
Sensitivity
Define
High sensitivity means what?
Good for ruling in or out disease?
Mnemonic
 Pt tests (+) for a condition they have
 High = A lot of true (+) bc it's so sensitive / Few false ()
 Good for ruling OUT who has disease
 SnNout  Test w/high Sensitivity, ( N) result rules OUT diagnosis

Specificity
Define
High Specificity means what?
Good for ruling in or out disease?
Mnemonic
 Pt tests () for a condition they DO NOT have
 High = pts don’t have disease  Many true ()/ few false (+)
 Good for ruling IN who has disease
 SpPin = Test w/high Specificity, (+P) result rules IN diagnosis

What is included in written informed consent?  4
 Info about general nature of what is to take place
 Any risks to individual & what will be done to minimize the risk
 Possible benefits
 Ethical disclosure

Independent variable  2
 Variable being manipulated (changed or controlled) to test effects on dependent variable
 Treatment/intervention administered to subjects

Dependent variable  define; ex
 Change, outcome or difference in behavior resulting from IND variable
 Variable being tested/measured to determine outcome
 Ex: ROM

Analysis of Variance (ANOVA)  define; used when
 Statistical test
 Multiple variables  Differences bet 3 or MORE groups

Chi square test  define; used when? (2)
 Statistical test
 Used  Genetics + Population

Ttest  define; used when? w/what variables?
 Statistical test
 Difference between TWO groups
 Only used if there is 1 dependent & 1 independent variable
 2 different groups are matched & tested OR 1 group is tested twice

Interrater vs intrarater reliability
 Interrater = 2 or more different testers
 Intrarater = Same tester

Bell curve  %s fall above/below 1, 2 3 standard deviations
 +1 or 1 SD = 68%
 +2 or 2 SD = 95%
 +3 or 3 SD = 99%

Null hypothesis  define; rejected; not rejected
 Predicts a finding of no significant difference between groups or treatment studies
 Rejected  Significant difference was observed  P value Not rejected  No significant difference = P value > level of significance

Alternative hypothesis  define; when is it accepted?
 Predicts significant difference between groups or treatment studies
 Accepted when null hypothesis is rejected

Statistical significance  define; P values & what they mean
 Hypothesis is statistically significant = test results in pvalue less than significance level
 Pvalue = Probability whether or not results of experiment happened by chance
 P = 0.05 = results occur by chance 5% of time & occur from tx 95% of time
 P = 0.01 = Higher statistical significance bc only 1% probability results occurred by chance

Descriptive research  Define; purpose; ex (5)
 Describe, record, analyze characteristics of what's being studied
 It desn't answer questions about how/when/why characteristics occurred
 Purpose  Classify & understand clinical phenomenon
 Ex  Developmental, normative, qualitative + case report & case series

Experimental research  Define; purpose; ex (2)
 Comparing 2 or more conditions
 Treatment is intentionally introduced & result or outcome is observed
 Purpose  cause & effect bet dependent & independent variables
 Ex  RCT; singlesubject designs

Exploratory research  Define; purpose; ex (4)
 Examines research question & its relationship to other factors
 It does not intend to offer conclusive solutions to existing problems
 Helps in determining research design, sampling methodology & data collection method
 Ex  cohort; case control; historical research; methodological studies

Scales of measurement  Nominal  another name; define; ex
 Classification scale (category)
 Each object/person can be assigned to ONLY 1 category  MUTUALLY EXCLUSIVE
 Ex  blood type; breath sound; type of arthritis; male vs female

Scales of measurement  Ordinal  another name; define; ex
 Ranking scale
 Degree to which difference bet scores is not quantifiable
 Ex  MMT; levels of assistance; pain; jt laxity grades

Scales of measurement  Interval  another name; define; ex
 Measurement scale
 Based on predetermined = intervals w/NO true 0
 Ex  Temp C vs F (body, skin, whirlpool); development tests; functional status tests; standard deviation

Scales of measurement  Ratio  another name; define; ex
 Measurement scale  highest, most precise level of measurement
 Based on = intervals w/true 0
 Ex  ROM (deg); distance walked (m); nn conduction velocity (m/sec); time to complete activity (s)

Validity refers to what?
How well a test measures what it is purported to measure

Reliability  define
Degree to which an assessment tool produces stable & consistent results

Internal validity  define
Degree to which observed differences on dependent variable are direct result of manipulation of independent variable & NOT result of extraneous factors
Basically how well an experiment is done, especially whether it avoids confounding (more than one possible independent variable [cause] acting at same time)

Internal validity  threats to it (7)
 History  Events that occur besides tx (events in environment)
 Maturation  Physical/psychological changes in participants
 Attrition  Participants lost from study
 Instrumentation  Learning gain from pre to posttest
 Testing  experience w/pretest  become test wise
 Regression toward mean
 Selection bias
Confounding  less chance for confounding in a study> higher its internal validity

External validity  define; threats to it (3)
 Generalizability of study to a particular population
 Threats  Interaction of treatment w/subject / Place (setting) / Time

Hawthorne effect  define; When is it present?
 Untreated subject experiences change simply from participating in study
 Present when pts are being observed

Type I error  define; leads to what finding
 Null hypothesis REJECTED when it is true
 Thinks presence difference/relationship when there is none
 False (+) finding (Test indicates pt has disease but in reality pt doesn’t)

Type 2 error  define; leads to what finding
 Null hypothesis NOT REJECTED when it's false  alternate hypothesis is true
 Thinks no difference/relationship when there is
 False () finding (Test indicates pt doesn't have disease but in reality they do)

Power  3 primary factors + 1 secondary factor
 Primary  sample size; effect size; significance level (p value)
 Secondary  statistics

Effect size (ES)  define; INC ES means what?
 Measure of strength of a phenomenon
 Measuring magnitude of effect  quantifying DIFFERENCE bet 2 treatments
INC ES = Larger value always indicates a stronger effect = statistically significant

Repeated measures design  does what
controls for differences between subjects

Reliability coefficient  good; moderate; poor
 Good = Above 0.75
 Moderate = 0.500.75
 Poor = below 0.50

Incidence vs prevalence
 Incidence  # of new cases
 Prevalence  # of existing cases (new & preexisitng)

Levels of evidence  Systematic reviews  define; best for what
 Comprehensive review of medical literature (Ex: Cochane)
 BEST source of information upon which to make clinical decisions about therapy for an individual patient

Levels of evidence  Meta analysis  define
Combined results of several RCTs

Levels of evidence  Randomized control trials
STRONGEST & most effective

Levels of evidence  Cohort studies  define; effectiveness
 Pts followed over time
 LEAST EFFECTIVE

Levels of evidence  Case studies  define; effectiveness
LEAST EFFECTIVE

Levels of evidence (Sackett)  from best to worst  5 + 1
 Systematic reviews + Meta analysis
 I  Multicenter RCTs
 II  Singlecenter RCTs
 III  Cohort studies
 IV  Case control studies
 V  Case reports

PICO  define
 Used to ask a focused clinical question
 P = pt or problem
 I = intervention
 C = comparison
 O = outcome

Construct Validity  define
 Degree to which test measures what it claims to measure
 "Nonobservable behaviors or ideas"
 "Does the measure behave like theory says a measure of that construct should behave?"

Content Validity  define
 Degree to which an instrument measures an intended content area
 Determined by expert judges

Criterion Validity  define
Extent to which a measure is related to outcome

RESEARCH  Correlation  .89; .68  define
 .89  high correlation
 .68  moderate correlation

When you are confident that the experimental manipulation produced the changes you measured in the dependent variable, your study probably has good ________ validity
a.external
b.internal
c.construct
Internal

The extent to which we can generalize the results of a study to other participants is called:
a.sampling validity
b.internal validity
c.external validity
d.construct validity
External validity

The validity of a measure refers to the
a.particular type of construct specification
b.consistency of measurement
c.comprehensiveness w/which it measures construct
d.accuracy w/which it measures construct
Accuracy w/which it measures construct

If a measure is consistent over multiple occasions, it has:
a.internal validity
b.construct validity
c.testretest reliability
d.interrater reliability
Testretest reliability

Measurement reliability refers to the:
a.comprehensiveness of scores
b.consistency of scores
c.dependency of scores
d.accuracy of scores
Consistency of scores

Bell Curve  % between standard deviations:
1 standard deviation = what %
2 standard deviations = what %
3 standard deviations = what %
 1 standard deviation = 34.1% + 34.1% = 68.2%
 2 standard deviations = 2 x (34.1% + 13.6%) = 95.4%
 3 standard deviations = 2 x (34.1% + 13.6% + 2.14%) = 99.7%


Bell curve & Osteoporosis
Normal
Osteopenia
Osteoporosis
 Normal = bone density within 1 SD (+1 or 1)
 Osteopenia = bone density bet 1 & 2.5 SD (1 to 2.5)
 Osteoporosis = bone density is 2.5 SD (2.5 or lower)

What is ICC?
High ICC # means what?
 Intraclass correlation coefficient is a reliability coefficient based on ANOVA  from 01
 High ICC = very reliable

