirm a **null hypothesis**
Allow for **meta-analysis**
Minimize **bias**
32
RCTs
1. Disadvantages
Expensive
Require long-term study (months to years)
Could introduce **hidden bias**
33
When is **randomization** _inappropriate_?
When the study involves **_prognosis_** of a disease
Investigating the **_validity_** of a diagnostic or screening test
Investigating quality of care issues when criteria for **success** are _not known yet_
34
Cohort Studies
1. Design
2. Prognosis Studies
3. Bias
1. **Observational - NO intervention**
- Two or more subject groups selected based on exposure or no exposure to a particular substance/organism to compare outcome
2. **_Prognosis Studies:_** special case - usually intended to determine **what happens** to people with a certain diagnosis or problem (e.g. heart failure) over time
- Require long-term follow up
3. Can be **substantially bias**
35
Case-Control Studies
1. Design
2. Bias
3. Concerned with
1. **Observational**, _not **randomized**_
- Patients with certain condition, "matched" with controls data usually obtained ***_retrospectively_*** for exposure to a disease-causing agent or circumstance
2. **More prone to _bias_** than cohort studies
3. Generally concerned with **harm** or **etiology**
36
Cross-Sectional Surveys
1. Design
2. Bias
3. Other uses
1. Representative sample interviewed, examined, or evaluated about a **specific question**
- Data are collected at a specific time but often include **retrospective information**
2. Potential for **significant _bias_**
3. Also used in questions of **Etiology**
37
Case Reports
1. Design
Simple report of a medical history
- Sometimes a **series** of patients' histories are reviewed and analyzed together
- Weak statistically but may give occasional insights into rare or unusual conditions
**Weakest:** report of a _single_ case
38
Systematic Reviews
1. Design
2. Key Features
1. Commonly employed in evaluating/comparing **RCTs**; can be used to compare others as well
- Conducted according to **strict methodology**
2. Key Features:
- Statement of objectives, materials, and methods
- Include **all** original reports available, **_globally_**, even if unpublished
- Individual reports each **critically evaluated**
- Conclusions based only on studies meeting **pre-set quality criteria**
39
Systematic Reviews
1. Advantages
- Large amounts of information assimilated quickly
- Explicitly **limit bias** when selecting studies for review
- Compare studies for **consistency** and **generalizability**
- Inconsistency **(heterogeneity)** is easily identified and new hypotheses can be formulated
- Conclusions are **more reliable**
- Quantitative reviews (**_Meta-Analyses_**) increase precission of overall results
40
Meta-Analysis
1. Design
A statistical synthesis of numerical results of **several studies** which all addressed the **_same question_**
- Incorporate the **advantages** of systematic reviews
- **Powerful _statistical analysis_**
41
Diagnosis Studies
1. Uses
2. Randomization
1. **Screening** a general or specific population, **Diagnosis** in an individual, Patient **follow-up** and management
2. **NOT** **randomized**
42
Three Preliminary Questions to ask for **all papers**:
1. _Why_ was the study done? (purpose or hypothesis? background?)
2. _Primary vs Secondary_
3. Appropriate _Design_ (e.g. RCT, Cohort, Case-Control, etc.)
\*\*\*Check the **abstract** first\*\*\*
43
What should be assessed first when reviewing the **results** of **Diagnosis Studies**?
Validity
44
Questions to ask when verifying **Validity**?
1. Research question clearly defined?
2. Presence or absence of disorder confirmed by **reference test** (aka the **Gold Standard**)
- How did investigators know the disorder was present?
- Was comparison indepndent and blinded?
3. Was test evaluated on an appropriate spectrum of patients?
4. Reference standard applied to all patients?
-
45
Diagnosis Studies: **Clinical Important**
1. Questions to Ask
2. Ideal Test
How accurate is the test/What is its **_sensitivity and specificity_**?
\*\*\*Ideal test will produce a **high proportion** of **true positives** and **true negatives**
46
What is the **usefulness** of any **Diagnostic Test** based on?
