Drug Lit Exam 2

Front 1

observational study where 2 existing groups differing in outcome are identified and compared on the basis of some supposed causal attribution

Back 1

Case Control trials

Front 2

Advantages of case control studies (7)

Back 2

1. Useful for hypothesis generation
2. Can study rare events or outcomes such as adverse drug reactions
3. Can study events or outcomes that develop only after many years
4. Frequently less expensive to conduct
5. Takes less to complete
6. May be able to find association with a relatively small sample
7. Avoids ethical problems of gathering data

Front 3

Disadvantages of case control studies (6)

Back 3

1. Can only study one outcome
2. Subject to multiple types of sampling and measurement biases
3. cannot directly measure either incidence or prevalence
4. information available is often limited
5. cause and effect may be difficult to determine because time sequencing and confounding
6. numerous examples of false associations have been documented using case control studies

Front 4

Selection of cases for Case Control studies

Ideal is a _____ sample of the total population with the disease

Back 4

random

Front 5

Selection of controls for Case Control should ideally have these characteristics (3)

Back 5

1. Same baseline risk for the development of the disease.
2. Same probability of exposure to the risk factor as the cases
4. Same degree of diagnostic surveillance as cases

Front 6

Advantages to population controls in a case control study (2)

Back 6

1. cases and controls drawn from the same population should be comparable
2. distribution of risk factors in the sample should be representative of the population

Front 7

disadvantages of population controls in a case control study (3)

Back 7

1. if the population is defined too narrowly it may stil not represent the entire population of interest
2. may be less convenient than simple clinic or hospital sample
3. population controls may be less motivated to cooperate

Front 8

disadvantages to hospital/medical practice controls in case control study (2)

Back 8

1. cases and controls may have other diseases implying frequency of baseline exposure to the risk factor may not be the same as a randome sample of the base population that produced the cases
2. catchment area (i.e. population drawn) for different diseases in a hospital may not be the same resulting in mixed demographics

Front 9

case control studies

Use____ controls per case because power to detect differences is increased (rare diseases). Beyond this number offers little statistical advantage.

Back 9

Up to 4

Front 10

A proportion that compares the number of times that an event occurred to the number of times that it did not occur

Back 10

Odds

Front 11

Total number of new events over a period of time

Back 11

Incidence aka risk

Front 12

Given a rare condition...Odds ration _____ the relative risk

Back 12

odds ration approximates the relative risk

Front 13

estimates the magnitude of the association between exposure and disease

Back 13

odds ratio

Front 14

Odds ratio...

A value of _____ indicates the ratio of the odds of exposure are identical in the case and control group (i.e. no association)

Back 14

1

Front 15

A value of ____ indicates an increased odds of exposure among cases as compared to controls (i.e. harmful exposure effect)

Back 15

> 1

Front 16

A value of ____ indicates a decreased odds of exposure among cases as compared to controls (i.e. protective exposure effect)

Back 16

<1

Front 17

potential problems with observational studies (4)

Back 17

1. confounding
2. effect modification
3. bias
4. chance

Front 18

confounding creates (2)

Back 18

1. bias
2. systematic error

Front 19

approaches to control confounding in the design of the study (2)

Back 19

1. exclude subjects with known confounding factors
2. match cases and controls based on potential confounding factors

Front 20

approaches to control confounding in the analysis of the study (2)

Back 20

1. stratify on the confounding factors (eg. age, weight, sex, disease severity)
2. use multivariate analysis to adjust results

Front 21

Advantages to matching (3)

Back 21

1. May reduce confounding by strong factors such as age, race, sex
2. May increase precision (power)
3. May be sampling convenience

Front 22

Disadvantages for matching (4)

Back 22

1. Exclusion of cases that don't have a match
2. Longer study duration
3. Increased cost
4. Matched variables cannot be analyzed as a risk factor

Front 23

Reason for self-referral may be associated with the outcome under study

Back 23

self selection sampling bias

Front 24

Membership in a group (employed, joggers, clubs) may be a marker for other health-related factors or factors related to access to health care

Back 24

Membership sampling bias

Front 25

Time gap between exposure and selection of cases, some actual cases may be missed if the condition is potentially clinically silent and short duration, or if rapidly fatal and not properly recorded, or evidence of exposure is lost with time

Back 25

Prevalence-incidence (Neyman) sampling bias

Front 26

Exposure to the risk factor causes a sign or symptom that leads to increased diagnostic surveillance--which increases the chance of being diagnosed

Back 26

Unmasking (detection signal) sampling bias

Front 27

May occur in hospital or medical practice controls, if perosns who are with disease, without the disease and with the characteristic of interest have different probabilities of being admitted

