Clinical Investigation Final

0.0(0)
studied byStudied by 0 people
GameKnowt Play
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
Card Sorting

1/152

encourage image

There's no tags or description

Looks like no tags are added yet.

Study Analytics
Name
Mastery
Learn
Test
Matching
Spaced

No study sessions yet.

153 Terms

1
New cards

DEVELOPING GUIDELINES

......

2
New cards

What is a CPG?

statements informed by a systematic review of evidence and an assessment of benefit and harms

3
New cards

What are the steps used to develop evidence based CPGs (6)

1. Identify guideline objectives and scope (target population and questions)

2. convene an expert panel including clinicians, researchers, stakeholders

3. Assess the evidence (systematc reviews, benefit vs harm analysis)

4. Develop recommendations (strength and level of evidence)

5. Review (internal and external peer review)

6. Disseminate to clinicials, organizations, and stakeholders

4
New cards

Step 1

1. Identify target population and the specific outcomes to be addressed

- condition should be common

- have high morbidity and mortality

- poor patient outcomes

2. create specific questions to be answered

5
New cards

Step 2

convene subject matter experts and stakeholders

- can include clinicians, researchers, patients, caregivers

- have minimal conflicts of interest

6
New cards

Step 3

assess evidence

1. use hierarchy of evidence - meta analysis, systematic reviews, RCTs are best

- rating systems are used to assign:

1. level of confidence

2. strength of recommendation

7
New cards

Step 4

translate evidence

- provide rubric for grading level of evidence, strength of rec., and rationale

- Prepare evidence table that includes:

benefits, harms, burden of therapy

8
New cards

Step 5

review of guideline by internal and external processes

- content experts

- professional associations

- process experts

9
New cards

Step 6

Disseminate guideline to

- national orgs

-government

- providers

-payers

- patient advocates

10
New cards

What is a clinical practice guideline

systematically developed evidence based statements to guide healthcare decisions

they help translate research into real world effectiveness

11
New cards

What are the limitations of CGPs (8)

1. complex and length

2. lack of high quality RCTs for some topics

3. Timing - difficult to evolve rapidly changing evidence

4. limited applicability to underrepresented groups

5. lack of individualization to patient comorbidities

6. no independent body to oversee CPG development

7. many available guidelines on one topic

8. conflicting recommendations between guidelines from different organization

12
New cards

What is the purpose/rationale of CPGs (5)

1. reduce practice variability

2. increase percentages of patients with their condition Controlled

3. improve quality of care

4. improve patient outcomes

5. serve as educational tools

13
New cards

what is a key point of CPGs

several types of publications aim to translate research findings into recommendations for improving patient care, but that alone does not make it a cpg

14
New cards

Strength of recommendation ratings

Class I - strong benefit

Class IIa - moderate

Class IIb - weak

Class III - no benefit

Class IV - Harm

15
New cards

Level of evidence ratings

Level A - high quality evidece

Level B-R - moderate quality evidence

Level B-NR - non randomized

Level C-LD - limited data

Level C-EO - No evidence/ expert opinion

16
New cards

P< a

the null is rejected

the results are statistically significant

17
New cards

what does a p<0.05 mean?

there is a >95% chance that you correctly rejected the H0

18
New cards

the smaller the p value __________________

the harder it is to say the study results occurred just by chance alone

19
New cards

a statistically significant p value ______________

does NOT tell you anything about magnitide or clinical relevance

20
New cards

one sided hypothesis

specifies a direction

- more, greater, bigger, less, increased, etc

21
New cards

two sided hypothesis

specifies an effect exist, but does not predict what direction

22
New cards

type 1 error

false positive

you said there was a difference when actually there wasnt

alpha

23
New cards

type 2 error

false negative

you said there was no difference, when actually there was

beta

24
New cards

what is alpha

probability of making a type 1 error

"level of statistical significance"

25
New cards

statistical power

1-beta

liklihood you will observe an effect in your sample if one actually exists

- if not enough subjects are enrolled, the study is underpowered

26
New cards

what is effect size? ∆

statistical calculation that describes the size of a difference in outcomes between treatment groups

27
New cards

what does effect size tell us? (3)

1. magnitude

2. direction

3. practical significance / investigator determines if clinically meaningful

28
New cards

what is sample size?

describes the number of subjects necessary for the experimental and control groups

29
New cards

analyzing results - Intention to treat

includes data from every subject who was randomized

30
New cards

analyzing results - per protocol

only subjects with complete data sets are included

31
New cards

phase I of clinical trial

- low number of healthy volunteers

- unblinded, uncontrolled

- goal: test safety, PK, and max tolerable dose

32
New cards

Phase II of clinical trial

- hundreds of diseased patients

- can be single arm or comparative

33
New cards

phase III clinical trials

- thousands of patients

- large, randomized, controlled

- compare with the standard of care

34
New cards

Phase IV clinical trial

- post market surveillence

- evaluate side effects and additional therapeutic uses

35
New cards

hierarchy of evidence

1. meta analyses

2. systematic reviews

3. RCTs

4. cohort studies

5. case control studies

6. case series/case reports

36
New cards

what is a prospective study

- move forward in time

37
New cards

examples of interventional trials that are prospective

1. parallel

2. crossover

3. superiority

4. noninferiority

5. pragmatic

38
New cards

examples of observational studies that are prospective

prospective cohort studies

39
New cards

examples of observational retrospective studies

- retrospective cohort studies

- case control studies

40
New cards

RCTS are ______________, __________________

prospective, quantitative research studies

41
New cards

efficacy

- intervention under ideal, highly controlled conditions

-also known as explanatory studies

42
New cards

effectiveness

- intervention under real world conditions

- known as pragmatic studies

43
New cards

what factors are on the continuum of intervention studies

1. access to intervention

2. cost

3. patient adherence/acceptance

4. physician recommendation

44
New cards

populations in an efficacy study

highly specific study population that means all inclusion/exclusion critera

45
New cards

populations in an effectiveness study

broad, clinically relevant population

46
New cards

purpose of randomization

prevent, eliminate bias

47
New cards

types of randomization

1. simple

2. blocked

3. stratified

4. cluster

48
New cards

blocked randomization

evenly assign allocations in the blocks, randomly select blocks to fill the study

