Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
97 Cards in this Set
- Front
- Back
|
what are two ways to measure the evidence of published research documents?
|
can look at DOEs (disease orientated evidence) or POEMs (patient orientated evidence)
|
|
|
Level of evidence is based on:
|
experimental design
|
|
|
Strength of recommendation is based on:
|
a large body of evidence
|
|
|
Strong evidence from at least one properly designed randomized controlled trial of appropriate size is what type of evidence strength?
|
Type 4
|
|
|
What are the criteria for a good research question?
|
FINER
- Feasible - Interesting - Novel - Ethical - Relevant |
|
|
Experimenta design refers to the structure of an experiment, specifically though to 3 things?
|
1. experimental procedures
2. selection of subjects 3. allocations of subjects to different experimental procedures |
|
|
What is the difference between a single blind and double blind study?
|
in single blind the researcher does not know what subjects are controls and which are being treated; in a double blind, neither the researcher nor the subjects know.
|
|
|
LOEs studies assign what type of grading system to their evidence?
|
Numerical rank (1,2,3)
|
|
|
SORs studies assign what type of grading system to their evidence?
|
ABCs
- based on body of evidence, - types of outcomes, - number, - consistency, - coherence of evidence as a whole, - relationship between benefits, harm & cost |
|
|
what is the sequence and cycle of research?
|
1. Choosing the research question
2. Developing the protocol 3. Pretesting and revising the protocol 4. Carrying out the study 5. Analyzing the findings 6. Drawing and disseminating conclusions |
|
|
A sham procedure is:
|
a procedure that resembles the intervention but lacks any effective treatment thus elminating a confounder
|
|
|
what type of design validity addresses the study impact on the larger population as a whole?
|
external validty
|
|
|
Internal validity is a measure of:
|
how well the study was carried out - the degree to which conclusions correctly apply to the study and its purpose
|
|
|
what is the goal of sound experimental design?
|
elimination of, or at least an accounting for, confounded effects
|
|
|
What are the keys to avoiding confounded effects?
|
1. sampling technique
2. sample size 3. randomization |
|
|
If all confounding effects are equalized among groups, then only the ________ will affect the _____________.
|
predictor variable will affect the outcome variable
|
|
|
When is randomization not a helpful design component?
|
when a certain population needs to be studied - i.e. emphysema rates in smokers. This is an opportunity to use experimental model
|
|
|
What is viewed as a stronger study - a retrospective or prospective study?
|
prospective
|
|
|
What are 2 key strengths to a prospective-style study design?
|
1. can assess more extraneous variables
2. employs randomization & possible control |
|
|
What is the strongest study design?
|
randomized control trial
|
|
|
What type of study design is the best accounting for cofounders?
|
RCT
|
|
|
T/F Prospective cohort studies have fewer ethical constraints than RCTs.
|
True
|
|
|
What are 4 criticisms of (prospective) cohort studies?
|
1. not necessarily random nor representative
2. other associated factors may be operative since a cohort may already be self selected 3. control groups may be small 4. no subject blindness |
|
|
what are the limitations of a retrospective study?
|
cohort limitations
i.e. there could be many other extraneous factors that influence your cohort & the outcomes that you don't even know about or know how to assess. Also, no randomization |
|
|
How does a prospective study differ from a prospective cohort study?
|
Prospective study follows one group for a defined period of time with an observation of outcomes. There is no manipulation of a variable or of experimental or control groups
|
|
|
Give an example of a prospective survey:
|
using an HMOs database to identify pts with a dx of alcohol abuse to track these patients to see if they develop cirrhosis as reflected by their diagnosis codes
|
|
|
What type of study offers a feasible approach when outcome is rare or the time period is long?
|
Case-control study
|
|
|
A case control study is a _________ study
|
retrospective
|
|
|
A case control study is looking for the identification of:
|
predictors by association
|
|
|
Criticism of case-control studies includes:
|
problematic confounders
small sample size |
|
|
What is an example of a case-control study?
|
looking at a skin cancer vs. control group to determine past associations with tanning bed use, other associations with perceived risk facts
|
|
|
A cross-sectional survey is___ group, ____ time.
|
one group one time
|
|
|
Two benefits of a cross sectional survey are _____ & _____ which lead to it being performed frequently.
|
lower cost, easier logistics
|
|
|
What is an example of a cross-sectional survey?
|
a mailed questionnaire study assessing the association of gender and health care services utilization
|
|
|
Which of these designs is not retrospective?
a. RCT b. Case control c. Retrospective cohort |
A. RCT
|
|
|
A before and after study design measures:
|
outcomes before and after an intervention is performed
|
|
|
T/F In before and after design studies, the subject is also the control.
|
true
|
|
|
Does before and after study designs allow for good accounting of cofounders?
|
No
|
|
|
What are necessary components in a study design that should be reported in the methods section?
|
design type
population of interest subject selection (included & excluded) variable definitions (predictor [independent], outcome [dependent) assessment for baseline demographic characteristics/extraneous variables (including possible cofounders) |
|
|
A cohort study can be either prospective or retrospective.
|
True
|
|
|
Which of the following designs cannot be retrospective?
a. cohort b. case-control c. RCT |
c. RCT
|
|
|
For a given research question, one of a few experimental designs might be employed. T/F
|
True
|
|
|
Which study design best accounts for confounding variables?
a. RCT b. prospective cohort c. survey |
a. RCT
|
|
|
How can variables be classified?