Its **accuracy in identifying the disorder**
47
Diagnostic Studies: **Likelihood Ratios**
Measure of **accuracy**
Predict likelihood of a certain result in a patient **with** the target disorder when compared to the likelihood of the same result in someone **without it**
48
Diagnostic Studies: Sensitivity vs Specificity
**Sensitivity:** proportion **_with**_ the disease testing _**positive_**
**Specificity:** proportion **_without**_ the disease testing _**negative_**
49
Diagnostic Studies: SpPin and SnNout
**SpPin:** if test with _high specificity_ is _positive_, rules the condition _IN_
**SnNout****:** if test has _high sensitivity_ and is _negative_, rules the condition _OUT_
50
Diagnostic Studies
1. Pre-Test Probability
2. Post-Test Probability
1. **Pre**: based on **estimates of prevalence** (how often a condition is found in a _specific population_)
2. **Post**: a good test **increases** post-test probability significantly
51
Diagnostic Studies
What does a **likelihood ratio (LR)** of 1 mean?
LR \> 1?
LR \< 1?
LR = 1 means that post-test and pre-test probability is the same
LR \> 1 **increases** the probability; says the test is **accurate and helpful**
LR \< 1 **decreases** the probability; says the test is not helpful
\*\*\*Represented with **Bayes' Theorem**\*\*\*
52
Diagnostic Studies
1. Will the results change your management?
**_If the LRs are close to 1_**, then post and pre test probabilities will be similar and therefore **_not very useful_**
53
Therapy Studies
1. Two key questions to ask
Do they have a **clear question**?
Is it **randomized**?
54
Therapy Studies
1. About
\*\*\*Most **_common_** kind of clinical papers, by far\*\*\*
- Ideally performed **_randomized_**
- Studies of **adverse reactions** to therapy (studies of **harm or risk**) are done **_differently_**
55
Preliminary Questions for **all** papers
Why was the study done?
Primary vs secondary?
Appropriate design?
56
Primary vs Secondary
Primary: trying to find **new info**
Secondary: **meta analyses** or **systematic reviews**
57
Therapy Studies
1. Validity
Clearly defined question?
Randomized?
**\*\*\*Blinded\*\*\***
***_Concealed Allocation_***
58
Therapy Studies: Validity
1. What is the most important thing to look for in RCTs?
**_Concealed Allocation_**
-Avoids selection bias if clinicians **_do not_** assign gropus
59
What is the **Rule of Thumb** for assessing the **Validity** of **Therapy Studies**?
If less than **80%** completed followup, likely **_invalid_**
60
Therapy Studies: Validity
1. Intention to Treat
Preserves randomization (minimizes bias)
-Means **all data** from subjects in the intervention group **is analyzed**
61
Therapy Studies: Validity
1. What happens if researchers fail completely analyze subjects **with their assigned group**?
Results can be skewed **toward the intervention**
62
Therapy Studies: Importance
1. Three questions to ask
How large is the effect of treatment? (**accuracy**)
How precise are the findings?
How useful is this in practice?
63
What is another term for **systematic error**?
**Bias**
64
What does a p-value of **\< 0.05** mean?
It means that there is less than a **1 in 20** probability that the result is due to **chance alone**
Smaller p-value = less likely due to chance
65
Therapy Studies: Importance
1. Confidence Intervals
Stated in Results section of a paper as a **range of values** around a demonstrated difference
When the CI crosses **1.0 ("Line of No Difference")** the result of interest is **_not statistically significant_**
66
Therapy Studies: Importance
1. Number Needed to Treat (NNT)
On average, if you do this intervention to this set of people who have this set of problems, if you treat x number, you will get a good result
\*\*\*Number of patients you would have to treat to achieve **one good result**\*\*\* (or avoid one negative)
\*\*\***Lower NNT = BETTER**\*\*\*
1 is the best, 100% will have an effect
67
Therapy Studies: Importance
1. What is the most important part of treatment?
It is important that treatments improve outcomes that are **important to _patients_**
**Better Approach:** Is there a compelling reason why results should **_not_** be applied to your practice?
68
Summary of Therapy Studies
1. Best Studies of Treatment
2. Determine Validity (5)
3. Determine Importance (3)
1. **RCTs**
2. Clearly defined research question
Randomization and blinding
Intention to treat
Equal treatments of groups
Comparable groups at the beginning
3. Accuracy
Precision (p values, \*\*\*CIs\*\*\* (Best))
Usefulness in **your** practice
69
Harm/Risk/Etiology Studies
1. Study Types
2. Definition
1. Cohort Studies (prospective) or Case-Control Studies (retrospective)
2. **Observational** and **Collecting Information**
\*\*\*These studies investigates ***_what happens (or happened) as a result of an exposure_***\*\*\*
70
Harm/Risk/Etiology: Importance
1. 3 Questions to Ask
How strong is the association between exposure and outcome?
