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;
75 Cards in this Set
- Front
- Back
|
Background questions
|
Assist clinicians in understanding that pathology, impairments, functional limitations, and disabilities, course of disease, prevalence, and incidence of a health problem
|
|
|
Foreground questions
|
Help clinicians and their clients make decisions about the specific PT management or issue
Ex. Diagnostic question, Intervention question, Prognostic question |
|
|
Diagnostic question
|
support you in determine the reliability and validity of measurement instruments
|
|
|
Intervention question
|
Focus on the comparison of the effects of a treatment condition compared to a control condition (no treatment) and/or a reference condition (other treatment)
|
|
|
Prognostic question
|
focus on the accuracy of factors that determine the course of a disease process and/or rehabilition outcome
|
|
|
Variable
|
A factor measurable and expected to change or have different levels/
|
|
|
Four Scale of Measurements
|
1) Nominal scale
2) Ordinal scale 3) Interval scale 4) Ratio scale |
|
|
Nominal scale
|
Ex. gender, color, yes/no
hot, warm, cool, cold |
|
|
Ordinal scale
|
Don't know the interval between each rank. Still have an order though.
Ex. pain (Visual analog scale) Ex. 1-10 with 1=cold 10=hot |
|
|
Interval scale
|
Fixed scale
Ex. Temp: Fahrenheit scale, negative to positive values (-10 to +70 degrees) |
|
|
Ratio scale
|
Absolute zero
Kelvin scale, 1 (absence of heat to positive values) |
|
|
If you have a measurement instrument that uses interval or ratio scale...
|
parametric statistics can be applied (assume normal distribution)
|
|
|
If you have a measurement instrument that uses nominal or ordinal scale...
|
non-parametric statistics (median and ranges)
|
|
|
Frequency of distribution
|
3 patients score high, 14 patients score low and 25 patients scored in the middle of the range.
(Data doesn't follow normal curve-- skewed) |
|
|
Mean
|
86 patients were treated by 12 different therapists in 1 day, with the result of 7.16 patients per therapist.
Avg of values of data points x = set of observations / # of observations |
|
|
Variance
|
James saw 12 patients in 1 day. He treated 4.84 patients more than the mean.
(Assume normal distribution) s^2 = sum (x^2)/ N --- variance for the entire population |
|
|
Standard deviation
|
Square root of variance
|
|
|
Normal curve
|
Bell shaped frequency distribution
|
|
|
Correlation Coefficient
|
The numeric relation between the number of patients treated on day 1 and day 2
less error: correlation co-efficient will be closer to 1 -1 (or negative value): inverse correlation |
|
|
SEM (standard error of measurement)
|
The SD of repeated measures of strength as measured on Kincom.
It is a measure of the "spread" of scores within a subject had the subject been tested repeatedly. How is a given score related to a "true" score How you can quantify error over time How much might a score vary with repeated measurements within one individual? SEM = square root (1-r) r is the intra-class correlation co-efficient between measures |
|
|
Frequency of distribution
|
3 patients score high, 14 patients score low and 25 patients scored in the middle of the range.
(Data doesn't follow normal curve-- skewed) |
|
|
Variance
|
James saw 12 patients in 1 day. He treated 4.84 patients more than the mean.
(Assume normal distribution) |
|
|
Standard deviation
|
Square root of variance
It is a measure of the "spread" of scores between subjects |
|
|
Normal curve
|
Bell shaped frequency distribution
|
|
|
Correlation Coefficient
|
The numeric relation between the number of patients treated on day 1 and day 2
|
|
|
SEM (standard error of measurement)
|
The SD of repeated measures of strength as measured on Kincom.
Kincom = measure muscle strength of interval scale It is a measure of the "spread" of scores within a subject had the subject been tested repeatedly. |
|
|
Mode
|
Most frequent value in a data set
|
|
|
Normal distribution or Gaussian curve
|
Mean = median = mode
Symmetrical variance left to right = SD Can quantify normal deviation if know SD and mean |
|
|
Parametric statics
|
1) Interval or ratio scale
2) If population is skewed, bell curve is not symmetrical. Then cannot use parametric stats. |
|
|
Non-normal (skewed) distribution
|
-Distribution of different age groups in western society (more elderly)
-Distribution according to age in the Parkinson's disease population (again more older adults than younger adults) -Distribution of age and gender in athletes with ACL injury |
|
|
Normal distribution or Gaussian curve
|
Mean = median = mode
Symmetrical variance left to right = SD Can quantify normal deviation if know SD and mean |
|
|
Parametric statics
|
1) Interval or ratio scale
2) If population is skewed, bell curve is not symmetrical. Then cannot use parametric stats. |
|
|
Non-normal (skewed) distribution
|
-Distribution of different age groups in western society (more elderly)
-Distribution according to age in the Parkinson's disease population (again more older adults than younger adults) -Distribution of age and gender in athletes with ACL injury |
|
|
Median
|
the point of a data set that divides the sample in half
|
|
|
Guidelines for normal distribution
|
68% of data will fall within 1 SD of mean
***95% of data will fall within 2 SD of the mean 97% of the data will fall within |
|
|
Construct validity
|
validity of the abstract construct that underlies measures
Ex. Strength measure: MMT (working against resistance). Does this capture construct of strength? No. 1-RM or max reps until fatigue. Does MMT correlate with leg strength for stairs? |
|
|
content validity
|
representation of concept of interest
Ex. writing a test on only one subject that was covered throughout the term |
|
|
criterion validity
|
extent to which one measure is systematically relation to another measure
Compares two measurement techniques to know if what it measure is valid. Ex. VAS and verbal sclae. Comparing to know if pain measurement scale is same |
|
|
predictive validity
|
extent to which a measure predicts outcome
|
|
|
Measuring errors
|
Observer