|
As categorical data (like nominal or ordinal data) or as interval data (continuous data)
|
|
|
A nominal variable has:
|
no implied or natural ordering
|
|
|
Give an example of a nominal variable.
|
gender
race marital status hand-dominance |
|
|
An ordinal variable is ______ categories.
|
ordered
|
|
|
T/F in ordinal variables, categories are present but differences between categories must be equal
|
false - differences between categories may not be equal. If equal, then interval variable
|
|
|
Which of the following variables is not interval data?
a. A1C levels b. K+ levels c. acceptable A1C |
c. Acceptable A1C
|
|
|
Interval variables have ________ between any two levels on a scale.
|
equal numerical distances
|
|
|
Data can be classified in what 2 ways?
|
discrete or continuous
|
|
|
when performing data analysis, what are 2 important characteristics to examine?
|
central tendency
measure of dispersion |
|
|
What data set characteristic means "average"?
|
central tendancy
|
|
|
What data set characteristic means "range"?
|
measure of dispersion
|
|
|
What are commonly used measures of central tendency for a distribution?
|
mean
mode median |
|
|
The mean is the:
|
average
|
|
|
The median is the:
|
middle number
|
|
|
The mode is the:
|
most commonly occurring category or value
|
|
|
A histogram differs from a bar graph because:
|
a histogram represents a range of data and a bar graph represents one data point
|
|
|
What does the best measure of central tendency for a data distribution depend on?
|
characteristics of the distribution! i.e. symmetry, outlier data
|
|
|
When will you find the median, mean and mode being equal?
|
symmetrical distribution
|
|
|
What is the best measure for:
data showing number of patients in Stages I, II, III, & IV of cancer? |
mode - the choices are categorical and not numerical
|
|
|
What is the best measure for: data summarizing numbers of 1,2,3,4 and 5 as responses for questionnaire item?
|
median - this is discrete data with a limited ordinal range
|
|
|
What is the best measure for: data distribution of serum sodium concentration?
|
mean - this is continuous data with an expected symmetrical distribution without significant outliers
|
|
|
What is the best measure for: census data on household income?
|
Median or mean might be acceptable depending on context. If outliers present, the median might be more useful
|
|
|
Ordinal data (ordered) is more likely to use what measure of central tendency?
|
median
|
|
|
When data is skewed, what is a good measure of central tendency?
|
median
|
|
|
What are commonly used measures of dispersion?
|
range
quartile standard deviation variance |
|
|
Range is the difference between:
|
the highest and lowest values
|
|
|
What are the two reasons that a quartile is commonly used to measure a dispersion?
|
- if distribution is for an ordinal variable
- distribution is not symmetrical |
|
|
Inter-quartile range is the difference between the:
|
75th and 25th percentile values
|
|
|
For discrete data for an ordinal variable, measures of central tendency and of dispersion are most likely to be:
|
median and quartiles
|
|
|
What is variance?
|
a measure of deviation of individual values from the mean
|
|
|
What is standard deviation?
|
a measure of average deviation of individual values from the mean
|
|
|
T/F The actual parameter values for the mean and standard deviation of the population of interest are easy to obtain and should be noted for every data set.
|
False - we typically do NOT know the actual parameter values for the mean and SD
|
|
|
The variance and standard deviation are common measures of dispersion for sets of:
|
continuous data for interval variables
|
|
|
T/F
Under certain conditions (like certain experimental designs, adequate scale width, adequate sample size, symmetry) data for an ordinal variable may be analyzed as continuous data. |
true
|
|
|
What is the difference between discreet and continuous data?
|
discrete data have values that may assume only whole numbers; continuous data may take any value within a defined range for interval variables
|
|
|
would nominal data be better displayed on a bar chart or histogram?
|
bar chart
|
|
|
would ordinal data be better displayed on a bar chart or histogram?
|
histogram
|
|
|
How is continuous data best displayed?
|
smooth curve (frequency polygon)
|
|
|
A bell-shaped symmetrical data distribution implies:
|
interval data and continuous data
|
|
|
For continuous data for an interval variable, measures of central tendency and of dispersion are likely to be:
|
mean and standard deviation
|
|
|
A standard normal curve is a member of the normal curve family with a mean of ___ and SD of ___.
|
Mean = 0.0
SD = 1.0 |
|
|
z =
|
z= (score - mean/SD)
|
|
|
taking data from numerical data to a distribution that is useful for estimating percents or probabilities based on a curve is termed:
|
data transformation
|
|
|
z score also is known as:
|
standard score
|
|
|
The standard score is how many:
|
SD units a point is from the mean
|
|
|
An area under a curve that refers to a specific proportion/percent is called:
|
alpha
|
|
|
One of the uses of a normal curve is to be ________ about the data that it contains
|
descriptive
|
|
|
The z-score is a result of:
a. data transformation b. division of a data point by the mean |
a. data transformation
|
|
|
The unit of "z" is:
a. SD, a measure of dispersion b. sample means c. population means |
a. SD, a measure of dispersion
|
|
|
The SND may be described as:
a. various bell-shaped curves with a known mean and SD b. various smooth curves with a mean of 0.0 c. a tool for transforming a data distribution to a distribution with known mean, dispersion, and probabilities |
c.
|
|
|
A retrospective cohort is made up of a group that is chosen for the presence or absence of...
|
a particular variable in the past
|
|
|
T/F
Cause and effect are less certain in cross-sectional surveys |
true
|
|
|
T/F There is no randomization in a before-after design
|
true
|
|
|
categorical variables are _____ data while interval variables are_____ data.
|
categorical = discrete
interval = continuous |
