Evidence Based Practice

what is evidence based practice?

the integration of best research evidence, clinical expertise, and patient values and circumstances used to make clinical decisions

what can influence EBP?

evidence can be influenced by the funding of research while patient preferences can be influenced by the culture, affordability, transportation, etc.

why are the benefits of EBP to the profession and patient care?

provides high quality care

decreasing unwarranted variation in practice

helps to deliver the best health outcome

enhance the value of our profession in order to increase health care quality, decrease medical errors, and decrease the consequences of unproven practices

what is the 5 step process for EBP?

1. Ask: specific questions

2. Find: evidence available

3. Appraise: identify strengths and weaknesses of article and be able to apply to the patient

4. Incorporate: how to use the evidence on patients

5. Self-evaluation: does it lead to another question

what are the barriers to EBP in clinical practice?

1. lack of time

2. decreased access to current research

3. lack of research skills/knowledge of research

4. poor quality evidence

5. confilicting evidence

6. barriers to application of evidence

7. remaining flexible and changing habits/attitude

8. lack of learning climate

what is evidence based practice NOT?

1. NOT a market-based practice: its not trying to create a lot of income or products

2. NOT a "cookbook" medicine: its not just taking the evidence and executing it. it involves taking your clinical expertise and patient preferences to adapt and figure out how to incorporate the evidence

what are the key points to consider when incorporating the evidence into clinical practice?

-is your patient similar to the participants in the study

-are the outcomes of the study consistent with your patients goals

-does the treatment align with your patients values and preferences

-is the treatment feasible in your setting?

-how does the treatmient fit into a comprehensive rehab program?

what are the two general approaches to how we ask a question in evidence based therapy?

1. Background: this is looking for information on the internet or in a textook about a certain topic, the diagnosis or medical information

2. foreground: looking for something direct or comparative or asking a more targeted question such as "what is the best intervention for my patient

What is PICO?

Patient or Problem: How would you describe the patient or their primary problem

Intervention: what main intervention are you considering

Comparison: what is the main alternatie to compare with the intervention; another treatment, control, placebo, etc

Outcomes: what are you trying to accomplish, improve or affect?

t/f questions may need to be redesigned multiple times to get an ideal amount of evidence

true

what is the evidence heirarchy?

From less to more bias

depending on the type of question, what is the best evidence?

-diagnosis: prospective, blind comaprison to a gold standard

-therapy/treatment: RCT>cohort>case control>case series

-prognosis: cohort study>case control>case series

-harm/etiology: RCT>cohort>case control>case series

-prevention: RCT>cohort study>case control>case series

-clinical exam: prospective, blind comaprison to a gold standard

what is the PICO 2.0 extension?

Patient or Problem: How would you describe the patient or their primary problem

Intervention: what main intervention are you considering

Comparison: what is the main alternatie to compare with the intervention; another treatment, control, placebo, etc

Outcomes: what are you trying to accomplish, improve or affect?

Type of question: is this a diagnosis, therapy, prognosis, etiology/harm, or prevention question

Type of study design: what study design would best answer this question?

what is research and why is it important?

• research is a process of gathering proof of the truth

• it’s important because clinicians rely on research as evidence for informing and supporting our clinical decisions. they pick a study design to best answer their specific question

discuss Clinical Practice Guidlines

• systematically developed statements to assist practitioners and patient decisions about appropriate health care for specific circumstances

• their goal is to provide clinicians with direct recommendations that they can use to make decisions based on available evidence grades

dicuss systematic reviews and meta-analyses.

systematic reviews: answer questions by systematically reviweing and describing ALL relevant available evidence. it attempts to synthesize available evidence to state what is known at this time

meta-analysis: synthesize findings from multipe studies to generate summary statistics. pools study samples together to create a larger sample and stronger evidence

discuss randomized control trials.

participants are randomly assigned to groups to compare interventions

pros: randomization helps isolate the effect of the intervention and reduce bias

cons: very specific focus and volunteer biasesd

discuss observational studies.

