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;
139 Cards in this Set
- Front
- Back
Reason for Measurements in PT
|
to provide the best clinical intervention possible
|
|
Assessments
|
"Tests" - provide the data for measurements - use them to evaluate changes in your patients for conditions (i.e. balance, strength, coordination...etc)
|
|
Measurements
|
numerical data - “Measurement is the act or process of quantifying some variable such as cognition, pain, blood pressure, force associated w/strength, liver enzyme levels, ROM, etc.”
|
|
Big Picture for tests and measures?
|
What you want to evaluate assessments (tests) measurements (data) statistical analysis evaluation & decide what to do
|
|
What are you ACTUALLY measuring?
|
Pounds, kilograms, ounces - NOT strength, force, or ROM
|
|
Critical Thinking
|
The art of analyzing and evaluating thinking with a view to improve it
|
|
Do journals and articles contain facts?
|
No - it is what we know at this time, and at this point of our research
|
|
Critical Thinking is?
|
Self-directed
Self-disciplined Self-monitored Follows a scientific method |
|
Egocentrism
|
The tendency to perceive, understand & interpret the world in terms of the self
|
|
Sociocentrism
|
The tendency to perceive, understand & interpret the world in terms of your society, culture or profession
|
|
Robert Fuller?
|
was a fool - did not use critical thinking in his alternative medicine study - body repairs itself - "healing by coincidence" - has no way to prove cause and effect
|
|
Scientific Method
|
1- empirical data is generated (objective)
2 - generation of hypothesis (null and alternative) 3 - experiments 4 - data is statistically analyzed 5 - reject or accept hypothesis 6 - generate theories |
|
Model
|
a series of steps or events that explains a process
|
|
Garrison's quote on models?
|
"Models are paradigms waiting to change & thus they are outdated once they are born"
|
|
Biggest problems with models?
|
you are trained to use certain models and you try to make all your thoughts and reasoning fit into those models (even though not everything always fits)
|
|
The 3 models for this course?
|
International Classification of Impairments, Disabilities, & Handicaps
Nagi Scheme ICF Model |
|
International Classification of Impairments, Disabilities, & Handicaps (ICIDH)
|
Disease -> Impairment -> Disability -> Handicap
|
|
Nagi Scheme
|
Active Pathology -> Impairment -> functional limitation -> disability
|
|
ICF Model
|
health conditions -> activities -> participation -> body structures/function -> environmental factors -> personal factors (circular in nature, everything relates)
|
|
Research
|
an objective, systematic investigation
|
|
analysis and interpretation of the data is done to?
|
gain new knowledge or add to existing knowledge
|
|
Null Hypothesis
|
there is no difference between the 2 conditions (hypothesis of no effect)
|
|
Alternate Hypothesis
|
there is a difference between the 2 conditions
|
|
Quantitative Research
|
considered a higher level of research than qualitative since the results can be generalized to the general population
|
|
Qualitative Research
|
descriptive of a population or sample being tested
|
|
Basic/pure/bench research
|
establish new knowledge in the development or refinement of theory
*quantitative/implies a laboratory situation |
|
Clinical research
|
involve human subjects
conduct clinical trials of new programs, products, drugs, and techniques |
|
Applied Research
|
quantitative research
designed to answer practical problems Ex: development of MRI machine |
|
Methodological Research
|
develop or design new changes between variables
*all tests and assessments are this type |
|
Descriptive Research
|
QUALITATIVE research
describe systematically a condition, observation, or area of interest |
|
Epidemiological Research
|
QUALITATIVE research
study the incidence, distribution, cause of disease, or impairment may describe 2 types of conditions |
|
Research Methodology
|
determines how you set up your experiment/research to evaluate the Null hypothesis
|
|
3 important pieces of research methodology
|
1. Manipulation
2. Control 3. Randomization |
|
Manipulation
|
the researcher changes (manipulates) 1 or more variables in connecting with the subject or condition
|
|
Variable
|
anything that can vary or change about the condition for the subject
|
|
Independent Variable
|
It is the variable which is manipulated which can be the experimental intervention/treatment variable
EX: hot/cold; drug |
|
Dependent Variable
|
this is the data (measurement) outcome, condition/appearance variable
EX: swelling of ankle; contusion |
|
Control
|
refers to the ability of the researcher to control of eliminate interfering and irrelevant influences from the study
(need to be able to say that the results of the experiment came from the change in the variable, not from an outside source) |
|
What is compromised if there is no control?
