Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

SCS > Research > Flashcards

Research Flashcards

1
Q

Cross-Sectional Study

A

A study that collects data about a phenomenon during a single point in time or once within a defined time interval

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Efficacy vs Effectiveness

A

Efficacy = the extent to which an intervention produces a desired outcome under ideal conditions
Effectiveness = the extent to which an intervention produces a desired outcome under usual clinical conditions.

For instance, if a treatment is performed for TBI’s in a quiet environment or if performed in a noisier, typical rehab setting. One is focused on efficacy, the other effectiveness.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Experimental vs Non-Experimental Design vs. Quasi-Experimental

A

Experimental = the behavior of randomly assigned groups is measured after purposeful manipulation of independent variable (cause-effect of independent variable and outcome0
Non-Experimental (Observational) = the groups already exist and the experimenter cannot or does not attempt to manipulate an independent variable. The experimenter is simply comparing the existing groups based on a variable that the researcher did not manipulate. The researcher simply compares what is already established.
Quasi-Experimental = an empirical interventional study used to estimate the causal impact of an intervention on target population without random assignment.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Prospective Study

A

Research design that follows subjects forward over time

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Retrospective Design

A

Use historical data from sources such as medical records or outcomes database

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Single-System Design

A

Quasi-experimental design in which one subject receives in an alternating fashion both the experimental and control condition

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Levels on the continuum of research?

A

1a. SR of RCT’s
1b. RCT with narrow confidence interval
2a. SR of cohort studies
2b. Individual cohort study
3a. SR of case-control studies
3b. Individual case-control studies
4. Case-series or low-quality case control studies
5. Expert opinion

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Best study design for a special test

A

Cross-sectional non-experimental study of people suspected of having the issue

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Cohort Study

A

Prospective research design. One receives intervention and one doesn’t and monitored over time. Could be labeled quasi or non experimental.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Case-Control Study

A

Retrospective research design used to evaluate the relationship between risk factor and disease

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Basic science vs case report vs expert opinion

A

Basic science higher than the other 2 (which are on par)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

2x2 Contingency table comparing diagnostic test and reference standard

A

Top Left: True +
Top Right: False +
Bottom Left: False -
Bottom Right: True -

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Sensitivity (Def and Calculation)

A

True positive rate.

Tells us the probability a test is positive in someone we already know has the condition.

A/(A+C)

Very sensitive has relatively few false negative results.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Specificity

A

True negative rate.

Tells us the probability a test is negative in someone we already know does not have the condition.

D/(D+B)

Highly specific has relatively few false positive results.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Positive predictive value

A

A/(A+B)

Given a positive test result, the probability that the individual has the condition

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Negative Predictive Value

A

D/(C+D)

Given a negative test result, the probability that the individual does not have the condition

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Predictive values versus Sn/Sp

A

Predictive values seem to work like things work in the clinic: given a test result, what is the probability the result is correct. Sn/Sp work in the opposite direction: given the presence/absence of a condition (based on reference standard), what is the probability of a correct test?

Predictive values are highly dependent on the prevalence of the condition of interest in the sample. Sn and Sp remain fairly consistent across prevalence levels. Positive predictive values will be lower and negative predictive values higher in a sample with low prevalence (vice versa for high prevalence). Prevalence changes drastically depending on the setting where the test is performed.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

Likelihood Ratios

A

Tells us how likely a given test result is in people with with the condition compared with how likely it is in people without the condition.

Quantifies shift in probability once the diagnostic test results are known.
+LR: Sn/(1-Sp)
-LR: (1-Sn)/Sp

LR of 1 indicates the test does nothing to change odds.

Large shift: >10 or <0.1
Moderate shift: 5-10 or 0.1-0.2
Small Shift: 2-5 or 0.2-0.5

High + LR corresponds with high specificity. Low -LR corresponds with low sensitivity.

These are the best stats for clinical practice because they are used to quantify probability revision based on positive or negative test results.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

95% Confidence Interval

A

This is most common and indicates a range of values within which the population value would lie with 95% certainty.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

How are pre-test probabilities obtained?

A

From published data or the clinicians subjective impression/personal experience.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

How to calculate post-test probability instead of using nomogram?

A
  1. Estimate pre-test probability (say 70%)
  2. Convert this to pre-test odds by dividing probability by 1-probability 0.70/(1-0.70) = 2.3
  3. Multiply this by LR 2.3 x 42 = 96.6
  4. Convert post-test odds to probability by dividing odds by 1+odds 96.6/(96.1+1) = 0.99
22
Q

What are the 3 types of statistical studies?

A
  1. Sample
    - Estimating population parameter
  2. Observational
    - Seeing if there is a correlation between 2 things
  3. Experiment
    - Trying to establish causality
23
Q

PICO

A

A mnemonic for the key components of a well-focused foreground question

P = patient problem
I = intervention
C = comparison
O = Outcome

24
Q

Best Types of Studies For:
1. Diagnosis
2. Therapy
3. Prognosis
4. Harm/Etiology

A
  1. Cross-sectional study or blind comparison to gold standard
  2. SR, RCT, or Cohort Study
  3. Cohort study, case control, and case series
  4. Cohort, case control, or case series
25
Q

Hierarchy of Evidence

A
  1. MA (statistical technique for quantitatively combining the results of multiple studies that measure the same outcome.
  2. SR
  3. RCT
  4. Cohort Study (Can be prospective, retrospective, or combination of both. A cohort of individuals who don’t have an outcome are exposed to something thought to cause the outcome vs something not though to cause it and then followed for a time. Can look back at databases too for restrospective version)
  5. Case-Control Studies (looks to determine the association between an exposure and outcome. Those with outcome are compared to those without with respect to the exposure of the suspected harmful agent)
  6. Case Series/Reports (no control group)
  7. Animal / Lab Studies
26
Q

What are Boolean Terms

A

Use “AND” or “OR” to find the intersection of 2 or more sets or find results that include either of the criteria. Use parentheses if needed:

(Heart attack OR myocardial infarction) AND aspirin

27
Q

Important points for RCT?

