EBP Midterm update

• Evidence-Based Practice (EBP) combines scientific research, expertise, and patient values to inform patient care and optimize therapy benefits
• Key vocabulary in EBP involves understanding scientific research, clinical expertise, patient values, and shared informed decisions

The EBP Process

• Identifying a clinical question
• Searching for evidence
• Appraising the evidence
• Integrating the evidence
• Evaluating outcomes.

Barriers to EBP

• Limited time

• Lack of generalizability

• Lack of research skills

• Understanding of statistics

• Search and appraisal skills

• Information resources

• Inconsistent culture of EBP

• EBP is important for enhanced confidence and assisting patients in choosing care options and moves the field toward evidence-based, systematic assessments

• Best evidence may not suit every patient due to individual circumstances and when evidence lacks, consider the scientific rationale, advantages, and disadvantages

• Strive to make optimal decisions with what you know, and appraise expertise like research evidence

• Patient values are key in decisions

Evidence-Based Practice Model

• Step 1: Identify need for information and form searchable clinical question

• Step 2: Search for the best research evidence

• Step 3: Appraise research evidence for validity and applicability

• Step 4: Integrate research evidence with expertise and patient values

• Step 5: Evaluate effectiveness and improve for the future

• Optimal patient outcomes depend on scientific research, clinical expertise, and patient values and circumstances

• These three form the foundation that therapists and patients use to determine optimal care strategies and are best for readable clinical questions

Evidence Pyramid

• Systematic reviews (SRs) are at the top, combining studies
• Individual studies like randomized clinical trials (RCTs) and cohort studies are below SRs
• Evidence-based summaries, like clinical practice guidelines (CPGs), are above SRs and combine research, expertise, and patient perspectives

Key Concepts

• "Searchable Clinical Question," is a foreground question about a patient to efficiently find research evidence and its elements are patient and patient management characteristics, and expected outcomes
• Background Questions - Seek general condition information, typically answered with texts
• Foreground Questions - Explore specific condition outcomes, usually with research or guidance, this is a study

Boolean Search Terms

• "AND" Narrows search to articles that contain all specified terms

• "OR" Broadens search to articles containing at least one specified term

• MeSH (Medical Subject Headings) indexes articles in MEDLINE, use keywords when better

• Database - Collection of articles and guidelines

• MEDLINE - Biomedical research database via PubMed

• Clinical Queries - Tool on PubMed to find relevant articles

• My NCBI - PubMed tool to save searches and customize filters.

• Cited By - A Google Scholar tool tracking publications citing a study.

PICO

• P (Patient/Population)
• I (Intervention)
• C (Comparison)
• O (Outcome)

Search Engines

• Google Scholar is known for helpful search tools
• PubMed provides access to MEDLINE and narrows searches using MeSH terms
• TRIP Database helps find relevant research, assesses bias, and identify various research types
• PEDro is a search engine used for physical therapy literature

Types of Research

• Systematic Review synthesizes information from primary studies
• Randomized Controlled Trial (RCT) assigns participants to intervention groups randomly
• Clinical Practice Guidelines (CPGs) are based on evidence and created by experts
• Descriptive Research seeks to describe a situation
• Causal Research determines intervention-caused change

Additional Concepts

• Efficiency- strategies to quickly locate information
• Full Text - the Complete version of a research article

Filters

• Tools to narrow search results
• Synonyms can be used to broaden a search
• Activity limitations - problems such as walking and climbing
• Participation restrictions - problems with community involvement.

Searching Steps

• Identify need for information, list keywords with synonyms and reorder terms with importance
• Refine terms with the MeSH database and enter terms in individually
• Combine terms with OR for similarity, with AND for narrowing, and use filters

Applicability

• Evaluating if its findings can benefit similar patients or clinical questions and that "close enough" is useful with the understanding that total accuracy is not always possible

Quality

• Relates to study rigor and validation controls

Study Design and Bias

• Randomization reduces bias, making groups similar
• Stratification divides subjects that might affect outcomes, and then randomizes within subgroups
• Blind treatment assignment prevents prejudice
• Placebo Effect- When belief in a treatment affects outcome
• Intention-to-Treat Analysis- When a study analyzes participants within originally assigned groups
• Participant attrition and the preservation of original groups also affect the integrity of studies
• Control and comparison groups also affect the integrity of studies

Study Factors

• Interventions- Need to be clinically realistic
• Outcome Measures- Tools used must be relevant to both goals of study and of subjects
• Consider surrogate measures, clinically meaningful vsriables, the length time of a study as well as its realism