|
sensitivity, validity, reliability, and predictive value
|
|
How do you get around not being able to control everything?
|
Add a control group
|
|
Randomization
|
A process designed to reduce the risk of systematic bias from creeping into the study.
|
|
Internal Validity
|
the chance we are changing & measuring what we think we are changing & measuring
|
|
External Validity
|
the chance that results found in subjects can be applied to groups outside of the groups we are studying
|
|
3 categories of research protocols (methodologies)
|
True Experimental Designs
Quasi-Experimental Designs Non-Experimental Designs (qualitative) |
|
True Experimental Designs
|
must have: manipulations, randomization, and control
"cause and effect" research Can be double-blind studies |
|
Quasi-Experimental Designs
|
must have: manipulations, but not control or randomization
*opens study up to outside influence Typically, case studies, groups of people |
|
Non-Experimental Designs
|
no manipulations, randomizations, or control
generates questions for research good correlational studies |
|
Data Collection (measurements)
|
The numeric value (number) assigned to an object, event, interaction, observation or person according to rules (OPERATIONAL CRITERIA)
|
|
if you have rules...
|
you should be able to measure everything
|
|
Categories of Measurements
|
Fundamental
Derived Change |
|
Fundamental Measurements
|
obtained w/o the need for derivation (no math!)
EX: measuring ROM |
|
Derived Measurements
|
measurements of a variable (dependent) that are obtained as a results of math applied to the existing measurements
EX: Femur is 18 inches on L, 18.5 on R = 0.5 difference |
|
Change Measurements
|
mathematical difference between 2 of the same kinds of measurements taken on the same person at 2 points in time
EX: Pre and Post treatment data |
|
3 types of purposes for measurements
|
Evaluative
Discriminative Predictive |
|
Evaluative Purpose
|
can evaluate the effect of an intervention over time
"outcome measures" EX: Berg Balance Scale |
|
Discriminative Purpose
|
to discriminate some function, variable or activity among subjects or groups
EX: cognitive function among subjects - with a test |
|
Predictive Purpose
|
using a measurement to say something about future events or creating a prognosis
EX: Berg Balance can predict balance in the future |
|
Qualitative Data
|
alphanumeric
comprised of letters or characters which may be digits "Character or Categorical" does not support anything descriptive stats - mean, mode, etc. for categories |
|
Quantitative Data
|
always numbers with quantities
does NOT have to be whole numbers measures should be standardized (reliable) discrete/cardinal (whole numbers!) continuous (any value along a continuum w/in a range) |
|
Scales used to measure?
|
Nominal
Ordinal Interval Ratio |
|
Nominal Scale
|
Qualitative
lowest level of refinement descriptive stats EX: person can stand, or not NONPARAMETRIC DATA |
|
Ordinal Scale
|
ranking scale
implies a greater or lesser degree of something EX: hate a lot, hate a little, its OK, like a little, love it *no equal increments! |
|
Interval Scale
|
data ranked in a logical sequence
EQUAL increments in data EX: ROM (degrees), height (inches) there is NO absolute zero |
|
Ratio Scale
|
highest level of scales
continuum of values (like interval) has an ABSOLUTE ZERO EX: if you get a zero, you don't know anything about the subject PARAMETRIC DATA |
|
Validity
|
refers to the degree to which a test (assessment), intervention, or instrument measures what is supposed to be measuring
*a matter of spectrum (not an all or none thing) |
|
Stats Definition
|
to extract the maximum amount of information about a set of data (measurements)
|
|
External Validity
|
can we generalize the results of an assessment to a similar population
|
|
Internal Validity
|
concerned w/correctly concluding that an independent variable is, in fact, responsible for variation in the dependent variable
(needs good controls - randomization - manipulation) |
|
Construct Validity
|
based on the knowledge & intellectual underpinnings, which are considered the CONSTRUCT, upon which the test & measurements are developed
|
|
Content Validity
|
related to the extent to which a measurement reflects the specific intended domain of content
Ex: testing for UE body strength: cannot say that you are testing for body strength b/c you only tested the UE; the content validity is not good |
|
Criterion-Based Validity
|
"instrumental validity"
involves comparing the measurements being examined w/another measurement or a series of other measurement or procedures which have been demonstrated to be valid |
|
Three types of criterion-based validity?
