Evidence-Based Practice Notes

Evidence-Based Practice

• Evidence-based practice (EBP) involves using research findings to inform daily patient care and clinical decision-making.
• It combines personal knowledge, patient condition, and existing research to make informed decisions.
• For example, identifying a problem, finding reliable research to support a solution, and combining evidence with personal knowledge.

Evidence Research in Allied Health Roles

• Monitoring specific diseases and health conditions.
• Analyzing data collected during health interventions, such as adverse events in vaccination programs.
• Producing evaluation reports of health interventions.

Principles of Evidence-Based Practice

• Decisions must be supported by valid and clinically relevant research (internal or external).
• Professional and clinical expertise, client values, circumstances, and practice context information are important.
• Ongoing professional development is crucial for maintaining currency in EBP.

5-Step EBP Approach

1. Ask an answerable practice-related question.
2. Acquire relevant information to answer the question.
3. Appraise the acquired evidence to ensure research trustworthiness.
4. Apply the appraised evidence to practice, integrating expertise, patient needs, and context.
5. Assess performance and set learning goals for improvement.

Research Process

• Systematic investigation:
  • Formulating a research question.
  • Reviewing relevant literature.
  • Deciding on an approach to the question.
  • Using the method to collect and analyze data.
  • Interpreting data in terms of the research question.
• Research questions must be specific and directly answerable.
• Collection of original primary data is known as primary research; reviews of primary research are secondary research.

Quantitative Research

• Tests theories or hypotheses by analyzing relationships.
• Involves measuring study participant characteristics relevant to the research question.
• Measurement: quantifying health-related phenomena and assigning numbers to represent characteristics (physical, behavioral, psychological, or social tendencies).
• Quality is impacted by the reliability of measurements (consistency of results).

Qualitative Research

• Uses inductive logic for data collection and analysis rather than deductive logic.
• Answers "why" and "how" rather than "how much."
• Can lead to the generation of theories on health-related phenomena.
• Provides in-depth understanding of health-related phenomena.
• Helps health practitioners understand attitudes, feelings, perspectives, and beliefs associated with patient experiences.

Systematic Review

• A comprehensive review of existing research meeting specific criteria.
• Must provide in-depth details of the review process:
  • Strategy for finding evidence.
  • Criteria for study inclusion.
  • Criteria for study exclusion.
• Steps include:
  • Formulating a question.
  • Searching literature using predetermined criteria.
  • Critically appraising study quality.
  • Extracting relevant data.
  • Synthesizing data using statistics (if appropriate).
  • Summarizing results.
  • Discussing implications.

Narrative Review

• Conducted in a story-telling fashion.
• Can be a whole paper review or partly reviewed.
• Integrated into the introductory section of a research report.
• Quality varies based on the author's intellectual style.
• Areas of emphasis vary, leading to different conclusions.
• Difficult to control bias.

Clinical Guidelines

• Compiled using an extensive review of research evidence.
• Combine client input and expert opinion.
• Translate research into recommendations for practice.

Step 1 - ASK (PICO)

• Turning information needs into answerable research questions.
• PICO:
  • P = Population and/or clinical problem of interest (e.g., age, gender, diagnosis).
  • I = Intervention (e.g., exposure, test, prognostic factor, treatment).
  • C = Comparison (what the intervention is better/worse than).
  • O = Outcome of interest for the client (e.g., cure, side effects, symptoms).

Types of Clinical Questions

• Background questions: General knowledge about a condition (e.g., why do children have ASD?).
• Foreground questions: Specific knowledge of patient management (e.g., what is the most effective treatment for children with ASD and reading problems?).
• Effectiveness Questions
• Prevention Questions
• Assessment Questions
• Description Questions
• Risk Questions

PICO Questions - Population

• Include primary problem, disease, or co-existing conditions.
• Specify age and gender if relevant.
• Consider:
  • Describing a group with a similar problem.
  • Describing the client to a colleague.
  • Identifying important patient characteristics.

PICO Questions - Intervention

• Considered in its broadest sense.
• What you plan to do for the client or the factor of interest.
• Examples: specific test, treatment, medication, product, or procedure.
• Not necessarily something implemented by a health professional (e.g., comparing smokers to non-smokers).

PICO Questions - Comparison

• Main alternative being considered.
• May be the standard care treatment.
• Comparing the intervention.
• Some questions may only focus on the intervention itself.

PICO Questions - Outcome

• Results you plan to accomplish.
• Examples: relieving symptoms, improving function, preventing conditions.
• Must be measurable.
• Specify whether the outcome involves increasing or decreasing.

