Evidence-Based Practice (EBP)

• Evidence Based Practice (EBP) integrates best research/evidence, clinical expertise, and patient values.
• EBP Steps: Ask the question, acquire evidence, appraise the evidence, apply to patient care, and assess the process.

Importance of EBP

• Addresses high healthcare costs.
• Improves patient outcomes.
• Provides justification for treatment.
• Accounts for cultural/geographic variations.
• Improves public access to healthcare information.
• Bridges the gap between research evidence and clinical practice.
• Encourages respect and confidence as a clinician.

Obstacles to EBP

• Limited access to research/databases.
• Preconceived notions.
• Demographics may not be universally applicable.
• Outdated, irrelevant evidence.
• Biases can influence understanding.
• Limitations include needing strong literature searching skills and challenges to clinical expertise.

Types of Clinical Questions

• Background Questions: Address general knowledge about a condition, test, or treatment.
• Focus on etiology, general knowledge, or cause, history, signs and symptoms, anatomical or physiological mechanisms
• Helps to understand the physiology of a treatment or test.
• Examples: "What are common signs/symptoms of a lateral meniscus tear?"
• Uses resources like textbooks, research summaries, and credible websites.
• Foreground Questions: Specific and focused questions, about a particular patient or situation to inform a clinical decision.
• Guided by PICO(TS) (Patient/Problem, Intervention, Comparison, Outcome, Time, Study type)
• Uses resources like research articles, systematic reviews, and CPGs.

Searchable Clinical Questions (PICO(TS))

• Foreground questions are searchable.
• Patient/Population/Problem: Includes age, gender, key co-morbidities, activity level, race/ethnicity, and environmental factors.
• Intervention: Specifies the treatment of interest (e.g., manual therapy for LBP, home exercise program after stroke).
• Comparison: Details what the intervention is compared to (e.g., no therapy, usual care).
• Outcome: Specifies the measure of the effect (e.g., Pain, ROM, functional activity, quality of life, surgery).
• Time: Study design & time

Types of Hypotheses

• Hypotheses: Should be clearly stated, related to the question, consistent, answerable, reasonable, and testable.
• Null Hypothesis: There is no statistical difference between testing variables.
• Research Hypothesis: Predicts the difference in testing variables and the expected outcome.
• Results either accept or reject the hypothesis, but do not "prove" it.
• Accepting a hypothesis doesn't prove it, and rejecting one doesn't make an alternative outcome true.

Appraising Evidence

• Determine if scientific evidence is available and complete.
• Consider: Type of evidence, how it was created, and to whom and what situations it applies.

Levels of Evidence

• Level of evidence depends on the type of question being asked.
• Systematic Reviews: Gather all information and research to give a conclusion, compiling work from others and involving no new subjects.
• It is considered the strongest level for all types of questions.
• Must use systematic methods and multiple databases to gather and appraise articles and have similar studies.
• Randomized Controlled Trials (RCTs): Participants are randomized into different treatment groups to examine cause and effect.
• Always prospective and allows for comparison between groups.
• Cohort Studies: Involve one or more predetermined, non-randomly assigned groups.
• Can be prospective or retrospective, used to assess the development of disease or other outcomes.
• Cannot control for confounding factors.
• EX: following smokers vs. nonsmokers for lung cancer.
• Case Report/Case Series/Case-Control/Historical Controls: Observational studies, don't apply an intervention.
• Case Report: Detailed story of an individual patient, useful for rare conditions (N=1), hard to apply to another person.
• Case Series: Several case studies illustrating a single point (N>1 but <20), without group statistical analysis.
• Case control: Those with condition (cases) are compared to those without (cross-sectional or retrospective).
• Considered best evidence for etiology or rare conditions with long latency.
• Expert Opinions/Animal/In Vitro Studies: Not good bc they are biased
• Cannot determine whether the treatment is good for a person

Research Design Formats

• Experimental: Manipulating variables to observe affects on another while looking for relationship
• Correlational: Investigates relationship between variables to see if they are related without manipulating variables
• Descriptive: Focuses on describing characteristics of a population or phenomenon being studied, and does not manipulate variables

Applying Formats to Research Designs

• Systematic Review: Strongest level for all types of questions.
• RCT: Best study design for intervention and best way to know if intervention works.
• Cohort: Best for prognosis.
• Control Studies: Best evidence for etiology or rare/rare disease or conditions with long latency.

Time Elements in Research Design

• Duration:
• Cross-Sectional: looks at group of people as assuming they represent everyone at single time. Data collected at a single point in time, like a snapshot of the population.
• Longitudinal: same person over time, data collected at multiple timers
• Direction:
• Prospective: Best because can plan details: following group(s) over time.
• Retrospective: Split from one group (outcome) into 2 *opposite of prospective

Classifying Research

• Planned, controlled, not standard care methods require approval.
• Manipulation, observation, or other study of humans with the intent of developing new knowledge requires approval.
• This approval process must occur PRIOR to beginning

Human Rights in Research

• Researchers need to protect human rights and dignity.
• Minimize harm
• Inform subjects of the research purpose and allow refusal/withdrawal.
• Autonomy: Subjects can make own decision to participate
• Benefice: Obligation to tend to well being of the patient
• Justice: subject selection or group assignment.