|
Concurrent
Predictive Prescriptive |
|
Concurrent Validity
|
when an inferred interpretation is justified by comparing a measurement w/supporting evidence that was obtained at approx. the same time as the measurement being evaluated (i.e. concurrently)
*more precise than criterion due to time frame |
|
Predictive Validity
|
concerned w/using criterion to make predictions which are true
*used in many screening tests |
|
Prescriptive Validity
|
concerned w/using the inferred interpretation of criterion (measurement) from a test to prescribe a treatment
|
|
Face Validity
|
how a measure or assessment appears
non-statistical variety of validity "does this data seem reasonable?" |
|
Convergent Validity
|
it refers to the degree to which a measure is correlated w/other measures that it is theoretically predicted to correlate with
EX: multiple testing for the same outcome - should align results for the patient/population |
|
Reliability
|
the degree to which measurements of a test remain consistent over repeated test of the same subject under identical condition
(repeatability) |
|
Inter-tester Reliability
|
consistency between different people measuring the same thing
indicates a correlation between testers measures TESTERS - not test |
|
Intra-tester Reliability
|
consistency or equivalence when one person repeated measurements over a period of time
|
|
Test-Retest Reliability
|
consistency of repeated measurements in time
indicates stability in test measure TOOL/ASSESSMENT |
|
Population
|
total number of individuals, measurements, or units from which data will be collected or generalized about
|
|
Sample of Population
|
o Rarely can you deal with a whole population so you must use a sample.
o In general when dealing with statistics you are dealing with a sample of the population |
|
Three Levels of Data Analysis
|
Descriptive
Correlative/Trend Comparative |
|
Descriptive Data Analysis
|
lowest level
mean/mode/frequency etc... qualitative research |
|
Correlative/Trend Data Analysis
|
Describes relationship of changes of one variable with changes of another variable
Middle level *correlation of coefficients* can extrapolate results to population (quantitative) |
|
Comparative Data Analysis
|
Determines whether 2 or more groups of data are different or not
"cause and effect" highest level |
|
Parametric Statistical Test
|
Tests which are run when the data (measurements) comes from a normal distribution (Bell Shaped Curve)
"data clusters around the mean" mode and median are values around the mean standard deviations included in curve |
|
Non-Parametric Statistical Test
|
Tests which are run when the data (measurements) DO NOT come from a normal distribution (Bell Shaped Curve)
- data is ordinal or nominal - sample sizes are small - used when normal distribution cannot be assumed |
|
Both distributions can involve what type of stats?
|
descriptive and inferential
|
|
Descriptive Stats
|
(those which describe, organize & summarize data)
Describe things so you cannot compare things or extrapolate the data to anyone else |
|
Frequency
|
the # of occurrences of a repeating activity based on a unit of time
|
|
Percentages
|
a way of expressing a number as a fraction of 100
|
|
Percentiles
|
values such that a specified percent of the data falls above or below a value.
|
|
Prevalence
|
the total number of cases in the population or sample at a given time
usually a percentage |
|
Incidence
|
a measurement of the number of new individuals who develop a disease of condition w/in a particular period of time
normally a percent* |
|
Central Tendency
|
o Mean: arithmetic average
o Mode: the value which occurs most frequently o Median: the value which separates a sample from the upper half & the lower half. |
|
Relative Position
|
Range
Standard Deviation Standard Error of Mean |
|
Standard Deviation
|
measure of the variability of a population, sample or probability distribution
*want them to be low (low is closer to the mean) |
|
Standard Error of Mean
|
it quantifies the certainty with which the mean computed from a random sample estimates the true mean of the population from which the sample was drawn
*more accurate than SD (extra step) |
|
Inferential Stats Test
|
Test which uses data from samples drawn from a population to make inferences about the total population (3 types)
|
|
Student t-test
|
parametric test
need normal distribution 2 types: unpaired/paired |
|
Unpaired t-test
|
one sample t-test
used to test whether the mean drawn from a normal population differs from a hypothesized value |
|
Paired t-test
|
whether the means of 2 groups are different - samples drawn in pairs/ or are related
|
|
Analysis of Variance (ANOVA)
|
an extensive class of related statistical models (tests) & their associated procedures, in which the observed variance (SD & SEM) is separated into categories due to different independent variables
|
|
Statistical tests which normally involve 3 or more independent variable and only 1 dependant variable
|
ANOVA
|
|
Correlation Coefficients
|