Back 27

Admission rate (Berkson) sampling bias

Front 28

Knowledge of the subjects prior exposure to a suspected risk factor may influence the intensity and the conclusion of a diagnostic evaluation

Back 28

Diagnostic suspicion measurement bias

Front 29

Knowledge of the patients disease status may influence both the intensity and the conclusion of the search for exposure

Back 29

Exposure suspicion measurement bias

Front 30

cases and controls may have different levels of recall or events based on different level of interest or on different levels of questioning by the investigator

Back 30

Recall measurement bias

Front 31

Responses to family hisotry information has been shown to vary considerably depending on whether the subject was a case or control, cases recall much higher rates of disease or risk factors present

Back 31

Family information measurement bias

Front 32

Comparison of a standard or conventional therapy (active entity) to a new therapy

Back 32

Active Comparator Trial

Front 33

Why would you choose to do an active comparator trial? (2)

Back 33

1. placebo is not possible or is unethical
2. new treatment has a potential advantage over proven treatment (lower cost; less adverse effects; simpler dosing regimen; new treatment less invasive)

Front 34

To determine if new therapy is superior to standard or conventional therapy

Back 34

superiority trial

Front 35

to determine if new therapy is no worse than or as good as the standard therapy by pre-defined margin

Back 35

non-inferiority trial

Front 36

New treatment is similar to standard therapy (not better, not worse)

Back 36

equivalence trial

Front 37

The maximum allowable difference between two treatments; the difference in treatment cannot be outside of the boundaries of the margins to claim "non-inferiority"

Back 37

non-inferiority margin

Front 38

Factors that can dilute true differences between treatments (7)

Back 38

1. imprecise or poorly implemented entry criteria
2. poor blinding
3. poor compliance
4. drop-outs
5. recruitment of patients unlikely to respond
6. treatment crossovers
7. use of concomitant medications (treatments)

Front 39

objectives of a meta-analysis (7)

Back 39

1. summarize and integrate results from a number of individual studies
2. analyze the differences in results among studies
3. increase power by combining data from several studies
4. increase sample size when individual studies are too small to identify an effect
5. evaluate effects in subsets of patients
6. generate hypothesis for further research
7. avoid RCT when precluded by ethical issues

Front 40

A 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 the studies that are included in the review

Back 40

Systematic Review

Front 41

Use of statistical techniques to integrate and summarize the results of included studies. Many systematic reviews contain these, but not all. By combining information from all relevant studies, these can provide more precise estimates of the effects of healthcare than those derived from the individual studies included within a review

Back 41

Meta Analysis

Front 42

When analyzing a narrative or systematic review what should you look for? (3)

Back 42

1. flow diagram, checklist (PRISMA)
2. structured abstract/title
3. risk of bias (publication bias, selection bias, statistical heterogeneity, clinical heterogeneity, sensitivity analysis, subgroup vs. sensitivity analysis)

Front 43

bias towards publication of positive results

Back 43

publication bias

to minimize:
1. multiple databases should be searched--give search dates and specific search terms
2. abstracts
3. bibliographies
4. unpublished data
5. contact experts in the field

Front 44

what tool is used to check publication bias?

Back 44

funnel plots

Front 45

what tests are used to detect publication bias? (2)

Back 45

1. Egger weighted regression method
2. Begg rank correlation method

Front 46

studies selected for inclusion in the analysis influence results

Back 46

selection bias

Front 47

Describe the degree of between-study variability in a group of studies

Back 47

heterogeneity

Front 48

considers within-study variability; assumes that the true effect of treatment is the same in each study

Mantel-Haenszel method

Back 48

fixed-effects model

Front 49

considers both between-study and within-study variability; assumes that the studies used are a random sample from the universe of all possible studies

DerSimion-Laird Method

Back 49

Random-effects model

Front 50

Random effects models are more conservative and generate a _____ and are less likely to show a significant treatment effect compared to a fixed effects model

Back 50

confidence interval (wider)

Front 51

____ is assumed to mean that all variability in effect size estimate is due to sampling error within the studies

Back 51

I2 (squared) = 0

Front 52

___ means that total variability in effect size is not caused by sampling error but by true heterogeneity between studies

Back 52

I2 (squared) = 50

Front 53

heterogeneity may result from differences in (4)

Back 53

1. patients (patient selection procedure; baseline disease severity; different age groups; gender differences)
2. treatments (drugs/doses used)
3. duration of follow-up
4. disease states treated

**may also be related to unknown or unmeasured confounders