Sampling

• Includes all patients referred, and selective focuses on participants that fit the study
• Power- A study has adequate sample size must have the ability to detect if the groups had different results
• Measuring constructs in studies and measure using objective data to have better study power
• Bias can affect results, but not only observer rated measures but any data

Week 3 Terms

• It relies on things like biases and conflicts of interest

Types of Bias

• Selection in assigning groups; or unequal care
• Performance when outcomes knowledge affects assessments
• Attrition with subject loss is not accounted for
• Report unaccurate or untruthful results due to conflict
• Recall from differences reporting prior memory events
• sampling causes a too small representative sample
• Also consider placebo effect, blinding, intention-to-treat, moving participants and comparative trials
• They key is using good measures, understanding data and being responsible

Data Types

• Nominal Data- Represents categories with no order such as sex
• Ordinal Data- Category w/ ranking like likert scales
• Discrete Data- Specific and separate values with discrete outcomes such as illness
• Continuous Data- Value within a number of given ranges such as ranges of motion
• Descriptive Statistics- Summarizing data to measure values betweeen groups
• Mean- Average used for normally dispersed data
• Median- Midpoint used for skewed data
• Mode- The most common data
• Range- The difference between results
• Variance- The average variance of the set of scores

Statistics

• Inferential tries to determine what occurs due to chance for testing differences from a sampling
• Null Hypothesis is generally what we consider to be "true;" or that there is no difference during an experiment
• Also be conscious to avoid statistical errors

Clincial Analysis

• Important effect is Effect Size which involves calculating mean differences in comparisons against multiple groups
• It is also important to calculate the number it takes for effects of benefits for outcomes and understanding Clinically Important Difference

Blinding

• Involves hiding patients, groups, and testing
• Try to use intention to treat to test groups even after dropout
• Understanding data, measuring statistically with 0.05, not making value generalizations is all important
• Also be sure that the correct measure is applied in the methodology, so not to cause bad results.

Week 4 Terms

• Type of data, data collection and descriptive statistics
• They all help you look with certainties, baselines, and what could be due to errors/outside influence

Definitions

• Nominal Data- is Categorial and number coding
• Continuous Data- is measured over a set line
• Ordinal data- ranks data given to a line
• Statistics are calculated based on central tendency or range from one set of point to another

Basic Statistical Concepts

• Reliability and the value, accuracy and consistency of data
• This occurs because there is some expectation of the outcome that occurs as opposed to just chance
• Type -2 error is missing because the result of there is no effect even though it exists which is bad and common
• Power analysis is an evaluation of if the results can be seen statistically
• They are useful to look at because of how large the patient population is
• Consecutive and Random sampling all effect reliability

Statistical Concepts

• Descriptive describe but do not compare data
• Mean is average in distribution over interquartile ranges
• A confidence interval is 95% assured to contain the true population mean
• Inference makes inferences from studies p-value
• Null hypoth. is the assumptions of no differences alternative hypoth. asserts if changes in group
• Statistical significance with values as often 0.05 will not mean the study is clean only due to small chance Type 1 errors- are false positives Type 2 errors- are false negatives

Analysis

• Population and Sample study need baseline
• Randomization and power
• Small and pilot scale studies will have more errors so must be watched carefully

Interpreted

• The change and meaning, if possible, and generalities must be understood to judge the outcome, as a statistically clear number sometimes means not much
• It relies on cronbachs values
• Small bias and the effects, while useful, need to be applied critically

Summarized data

• The Mean is normally disributed averages
• Medians are in the center of a given distrubution range
• Interquartile and deviation ranges show how the range is dispersed across the data
• Confidence interval represents how confident the study is that the sample represents and contains the true mean of the larger population

Week Five

• Important statistical concepts

Statistical measures and data

• Alpha and Standard Levels- are usually represented by something not close like in non normal distributions.
• Must be careful as researchers report these as what was done, inclusion etc and selection bias that make things misleading

Week Six

• Outcome measures are characteristics or qualities that are measured to assess status, tracking and breaking down movements classifications

• Also important are the ability for it to measure consistently, and meaningfully

• Validity and construct also needs to be taken in mind

• There should be an integration of the above for results and study appraisal in terms of use and ability to implement

• They key to consider, given how limited that resources and time are, and how complex the data can get sometimes

• There can and can't be perfect validity, consistency or integration They can also take a lot of time and energy for use given the complex data