an index of the degree of association between 2 variables or the extent to which the order of individuals on 1 variable is similar to the order of individuals on a 2nd variable
|
|
Linear regression analysis
|
establishes a mathematical relationship between 2 or more variables
|
|
• Strong correlations of data do not necessarily prove
|
cause and effect
|
|
Pearson's Correlation Coefficient
|
quantifies the strength of association between 2 variables that are normally distributed
Parametric b/c it comes from a normal distribution Usually shown by ‘r’ in papers Used a lot with a true experimental design |
|
Interclass Correlation Coefficient (ICC)
|
it demonstrates the consistency of measurements when 1 or more raters takes the measurements
- consistency/conformity between multiple testers |
|
Spearmen Rank Correlation Coefficient
|
it is used to quantify the strength of association between 2 variables that are measured on an ordinal scale
Want to see if 2 variables are related in a correlation in a positive/negative manner. Non parametric test |
|
Cronbach's Alpha
|
frequently used as a measure of the internal consistency reliability of an assessment
How close is the dependent variable to being close each time you measure Measure of internal consistency, how accurate |
|
Cohen's Kappa Coefficient
|
it is a measure of inter-rater agreement for qualitative (categorical) items
QUALITATIVE only |
|
True Positives
|
sick subjects who have the disease
|
|
False Positives
|
healthy subjects wrongly identified as having the disease
|
|
True Negatives
|
healthy subjects correctly identified as not having the disease
|
|
False Negatives
|
sick subjects incorrectly identified as not having the disease
|
|
Sensitivity
|
A value which indicated the proportion (percentage) of actual positives which are correctly identified as being positive (TRUE POSITIVE)
Formula: No. of True Positives/No of True Positives + No. of False Negatives |
|
Specificity
|
A value which indicates the proportion of negatives which are correctly identified as being negative (TRUE NEGATIVE)
Formula: No. of True Negatives/No. of True Negatives + No. of False Positives |
|
Relative Standards
|
If you don’t want to use standardized tests you may choose to use relative standards, which means you use the subject as their own control & not some standardized chart
NOT relevant to a population - only to patient |
|
P value
|
In accepting the Null Hypothesis (no difference exists) you need to know ahead of time what kind of chance you are willing to take of being wrong. This is the P value.
|
|
Type 1 Error
|
(Alpha Error/False Positive - synonyms): you reject the null hypothesis when the null hypothesis is true
|
|
Type II Error
|
(Beta Error/False Negative - synonyms): You fail to reject the null hypothesis when the null hypothesis is false
|
|
Cut-off Point
|
Can be anything that we determine to be indicative of a problem or a non-issue in terms of patient presentation.
Subjective points used to base decisions about whether or not a person has a condition, is eligible for a specified intervention, or needs to be referred for further testing. |
|
Z-score
|
how many standard deviations a person is above or below the mean
|
|
Norm-referenced cut-off points are?
|
The BEST - then criterion, then arbitrary (i.e. pain)
|
|
Positive Predictive Value
|
A proportion of individuals identified by the cutoff as being abnormal who are classified as having a target condition by a criterion measure
|
|
Negative Predictive Value
|
A proportion of individuals identified by the cutoff as being normal who are classified as not having a target condition by a criterion measure
|
|
Sources of instability in a study
|
verification bias
small samples errors in criterion measure construct irrelevant variance |
|
Likelihood Ratio
|
Incorporates sensitivity and specificity
Provides a direct estimate of how much a positive or negative test result will change the likelihood of having a condition or disease. |
|
Positive Result of Likelihood Ratio (LR+)
|
how much the likelihood of the condition increase when a test is positive.
LR+ = Sensitivity/1-Specificity |
|
Negative Result of Likelihood Ratio (LR-)
|
how much the likelihood of the condition decrease when a test is negative.
LR- = 1-Sensitivity/Specificity |
|
Minimal Detectable Change
|
Minimal change that could be attributed to intervention vs. error
|
|
Minimally Important Difference
|
Clinically relevant change; magnitude of change that is meaningful or change in function beyond natural progression.
|
|
Clinical Significant Change
|
A change that is recognizable to peers and others
A proportion of persons who show improvement A proportion of elimination of the presenting problem Minimally clinically significant change |
|
Z-score for the mean is always?
|
0 (zero)
|
|
Z-score for the standard deviation is always?
|
1 (one)
EX: Z-Score of -1.4 = 1.4 standard deviations below the mean |
|
How do you calculate a Z-score?
|
X- x/sd
E.g. raw score (X) = 15, mean (x) = 10, standard deviation (sd) = 4 (15 -10/4) – DO NOT DO ORDER OF OPERATIONS • Z-Score = 1.25 |