Midterm EBP

what is evidence based practice

integration of best research evidence with clinical expertise and patient values

5 steps to practicing EBP

1. formulate an answerable question
2. track down the best evidence
3. critically appraise the evidence for validity, clinical relevance, and applicability
4. individualize, based on clinical expertise and patient concerns
5. evaluate your own performance

theory

overarching tenets that may or may not be correct, usually derived from basic science research in anatomy, physiology, biomechanics, and histology

evidence application

using clinically applied data, not theoretical arguments

evidence

actual data collected on patients

levels of evidence (highest to lowest)

1. systematic review of randomized trials
- systematic of cohort/cross
2. single randomized trial
3. systemic review of observational studies with patient-important outcomes
4. physiologic studies
5. unsystematic clinical observations
6. expert opinion

gold standard for evidence

peer review

abdication

decisions based on expert advice

induction

making decision based on experience or based on theory

deductive reasoning

decisions made based on prospective peer-review research based on clinical questions and hypotheses

5 A's

1. ask
2. acquire
3. appraise
4. apply
5. assess

background questions

- broader, disorder related
- may deal with pathology, physiology...

foreground questions

- ask about the specific patient
- easier to answer
- more straightforward
- more transparent

PICOS

- population/patient
- intervention
- comparison
- outcome/time frame
- study design

components of nagi

1. pathology
2. impairment
3. functional limitation
4. disability

WHO-ICF

1. health condition
2. impairment (body function/structure)
3. activity limitation
4. participation restriction

major perspectives to be considered with regards to disablement

1. how person defines disablement situations and their reaction to it
2. how others perceive the disability and react to it
3. characteristics of the environment

contextual factors

1. environmental
2. personal

prognosis of a patient

refers to its possible outcomes and the likelihood each one will occur

outcome of a patient

status of the patient at the time of discharge from PT

evaluation

process of diagnosing

examination

diagnostic test evidence, measurement evidence (reliability and validity), assessment for risk factors (etiology), and prognostic evidence

intervention

tremendous amount of data

bias control continuum (most bias control to least)

1. experimental designs
2. quasi-experimental designs
3. non experimental designs
4. case report/anecdote

primary sources of evidence

original research reports found in peer reviewed journals, proceedings from professional meetings, theses, and dissertations

secondary sources of evidence

summary review of works of others found in systematic and narrative review in peer reviewed journals, textbooks, practice guidelines

hierarchy of evidence

1. systematic review of RCT (intervention)
2. systematic review of cohort. cross sectional, or case control studies (non-intervention)
3. individual RCTs (interventions) or individual observational studies (other)
4. case series study (interventions)
5. case reports (any)
6. expert opinion (any)

characteristics of good clinical questions

- based on patient problems (patient focused)
- action focused
- specific to the issue you need to know about
- reasonable

measurement

numeral assigned to an object, event, or person or the class to which an object, event, or person is assigned according to rules

types of measurement

1. nominal
2. ordinal
3. interval
4. ratio

nominal

- based on names or categories
- dichotomous data

ordinal

allows for rank order

interval

- consistent degree of difference between items
- no absolute zero

ratio

possess a meaningful zero value

3 purposes for measurements

1. discriminative
2. predictive
3. evaluative

discriminative measurements

distinguish between individuals on the variable of interest when no gold standard exists

predictive measurements

distinguish between individuals on the variable of interest when a gold standard exists

evaluative measurments

used to measure the magnitude of change over time in an individual on the variable of interest

operational definition

sets of procedures that guide the process of obtaining measurements

two criteria for a sound operational definition

1. universality
2. south theoretical basis

universality

everyone understands it

sound theoretical basis

it's a good reason

reliability

consistency or repeatability of measurements, the degree to which repeated measurements will agree

sources of reliability error

- errors made by the examiners
- instrumentation flaws
- the measure may be inherently inconsistent

4 common types of reliability error

1. intratester
2. intertester
3. parallel (alternate) forms
4. population specific

intratester

measures repeated by same tester

intertester

measures repeated by different testers

parallel (alternate) forms

measures obtained using different forms of a test (interchangeability)

population specific

reliability may be population specific (normal vs diseased)

when should kappa or weighted kappa be used

for nominal and ordinal level data

when should ICC with the SEM be used

for interval and ratio scaled data

ICC

intraclass correlation coefficient

most common ICC

ICC2, 1

SEM

- standard error of measurement
- estimates the error associated with a single measure in the units of interest