cohort studies: retrospective or prospective study of exposures and outcomes

case-control study: observation study design where cases have condition of interest.

these study designs do not show cause and effect, rather they attempt to understand correlations

discuss case studies or case series

• typically have a smaller participant number from studying a rare condition

• can provide detailed explanations of clinical intervention descriptions

• provide evidence when there are barriers to other study design

• con: limited generalizability of results

discuss narrative reviews and expert opinions

they provide clinical expertise, vision casting or calls to action when there is/are:

1. little available research or evidence on a topic

2. new or emerging events

3. a need to highlight/bring attention to

4. raise awareness for a new philosophy of thought

research is a systematic process for answering a question. what is this process?

1. develop a question

2. create a hypothesis

3. select a study design and collect data

4. test the hypothesis

5. interpret the findings

what is the null hypothesis vs alternative hypothesis?

null: there is no difference between two treatments

alternative: there is a significant difference between two treatments

what is the goal of research regarding the null hypothesis?

to reject the null— we first have to assume there is no difference until we can prove there is

what are the two ways to test a hypothesis?

1. one sided (or 1 tailed) test: these are directional. for example intervention A> intervention B. however the limitation here is that you can only test in one direction

2. two-sided (or 2 tailed) test: does not specify direction so intervention A ≠ intervention B. the limitation here is that further examination is then needed to determine “effect” of treatment

researchers want to provide evidence taht something is true (or works) for a certain population. this population is referred to as…

the target population

what is a sample?

a representative subset of the population

how do researchers use samples?

to make inferences about the target. the closer the sample matches the target population the more generalizable the results

what are the limitations of samples?

samples are always approximations. therefore researchers need to be mindful of…

• the methods they use to select a sample

• statistical plan to maintain a samples ability to make inferences about the target population

what is a variable?

characteristics of individuals, objects, or the environment

what are the types of variables?

independent: predector or explanatory; X variable; intervention tested

dependent: response; y variable; outcome of interested

measurements can be measured with increasing levels of measurement complexity to give different types of data. what are these ways?

from least complex to most complex:

1. nominal: categories

2. ordinal: order

3. interval: distance between has meaning

4. ratio: absolute zero

discuss nominal data

this is categorical data and categories have no mathemetical properties, rank or value

can be described as frequences and summarized with a mode statistic

ex: sex, eye color, marital status, medical diagnosis, race, geographic location

describe ordinal data

this is categorical but they have a rank, order, or value BUT there is no set distance between them.

can be described as frequences and summarized with a mode statistic

ex: manual muscle tests, level of assistance (min, mod, max), survey responses

discuss interval data

this is continuous data that has numerical values on a scale that have equal distances between them. but there is no zero point (values can be negative or positive.

described as mean, median, and mode

ex: temperature, calender year

describe ratio data

this is continuous data that has numerical values on a scale that have equal distances between them. but there there IS a zero point (values cannot be negative)

described as mean, median, and mode

ex: height, walking speed, walking distance, time to complete task

________ data is the first step to understanding it.

visualizing

what is central tendency?

used to describe measures that tell us about the center of a distribution of data

what are the central tendencies"?

mean, median, mode

what is special about a normal distribution or parametric data?

the mean, median, and mode are all equal

it is representative of the variation in a population for a given characteristics

is is an assumption for many statistical tests

what is special about non-normal or non-parametric data?

the mean follows the tail while the mode is at the most frequent variable. the median is between these and best used to see the central portion of what most people demonstrate in a non-normal curve

non-normal distribution ≠ bad data— many variables in PT are not normally distributed and thats ok

what are the two types of skewed distributions?

left or negatively skewed: asymmetrical with mean shifted to the left

right or positively skewed: asymmetrical with the mean shifted to the right

_________ and _________ combine to better describe how data is distributed

measures of variability and measures of central tendency

what are the measures of variability?