Institutional Review Board (IRB)

• Any research must go through this designated group that looks at risks vs. benefits.
• Full Review: Studies involving more than minimal risk, sensitive topics, or minors.
• Expedited Review: Studies involving minimal risk, no sensitive topics, and no minors.
• Exempt Review: Studies posing little risk, not involving sensitive topics or minors.
• Research includes questionnaires, archival data, or education setting with normal curriculum.

Consent Forms Should Include

• Information regarding risks, benefits, time, etc.
• The purpose, procedures, and confidentiality of the research.

Consent Elements

• This must be voluntary and special considerations for vulnerable populations.
• Subjects can withdraw at any time.
• Consents come in written, verbal, electronic, or parental permission (minors) formats.

Research & Clinical Questions

• Research: descriptive, comparative, casual, exploratory
• Clinical questions: therapy/intervention, diagnosis, prognosis, etiology, prevention, cost-effectiveness

Level vs Quality of Evidence

• Can be a systemic review (high level of evidence but have poor methods/bad studies) so it can be high level but poor quality

Systematic Reviews

• Systematic review: highest level of evidence for all questions: will gather all the info and research and give a conclusion from a combination of all available research
• Meta-analysis: When data of the studies are compiled and analyzed as one big study (use similar studies to compare), a statistical technique for combining results of independent trials to give a quantitative estimate of effectiveness
• All systematic reviews do not contain a meta-analysis bc they may not include it for qualitative data, not similar studies, or poor quality

Forest Plots

• This is used in a meta-analysis showing results of multiple studies that can show common effects

Populations & Samples

• Target Population: All children w/ asthma (total group of interest -> not possible to get all people who the "population", but it is the group of interest)
• Accessible Population: potential subjects who are available for study (makes it easier). Ex: All subjects w/ a specific condition from the CT state
• Sample: subgroup of accessible population (Think world vs state vs few subjects to test)
• Inclusion*:
• Should be broad enough to maximize sample size but narrow enough to minimize extraneous variables.
• Narrow: High validity
• Broad: Low validity

Exclusion Criteria Characteristics

• This will make individuals ineligible with extraneous variables
• Should not be narrowed too much bc the more restrictive the sample size need

Sampling Methods

• Probabilistic: Randomization occurs at some point when getting sample: simple random, systematic sampling, stratified sampling, cluster sampling

Methods

• Simple Random: Each potential subject has an equal chance of being selected (random number generator used)
• Systematic Sampling: Subjects organized according to identifier. The first subject is selected randomly from the group; remaining subjects are based on numerical distance
• Stratified Random: Randomly select subjects in each stratum that's used to either preserve the proportions of subgroups in population with the sample (proportional) or to ensure subgroups are represented in adequate numbers (non-proportional)
• Cluster: Used when naturally occurring pockets of population are geographically dispersed, then select from there and the pockets

Sampling Method Examples

• Select patient from 3 cities from the same state, 3 categories, 25% of pts in each category: non-probabilistic
• No randomization in getting the participants
• May result in misrepresentation of populating subgroups within sample
• Faster, easier & cheaper to implement
• Convenience Samples: Readily available subjects

Random Assignment

• Assign ppl to a study
• Each participant has 50/50 chance of being in each group
• By block: predetermined # of subjects to be contained in each group: balanced sample size
• Systematic: select members of pop at regular intervals
• Matched: all groups have equal # of subjects with the specific characteristics arranged into subgroups
• Consecutive: random order of group assignments (most common method): study, so as to assign, based on composed list given before

Sample Size

• Larger the sample size, similar #s of participants
• To detect differences
• Researchers must anticipate and deal with attrition to minimize reduction in power
• Larger samples lend to be population representative than small sample sizes

Ways to Assess Risk for Bias of RCTs

• PEDro score and Cochrane
• Look for research validity
• For results to be valid

PEDro Scale

• Assesses levels of bias in articles; ranges from poor to excellent. 0-3 = poor 4-5 = fair 6-8 = good 9-10 = excellent

Registration

• The scale has 11 items but the total score is 10 bc item 1 is doesn't count; this as to do with external validity
• Register prospectively aka must prior register before recruitment of the first person to improve transparence & reduce bias and is mandatory for regulatory agencies: Clinicaltrials.gov

Guidelines

• This is a list of important items that a study must have.
• RCT: Consort Systematic review needs Pirsma

Boolean Operators

• And: Combines terms, finds studies that are combining all the words
• And & not: narrows down results
• Or: Expands results; combining all options.