How precise is the estimate of risk?
- If **cohort** study (Prospective), must use **RR (Relative Risk)** i.e. Hazard Ratio (HR)
- if **case-control** study (Retrospective), use **OR (Odds Ratio)**
Statistical Significance?
-Look for 95% CI around results - if RR or OR cross 1.0 then **no effect**
71
Prognosis Studies
1. Definitions
2. Prognostic Factors
1. **Possible** outcomes of a disease or condition and
- **Frequency** with which possible outcomes occur
Synonym is **_*\*\*\*natural history\*\*\**_** of an ailment
2. **Prognostic Factors**: characteristics or population that can be used to more accurately predict that patient or population's outcome
- Demographic (e.g. age, gender, etc.)
- Disease-specific (e.g. tumor stage)
- Comorbidity (e.g. diabetes in a cardiac patient)
72
Which type of study is best for studies of **prognosis**?
**Cohort Studies (prospective)**
Also can use Case-control studies
73
Prognosis Studies: Importance
1. Questions to ask
What is the risk of the outcome over time?
How preceise are the estimates of risk?
74
Prognosis Studies: Validity
1. 3 questions to ask
Was there a **representative and well-defined sample** of subjects? (recruied at common point in illness, inclusion/exclusion criteria, representative setting, similar to your own patients)
Was follow up sufficiently long and **complete**? (\>20% lost to followup, validity is suspect)
Were outcomes measured **"blind"**?
75
Prognosis Studies: Referral Bias
1. Definition
A _systematic error_ that can occur when patients are selected from **different care settings (e.g. tertiary care** compared to primary care centers**)** thus increasing the likelihood of **adverse or unfavorable outcomes**
76
Prognosis Studies
1. Were **objective** and **unbiased** criteria used?
- Outcome of interest is defined **_before_** the study starts
- Criteria for diagnosis of outcome of interest should be as **objective as possible**
- **_Blinding_** of individual(s) measuring outcome is best but not alwasy possible
77
Prognosis Studies
1. Reassuring a Concerned Patient/Relative
A valid, precise, and generalizable result of **good prognosis** is usually very helpful in reassuring
78
Systematic Review
1. Definition
Review of a clearly formulated question that uses systematic and explicit methods to identify, select, and critically appraise relevant research and to collect and analyze data from studies included
- Research synthesis
- Research integration
- Information synthesis
- Overview
79
Meta-Analysis
1. Definition
Statistical analysis of a collection of analysis results from individual studies for the purpose of integrating the findings
80
Systematic Reviews: Importance
Odds Ratios
Relative Risk
81
Systematic Reviews: Validity
1. **High quality studies** means...
2. How and what trials were found?
3. Are results consistent across studies?
1. RCTs
2. Should go **beyond** standard databases, Should include **all relevant studies (including worldwide and non-published)**
3. What criteria used to extract data (think **PICO**)
- Clinical **heterogeneity** makes studies **less useful**
82
Odds Ratio
1. Used
2. What does OR of 1 mean?
1. Used in **case-control studies** to copmare whether or not probability of a certain even tis the same in two groups
2. The event is equally likely in both groups
83
Relative Risk
1. Used
2. What does a RR of 1 mean?
1. Used in **RCT (randomized clinical trials****)**or in**cohort studies** to state risk of an event (or developing a disease) relative to exposure
2. Means that the treatment **did not do anything**, as it is equally probable in both groups
84
Systematic Reviews: Integrating the Findings
1. Single Studies
2. Studies can disagree with respect to:
1. _Single_ Study rarely provides definitive conclusions; reviewing a group of similar studies adds **power**
2. Can disagree with respect to:
- Clinical Significance: the **magnitude** of effect
- Statistical Significance: the likelihood that an observed effect occurred by **chance alone**
85
What does the **PICO** mnemonic help with?
Asking **Answerable Questions**
86
What does PICO stand for?
**P**atient or **p**opulation
**I**ntervention
**C**omparison intervention
**O**utcome of interest
87
What is the basis for evidence based medicine?
Asking **truly answerable questions**
88
What is the best design for a study of **Therapy**?
Double blind, randomized control trial
89
Studies on **prognosis** seek information on which of the following?
1. The _reason_ a particular condition has happened
2. The _results_ of a particular intervention
3. The _outcome_ of a test for a condition
4. The _possible outcomes_ of a disease or condition
5. Which patients _survive longest_
The possible outcomes of a disease or condition