1. range

2. standard deviation

3. coefficient of variation

4. z scores

5. standard error of the mean

6. confidence intervals

discuss range

the maximum value-the minumum value

its the simplist way to understand the spread of the data

discuss standard deviation

it describes the spread of the data around the mean value

small SD: very little variance in the sample so all values are close to the mean

large SD: a lot of variance in the sample so all values are spread out over a wide range

dicsuss coefficient of variation

the amount of variablility expressed as a percentage of the mean

advantages

• unit-less

• compare variablility between different measures of the same thing

  • compare variability between 2 different samples on the same measure

discuss z-score

the number of standard deviations a data point is from the mean of the sample or the population

informative for a singel data point or a group

allows comparison of data points across datasets

95% of z scores between the values of -2 and +2

discuss standard error of the mean

a measure of precision of the sample mean

it tells us how well your sample represents the population

smaller values are better bc it indicates the sample mean is more precise

the larger the sample, the smaller the standard error

discuss confidence intervals

a range of values that contain the true value with a given probability (95%)

a smaller interval is better bc we can be more confident that the mean shown is close to the target population mean for that measure

what is the purpose of measurement?

to provide a way to quantify something that is abstract in a meaningful way

what makes the perfect measurement tool and why do we need them?

1. standardized

2. sample free

3. equal distance between items or scores, providing a known distance

this allows for efficient and accurate understanding of how much exists

also provides knowledge for how much change is real, meaningful, and attainable

what are the measurement benefits?

a well selected measure will let you…

1. predict

2. diagnose

3. discriminate

4. assess change

measures increase the efficienty of your examination and improve communications

measured values should drive clinical decision making

what is a measured value?

true value + random error + systematic error

what are sources of error when measuring that could make a measured value change?

1. patient

2. examiner

3. environment

4. instrument

what is primary variance?

consistent changes in the measured value related to the primary effect

changes due to things like amount, duration, and quality of the independent variable

what is secondary variance?

consistent changes in the measured value due to factors other than the primary effect

changes due to things like environment, aspets of the participants, aspects of the raters

what is error variance?

inconsistent changes in the measured value due to things like measurement error or using incorrect statistics

what does it look like to control variance?

1. Maximize primary variance

2. control secondary variance

3. minimize error

what is measurement validity?

test/instrument measures what it is supposed t omeasure

a measure cannot be valid if it is not reliable but reliability does not imply validity

what are the types of validity?

face

content

criterion

construct

discuss face validity

“does the test instrument seem to be a good choice to measure this contstruct?”

clinical judgement, subjective, does it make sense

ex: pain scale

discuss content validity

used for outcome measures often

instrument covers all elements of the construct being measured but not does include irrelevant effects

ex: a balance questionare will not include questions about the wrist

discuss criterion validity

how a measure is related to an outcome

1. predictive: measure predicts an outcome of interest

2. concurrent: comparison with a gold standard at the same time. measure of interest and a measure with established validity are administered at the same time point and produce consistent results. Most common- orthopedic tests compared to medical imaging

discuss construct validity

degree to which a test measures the theoretical construct that it claims/intends to measure

ex: berg balance scale, patient reported outcomes

what is reliability?

repeated measures that are consistent

reproducibility, repeatability, consistency, dependability

______ tests/measures allow you to be more confident that a true change has occured between measurements

reliable

what are the types of reliability?

test-retest: measure is consistent when performed multiple times on same patient and construct has not changed. used in short time periods

intra-rater: measurements obtained by the same assessor are consistent

inter-rater: measurements obtained by 2 or more assessors are consistent with each other

What are the two types of validity in research?

Internal and External Validity

What are the types of research questions?

1. Does intervention A improve an outcome measure mroe than intervention B?

2. Does intervention A improve an outcome measure

3. Is variable C associated with Variable D?

4. Can the values of Variable C (and D, etc) predict the value of Variable E?

5. Is population F different from population G in terms of some characteristic or outcome measure?

6. none of the above (this is rare)

What is internal validity?

the degree to which the results of the study can be attributed to the study intervenention and not extraneous factors.

“how confident are you that the results truthfully answer the research question?”

what is internal validity dependent on?

study design

care with measurements

care with intervention delivery

what is the most important property of any study?

internal validity bc problems with internal validity can limit or discredit the findings of a study

what are the threats to internal validity?

1. selection bias

2. insufficient sample size

3. participant attrition

4. history

5. maturation

6. testing

7. assessor bias

8. conflict of interest

9. instrumentation

10. regression to the mean

11. multiple comparisons

12. diffusion of imitation of treatment

13. compensatory rivalry/resentful demoralization

14. multiple comparisons

what is selection bias?

the process of selecting subjects leads to a sample that is not representative of the target population

theprocess of placing subjects into groups leads to groups with different baseline characteristics

what is the potential problem with selection bias?

the samples characteristics could influence or mask the effect of the dependent variable

what are the potential solutions to selection bias?

1. randomization of subjects

2. matching of subjects

3. account for statistical analysis

what is insufficient sample size bias?

the sample size is determined by a calculation to detect an estimated effect while balancing the risk of error

what are the potential problems of an insufficient sample size bias?

increasing the sample size drastically increases the possilibyt of finding a significant difference

if the sample size is too low, a study is underpowered and researchers are at a higher risk of making a false negative (type II) finding

what is the solution to insufficient sample size bias?

sample size justification

what is participant attrition bias?

participants drop out of (or do not complete) a study due to injury, illness, death, family, transportation

what are the potetial problems with participant attrition bias?

can create bias, especially if non-random as the gruops may no longer by similar

influence study power

may be reflective or problems with the study such as intervention, timing, or burden

should be reported

what are the potential solutions of participant attrition bias?

1. enroll additional subjects

2. account for in statistical analysis

3. document drop-outs and reasons

what is history bias?

events that occur outside the study but influence the results of the study such as concurrent events with the study that are out of the control of the investigator and that woul likely increase the length of the study

what is the potential problem with history bias?

the outside event could influence of mask the effect of the dependent variable

what are the potential solutions to history bias?

1. random assignement with a control group

2. plan/schedule as able to avoid any known events

3. account for in statistical analysi

what is maturation bias?

changes over time that are internal to participants. not related to the study but may impact results

what is the potential problem with maturation bias?

the course of recovery or decline could influence or mask the effects of the dependent variable

what are the potential solutions to maturation?

1. random assignement with the control group

2. multiple baseline testing

what is testing bias?

the subjects become familiar with the test or there is inconsistent administration of testing procedures by the study team

what are the potential problems with testing bias?

something about the test could influence or mask the effect of the dependent variable

bas measurement gives unreliable results

what are the potential solutions to testing bias?

provide practice sessions with test

standardize protocols for test administration

what is assessor bias?

the person testing the participants has a particular bias or opinion about how the outcome of the study should look

what is the potential problems of assessor bias?

soemthing about the bias influences or masks the effect of the dependent variable

bad measurement gives unreliable results

what are the potential solutions of assessor bias?

blind the assessor to the group assignment

use a third party or technology to conduct testing

what is confluct of interest bias?

the persons involved in the study have an invested (apparent to them or not) interest in the outcome

what are the potential problems with conflict of interest bias?

something about the bias influences or masks the effect of the dependent variable

could result in ethical problems

what are the potential solutions of conflict of interest bias?

researchers declare conflict of interest

an independent team develops a plant to manage the conflict of interest

IRB approval and Data Safety Monitoring Boards can help prevent ethical concerns

what is instrumentation bias?

problems with measuring tools regarding selection, application, and calibration

what are the potential problems with instrumentation bias?

bad measurement gives unreliable results

what are the potential solutions with instrumentation bias?

select best, most appropriate measures

train study personnel on standardized testing procedures

calibrate equipment