A
  1. Was there randomization
  2. Was there allocation concealment
  3. Blinding (if you can’t blind clinician to Rx then a blinded researcher can get outcomes; if measures are very objective then it may be less important
  4. Attrition
28
Q

Drop out rates and how to resolve?

A

Good studies have >80% follow up with patients. When there is a large loss the lost patients should be assigned to “worst case” outcomes and results recalculated.

29
Q

ICF vs Nagi Model

A

The ICF shifts focus from the disease to the individual and a patient-centered approach. Nagi key terms were: pathology, pathophysiology, impairment, functional limitation, disability.

30
Q

Type 1 vs Type 2 Error

A

Type 1 is a false positive and type 2 is a false negative. So type 1 is when you reject the null hypothesis when it’s actually true. Type 2 is when you fail to reject a null hypothesis that is actually false.

31
Q

Null Hypothesis

A

The hypothesis that there is no significant difference between populations in a study.

32
Q

What is p-value?

A

The probability that a particular result is at least as extreme as the result observed when you assume the null hypothesis is correct.

33
Q

T-Test vs ANOVA?

A

T-test is used to compare the means of 2 groups, ANOVA is 3 or more.

34
Q

Types of t-tests?

A

Paired or Dependent t-test: the same group or item is tested twice (two thermometers on same people; same people’s scores on a test before and after an intervention)

Unpaired/Independent: compare 2 separate groups with equal variance (they have same starting point and you compare scores for the groups after intervention)

35
Q

What has to occur for t-test to produce valid results?

A
  1. Random sample
  2. Normal sampling distribution
36
Q

Types of ANOVA

A

One-way ANOVA: 3 or more groups for one factor
Two-way ANOVA: 3 or more groups but 2 factors (factor one: light, medium, intense exercise; factor two: males and females)

37
Q

Experimental vs Non-Experimental Research

A

Experimental = when an independent variable is manipulated

Non-Experimental = researcher does not manipulate independent variable.

38
Q

Types of Variables

A

Nominal = 2 or more categories but no order (Ethnicity)
Ordinal = 2 or more categories but ordered/ranked (Team rankings, Letter Grades)
Dichotomous = Only two choices (male/female; yes/no)
Interval = defines values along a scale and each point is equal distance but there is no zero point (temperature, time, year)
Ratio = is like interval but it does have a zero value (Height, weight, age; or a starting point that doesn’t have to be zero but you can’t get lower than this)

39
Q

Continuous vs Discrete Data

A

Continuous = it can take on any value in an interval and can be measured (temperature or speed of a car)

Discrete = it can only take on specifics values in an interval and can be counted (books on a shelf)

40
Q

What is the problem with using mean?

A
  1. It is particularly susceptible to outliers.
  2. Should not be used when data is skewed
41
Q

Normal Distribution vs Skewed

A

Normal distribution has mean, mode, median at the peak of the curve and both tails are completely symmetrical. Left/negative skew is longer tail to left and right/positive skew is longer tail to right. With skewed data the mode is at the peak of the curve, then median, then mean going down the longer tail.

42
Q

Mode vs Median

A

Mode is most frequent score
Median is middle mark

43
Q

Best measure of central tendency based on the variables?

A

Nominal = Mode
Ordinal = Median
Interval/Ratio = Mean (median when skewed)

44
Q

Normal Bell Curve and standard deviation.

A

This is for data that has a normal distribution with no bias left or right (heights of people, marks on a test, error in measurements).

68% of the values are within 1 standard deviation of the mean (34% on each side of center). 2 standard deviations has 95% of values within it. 3 SD = 99.7.

45
Q

Other names for standard deviation?

A
  1. Standard score
  2. Sigma
  3. z-score
46
Q

Standard deviation formula?

A

SD = (value to be standardized - mean) / The standard deviation

47
Q

Correlational Coefficient

A

Pearson’s “r”

Tells us how strong a relationship is between 2 variables. Goes from -1 to +1.

0-.2 = negligible relationship
.2-.3 = weak
0.3-.4 = moderate
.4-.7 = strong
.7+ = very strong

48
Q

What is power?

A

Refers to the number of patients required to avoid a type 1 or 2 error.

49
Q

Construct Validity

A

Does the test measure what it’s intended to measure? A construct refers to concepts/characteristics that can’t be directly observed but can be measured by observing other indicators associated with it. Does the test measure these indicators or something else. For instance, would a questionnaire looking at depression actually measure indicators of depression or having a bad mood?

Two subtypes are convergent and discriminant validity and both need to be established for construct validity:

Convergent: Does a test that is designed to measure a particular construct correlate with other tests that assess the same/similar constructs

Divergent: Whether two tests that should not be highly related are indeed not related.

50
Q

Content Validity

A

Assesses whether a test is representative of all the aspects of a construct. The measure must cover all relevant parts of the construct.

51
Q

Face Validity

A

This is subjective and considers how suitable the content of the measure seems on the surface. It’s similar to content validity but is more informal and subjective.

52
Q

Criterion Validity

A

How well a test can predict an outcome compared to another test (considered gold standard). There are 2 types of criterion validity:

Concurrent Validity: The scores of the test measure and the criterion are obtained at the same time.

Predictive Validity: The criterion variable are measured after the test measures are taken.

SCS (27 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions