psychometrics quiz

evidence based practice

open and thoughtful clinical decision making about PT management of a patient that integrates the best available evidence, clinical judgement, and patient’s preferences/values, and larger social context to optimize patient outcomes and quality of life

why is ebp important in healthcare?

• minimizes practice variation

• decreases healthcare costs

• decreases medical errors

• optimize research trends

evidence does not replace

clinical experience, expertise

what needs to occur in the process of ebp?

• challenge longstanding assumptions

• develop patient-specific questions

• perform literature search

• appraise evidence

• consider evidence in context of patient’s goals

what are barriers to ebp?

• lack of practice in implementing thought

• access to literature

• lack of confidence

evidence

any empirical observation about relation between events

• articles, clinical guidelines, records, personal experience

content criteria of evidence

• specific to question asked

• subjects with similar characteristics

• peer reviewed

• recent

levels of evidence

1-5 (1 is highest)

strongest type of evidence

1. systematic reviews

2. single RCT

3. systematic review with conflicting results

4. single quasi

weakest level of evidence

case reports, expert opinion, personal experience, physiologic studies

purpose of diagnostic test

• examine status of an anatomical structure

• predictive of response to particular interventions

a good diagnostic paper

• reference standard is used

• purpose of reference standard is explicit

• test procedure is reproducible

• population specificity

validity

throwing darts and hitting bullseyes

• ensures that test outcomes are aligned with gold standard

reliability

throwing darts in a consistently bad position

• correlation coefficient (-1-0-+1)

• intra-rater/inter-rater reliability

sensitivity

true positive rate- proportion of true positive subjects among all subjects who have the condition of interest

• prevalence levels does not affect sensitivity

TP/(TP+FN)

specificity

true negative rate: proportion of true-negative subjects among all subjects without the condition

• relatively stable across prevalence rates

TN/(TN+FP)

intra-rater reliability

one person would do the test two times

inter-rater reliability

2 different people would do the test

positive predictive value

proportion of all true-positive subjects among all subjects with a positive test

• PPV will be higher in a populations with low prevalence of disorder

  • will be lower in populations with high prevalence of disorder

TP/(TP+FP)

negative predictive value

proportion of true-negative subjects among all subjects with a negative test

• NPV will be lower in populations with low prevalence

• will be higher in populations with high prevalence

• not as valuable as sensitivity/specificity

TN/(TN+TP)

likelihood ratios

used to quantify shifts in probability that a disease/disorder is present once the result of a test is known

• if there is higher ratio for test and pt has positive test, we’d shift odds to higher

• if there is a lower ratio for test and patient is negative, we’d shift odds to lower

advantages of LR

• ratios can be calculated for all levels of a disorder

• not dependent on prevalence of condition within a population

  • can be applied to individual patients

floor effect

the scale no longer registers a further decline in performance

ceiling effect

the scale no longer registers a further performance improvement

Standard error of mean

a measurement of the variance about the mean score

minimal detectable change

amount of change just exceeding the SEM

• less than 5 degree distance in goni- error in measurement, not pt progress

minimal clinically important difference

the smallest treatment effect resulting in a change in patient management

• how much ROM would pt have to gain to change tx?

number needed to treat

number of pts needed to treat to get one good outcome or prevent one bad outcome

randomized clinical trial

• randomization

• 1+ group

• parametric variation

• measures at person level

  • application of scientific method

systematic review

appraising quality of evidence, making conclusions based on quality of evidence

quasi-experimental

lacking one of parameters of randomized clinical trial

single subject design

common in PT, when there is a pt with a rare condition

research based hypothesis prior, then run the intervention and see how it comes out

observational study

no parametric change in protocol, lower level of evidence

physiologic studies

far-removed, not high level of evidence

case reports

after rare case, write-up. more observational than single-subject

narrative reviews

picking and choosing articles to write about