49
New cards

stratified randomization

stratify participants by confounding variables

50
New cards

cluster randomization

unit of randomization, not on the individual level

ex: all patients in a healthcare system

51
New cards

parallel RCT design

whole pop-> randomize -> treatment A or Treatment B

<p>whole pop-&gt; randomize -&gt; treatment A or Treatment B</p>
52
New cards

key elements of a parallel design

- groups are similar and treated identically except for intervention

- compare effects between treatments

- can examine magnitude of effect between groups

53
New cards

crossover rct design

population -> randomization -> treatment A or B -> washout -> treatment B or A

<p>population -&gt; randomization -&gt; treatment A or B -&gt; washout -&gt; treatment B or A</p>
54
New cards

key elements of crossover design

- all subjects receive all interventions

- washout period to prevent carryover

- can compare between and within groups

55
New cards

overall goal of a superiority study

show a meaningful difference exists between treatments

56
New cards

Null hypothesis and statistical interpretation of a superiority study

no difference exists

p values

57
New cards

overall goal of a noninferiority study

show a meaningful difference does NOT exist between treatments

58
New cards

Null hypothesis and statistical interpretation of a noninferiroty study

difference exists

confidence intervals

59
New cards

noninferiority margin ∆

predefined unfavorable difference in outcomes between new intervention vs control

60
New cards

exposure

when variable is naturally determined and or not assigned

61
New cards

intervention

preferred term when researchers assign participants to receive a specific treatment

62
New cards

goal of a cohort study

determine if exposure to a risk factor is associated with an outcome of interest

63
New cards

cohort studies are_______________, they start with _____________

always observational,

exposure, and watch for outcome

64
New cards

key feature of a cohort study

participants do NOT have outcome of interest to begin with

65
New cards

Prevalence

number of people in a population that have disease AT A GIVEN TIME

- permanent conditions

- person counted once

66
New cards

Incidence

Number of new cases over time

- transient conditions

67
New cards

prospective cohort study

start in present, watch for the development of outcome

68
New cards

In prospective cohort studies can incidence be calculated?

yes

69
New cards

retrospective cohort study

everything has happened, look back at records to see any outcome based on exposure

70
New cards

case control studies

- observational

- always retrospective

- start with OUTCOME

- good for rare outcomes

- cannot calculate incidence

71
New cards

estimation of risk

-Prospective = RR or OR

-cohort = RR or OR

- case control = OR only

72
New cards

strategies to manage sampling bias in case control studies

1. clinic or hospital nased controls

2. population based controls

3. 2+ control groups

4. matching

73
New cards

Risk ratio

number of events/ TOTAl Number

74
New cards

odds ratio

number of events / number of events not of interest

75
New cards

what does an odds ratio of 0.7 tell us

the outcome is less likely with exposure

there is a 30% reduction in risk of outcome based on exposure

76
New cards

Meta analysis

a systemic review that uses quantitative statistica methods to combine results of the separate studies

- calculate a summary statistic

- primary literature

77
New cards

what does a funnel plot assess in meta anlayses

publication bias

78
New cards

What are the benefits of matching (3)

1. eliminate confounding variable influence

2. increase power

3. provide convenience when selecting controls

79
New cards

waht are the common biases in cohort studies

attrition, selection, confounding

80
New cards

what are the common biases in case control studies

recall, sampling, confounding

81
New cards

forest plots

- box= point of estimate of effect size

- 95% CI

- size = weight given

- diamond = pooled results

82
New cards

heterogeneity

differences between studies

83
New cards

clinical heterogeneity

- participants

- interventions

- outcomes studied

84
New cards

methodological heterogeneity

- study desing

- risk of bias

85
New cards

tests of heterogeneity

- Cochrans Q test

- I^2

86
New cards

for X^2 test interpretation

p> alpha = insignificant heterogeneity

p

87
New cards

0-40%

might not be important

88
New cards

30-60%

may be moderate heterogeneity

89
New cards

50-90%

may be substantial heterogeneity

90
New cards

75-100%

considerable heterogeneity

91
New cards

sensitivity analysis

repeat the meta analysis, substituting alternative decision ranges to prove the findings are not dependent on arbitrary or unclear decisions

92
New cards

EER

cases/ total subjects in experimental group

93
New cards

CER

cases / total subjects in control group

94
New cards

RR

EER/CER

95
New cards

ARR

CER-EER

- usually reported as a %

96
New cards

RRR

1-(EER/CER)

97
New cards

NNT

1/ARR

Number of patients that would need to be treated in order to observe a specific health benefit

98
New cards

NNH

1/ARI

Number of patients that would need to be treated before one person is harmed

99
New cards

RR < 1

experimental group had lower rate of primary outcome

100
New cards

RR>1

experimental group had a higher rate of primary outcome