-Inter-observer reliability -Intra-observer reliability Patient/Client -Test/retest reliability with relevant time Instrumentation -Validity and Reliability: comparision to gold standard Machines |
|
|
Inter-rater reliability
|
Protocol to apply object the same way on many people
(by same observer and many trials over time) |
|
|
Inter-rater reliability
|
Test on same person over and over
(by same observer and many trials over time) |
|
|
Pre-experiment
|
"The PT/AT measures ROM before and after 6 knee knee mobilization session in a young athlete"
No control group |
|
|
Quasi experiement
|
"Effectiveness of manual therapy in comparison to home exercises in patients with low back pain"
Control group, no RCT |
|
|
Cohort study
|
"The recovery of knee stability during walking after ACL surgery using the allograph"
Follow over time. Group study. Impact of factor on recover. No control |
|
|
Cross-over design
|
"The order of a functional exercise condition and a refernce condition is randomized in one grop of patients PF pain"
repeadly done over and over the subjects get both treatments in sequence. Contrast this with a parallel groups design where some subjects get the first treatment and different subjects get the second treatment. The crossover design represents a special situation where there is not a separate comparison group. In effect, each subject serves as his/her own control |
|
|
Multiple factor design
|
"A design focused on the interaction effects of 3 different intensities of function training and gender after ACL surgery"
|
|
|
Post-test design
|
"The comparison of the effects of mobilzation techniques and fucntion exercises starting immediately after ACL surgery on walking speed"
|
|
|
True experiement
|
Control group and RCT
"101 persons after a stroke were randomly distributed across 3 treatment intentsity conditions in pre-test/post-test design" |
|
|
Single subject design
|
"The comparision of teh effects of a strength training program and a control condition within one subject one year after ACL surgery"
subject serves as his/her own control, rather than using another individual/group Often there will be large numbers of subjects in a research study using single-subject design, however--because the subject serves as their own control, this is still a single-subject design |
|
|
Statistical validity
|
Could the change observed be due to chance?
Is the outcome between 2 groups due to chance? Or true difference? |
|
|
Interval validity
|
What would have happened without the treatment?
Is this difference due to the treatment? |
|
|
External validity
|
May the results be generalized?
Generalized to practice, other patienst, other populations and enviornments? |
|
|
Construct validity
|
In how far are the results theoretically relevant?
Can we explain them in a theoretical way? |
|
|
Dependent variables
|
Outcome measures
Ex. Womac, 6-min walk |
|
|
Independent variables
|
-Intervention
Ex. 2 levels: clinic program, HEP -Time (how many repeated observations?) Ex. 4 repeated: 0, 4, 8 weeks and 1 year |
|
|
Independent factors: baseline variables
|
Gender, OA at the knee, age, weight, duration of symptoms, meds, self-rating of physical activity, days/week of activity, serveity of radiographic findings, etc
|
|
|
Attrition
|
Subjects drop out because the intervention made them worse
|
|
|
Mortality
|
Subjects drop out because they cannot complete the intervention
|
|
|
Drop-outs
|
Subects drop out from one group than another and this creates bias
|
|
|
Instrumentation effect
|
The measurement is biased. Device causes flaws
|
|
|
Unequal groups at baseline
|
One group has characteristics that predispose them to benefit (or not) from treatment; randomization of subjects has failed to balance subjects
|
|
|
Regression to the mean
|
Subjects with extreme score will tend to change toward the average, even without treatment
|
|
|
Testing effect
|
The testing of subjects particulary if it occurs often can effect the outcomes.
Learning curve |
|
|
Randomization of subjects in treatment
|
Reduces bias; sample characteristics should equally represented in all treatment groups
|
|
|
Blinding (randomization concealed, maskng)
|
Reduces bias in outcome assessment
|
|
|
Non-specific treatment items are qual for groups
|
Ex. same amount of attention
Groups should recieve the same treatment outside of experiemntal treatments |
|
|
Characteristics of Internal Validity
|
-Selection bias (patient characteristics, maturation, statistical regression to the mean, assignment, mortality, drop-out)
-Contaimination (circumstances, history) -Confounding (non-specifi parts of treatment, diffusion or imitation of treatment; compensatory equalization of treaments, compensatory rivalry of resentful demolarization) -Assessments (instrumentation) (Test retest reliability, intra/inter rater reliability, blinding observer, construct validitiy) -Carry over effect (cross over designs, effects previous to treatment) -Reactive effects (repeated measurements; testing) |
|
|
Maturation
|
What patient does over time
|
|
|
compensatory rivalry of resentful demolarization
|
Pt resistance towards treatment
|
|
|
Heirarchy of research design
|
1) True experiment- RCT (gold standard)
2) Quasi experiment (no randomization, but has control group) 3) Pre-experiment (no control) |
|
|
External validity
|
-Different assessment instruments (impairments, functional limiations, disabilities)-- ecological validity
-Different patient (sub) populations -Different circumstances (setting/time) |
|
|
Construct validity
|
-Construct underrepresenation (not clear what you are testing)
-Experimenter expectancies (if not well blinded) -Interactions between different treatments (group designs) -Interation between testing and treatment (if test w/ treatment: use endurance exercise to test endurance) |
|
|
Probability sampling
|
-Simple random sampling (preferred)
-Systematic sampling (take specific number off 1st) -Stratified sampling (select number of patients from each institution or severity of disease state) -Cluster sampling |
|
|
Non-probability sampling
|
-Sample of convenience (may not represent total population)
-Snowball sampling (word by mouth-- highly motivated subjects) -Purposive sample (choose a case- loose generalization) |