MDC

minimal detectable change

MCID

- minimally clinically important difference
- change in a measure necessary to indicate that the change is meaningful to the patient
- based on another patient-based measure
- usually approximately equal to the 1 or 2 SEM

validity

does the test measure what it's supposed to measure AND what can be inferred from the measurement

types of validity

1. face validity
2. construct validity
3. content validity
4. criterion related validity

face validity

does the measurement appear, on the face of it, to assess what is intended

construct validity

conceptual/theoretical basis for using a measurement to make an inferred interpretation, evidence for construct validity is through logical arguments based on theoretical and research evidence

criterion related validity (CRV)

requires direct comparison of measurement of interest with a standard (criterion) measurement

concurrent (CRV)

inferred interpretation is justified by comparing a measurement with supporting evidence that was obtained at approximately the same time as the measure being validated

predictive (CRV)

inferred interpretation is justified by comparing a measurement with supporting evidence that is obtained at a later point in time

etiology

deals with risk

what types of questions are etiological questions

background questions, not directly applicable to individual patients decision making

issue with etiological questions

- the cause
- it's necessary to study people before the disease starts

STROBE

- strengthening the reporting of observational studies in epidemiology
- an instrument to help determine the quality of longitudinal study
- large, representative cohort

prognostic research

if concerned with prediction, the investigator does not need to differentiate between causal and non-causal associations because any association meets the goal of predicting outcomes

what is necessary to infer causation

1. consistency
2. strength
3. specificity
4. dose response relationship
5. temporal relationship
6. biological plausibility

motivation for etiological studies

the primary motivation for studying the etiology of a health condition is to suggest mechanisms through which health interventions might act and, in that way, to inform development of new interventions

bottom line of etiological evidence

1. study should be prospective
2. measures of risk factors and outcomes are replicable and sound
3. minimal loss to follow up

prognostic evidence

knowing how to estimate our patient's likely clinical course and outcome over time (temporal and outcome driven)

components of prognostic decisions

1. functional status
2. pain intensity
3. potential need for assistance
4. destination
5. return to societally expected activity

prognostic studies require the following

- type of outcome be predicted
- probability of the outcome actually happening
- time frame of interest

study design for prognostic research

- prospective longitudinal designs are optimal
- inception cohorts are optimal
- follow up in prospective fashion (remember STROBE)
- some are cross sectional and attempt to look back but those are weaker

bias domains in prognostic studies

- study participation
- study attrition
- prognostic factor measurement
- outcome measurement
- study confounding
- statistical analysis and reporting

bottom line of prognostic research

1. study attrition (minimal loss to follow up)
2. prognostic factor measurement (reliable)
3. outcome measurement (replicable)
4. study confounding (all important confounders addressed)

most common estimate from a prognostic study

relative risk (RR)

relative risk

- the gold standard
- cohort study design
- exposure status known
- all subjects followed into future
- outcome measured later
- compare probabilities of those with and without outcome

RR>1

increased probability of outcome among those with prognostic factor

RR

decreased probability of outcome among exposed with prognostic factor

RR=1

probability outcome is similar in all subjects
- not significant if 1 is included in the range

RR equation

[a/(a+b)]/[c/(c+d)]

point estimate of RR

describes magnitude of effect

interval estimate of RR/OR

describes precision of point estimate

odds ratio

cross sectional prognostic study which starts with outcome and follows backward to the prognostic factor

RR vs OR

- as long as prevalence of outcome of interest is low (

strengths of cross sectional studies

- relatively quick and inexpensive to conduct
- no ethical difficulties
- data on all variables are only collected at one time point
- easy for generating hypotheses
- many findings can be used to create an in-depth research study

weaknesses of cross sectional studies

- unable to measure the incidence
- difficult to make a casual inference
- associations identified might be difficult to interpret
- unable to investigate the temporal relation between outcomes and risk factors
- not good for studying diseases
- susceptible to biases such as non responsive bias and recall bias

simple random sampling

every member of the population has the same probability of being randomly selected into the sample

systematic sampling

one selects every nth subject in the population to be the sample

stratified sampling

the population is divided into non-overlapping groups, or strata; a random sample of population members is then collected from within each stratum

clustered sampling

the researcher divides the population into separate groups (clusters), then a simple random sample of clusters is selected from the population

convenience sampling

participants are selected based on availability and willingness to start

quota sampling

a tailored sample that is in the proportion to some characterizing or trait of a population

snowball sampling

existing study subjects recruit future subjects from among their acquaintances

sampling bias

some individuals within a target population are more likely to be selected from inclusion than others

allocation bias

there is a systematic difference between participants in exposed and unexposed groups

loss-to-follow-up bias

some individuals lost to follow up differ from those who were not lost to follow up with respect to the exposure and outcome

nonresponse bias

there is a systematic difference between responders and nonresponders