The Research Process

1. Develop a hypothesis:
   • Expressed in terms of two variables: proposed cause and proposed outcome.
   • Variables can be independent or dependent.
   • Independent variable: value does not depend on another variable.
   • Dependent variable: value depends on the cause.
   • Manipulate the independent variable and measure the effect on the dependent variable.

Randomized Controlled Trial (RCT)

• Aims to ensure observed effects result from the intervention.
• Sample participants are randomly assigned to intervention or control groups.
• Researchers aim for groups to be similar in characteristics.
• Placebo Effect: Expectation of treatment leading to improvement can change brain neurochemistry.
• RCTs are best for answering intervention questions.
• Limitations: unethical to expose intervention group to suspected causes of disease.

Prospective Cohort Studies

• Participants are followed over time to determine exposure to causal factors and development of disease.
• Best for questions about the cause and prognosis of disease.

Considerations in Research

• What type of question is it (intervention or etiology)?
• Maximize internal validity (certainty about conclusions).
• Strength of evidence depends on study design and methodological quality.

Qualitative Research Principles

• Relies on stories to understand an experience or perception.
• Inductive reasoning to build theory.
• Draw out meanings, perceptions, and understandings that individuals attach to behaviors, experiences, and social phenomena.
• Objectivity: acknowledges values and biases.
• Thick description (behavior and context) over explanation.
• Data reported in words, using stories and quotes to build theory.

Approaches to Qualitative Research

• Narrative Inquiry:
  • Stories or accounts of events from the person involved.
  • Discover meanings individuals give to life episodes.
  • Examine how stories compare and contrast.
• Phenomenological Study:
  • Describes the meaning of lived experiences for a group about a concept or phenomenon.
  • Identify phenomena through individual perceptions.
  • Capture subjective nature of the phenomenon.
• Ethnographic Research:
  • Prolonged observation, often through participant observation.
  • Immersion in social setting, conversations, in-depth interviews, and document gathering.
• Grounded Theory:
  • Focuses on generating data to develop a theory.
  • Useful when little is known about the topic.

Methods of Data Collection

• Individual In-Depth Interview:
  • Researcher talks to participant, inviting them to discuss their life or perspectives.
  • Unstructured: informal tone, brief prompts.
  • Semi-structured: pre-set questions with allowance for expansion.
  • Standardized open-ended: same questions for all participants in a structured way.
• Focus Group:
  • In-depth discussion with a group.
  • Participants share common interest or experience.
  • Generate collective and differing perspectives.
  • Skilled moderator needed to manage dynamics and ensure all viewpoints are heard.
• Clinical Data Mining:
  • Extracting and analyzing data to find patterns.
  • Uses secondary data analysis.
  • Examples: patient characteristics, treatment outcomes.
• To find the appropriate research method for the question ask yourself am I theory testing or theory developing?

Research Structure

• Quantitative Research Structure
• Qualitative Research Structure
• Prospective studies: investigate individuals over time, observing outcomes.
• Retrospective studies: examine past data to understand relationships between exposures and outcomes.

Rigor and Trustworthiness in Qualitative Research

• Criticisms: lacks reproducibility and generalizability, potential for researcher bias.

Qualitative Articles Should Report

• Relevance of question and rationale for approach
• Sampling strategy (selection, saturation)
• Data Collection (interview guide, depth of data)
• Analysis (number of coders, training, code structure, systematic process, software, divergent class, participant confirmation, audit trail)
• Credibility Evaluation
• Transferability
• Dependability
• Confirmability

Qualitative Research - Inductive Inquiry

• Studying a sample and generating a description to enhance understanding of the phenomenon.
• Allows for multiple truths (replicability).
• Crucial idea is that the sample needs to give us new insight into the phenomenon being studied
• Rigour - is the research conducted in a way that allows us to draw on new conclusions and advance our understanding?, is the researchers description of what is happening in this particular sample trustworthy?

Qualitative Research - Deductive inquiry

• Research that begins by assuming a hypothesis about a wider population and taking a sample thought to represent that population and looking for evidence that results in the sample generalized to the wider population
• Crucial idea is that the sample needs to be representative of the population
• Expected different people looking at the same data will reach the same/similar conclusions (replicability)
• Point of both methods of doing research is to advance theory

CASP Qualitative Checklist

• Narrative Analysis
• Thematic Analysis
• Visual Representation
• Matrices or Frameworks
• Graphs and Charts

Critical Appraisal

• Looking at published research papers and determining whether it is high/low quality research
• Non-inferiority trials are designed to show that a treatment is equivalent to another treatment
• N-of-1/single case study involves a single participant

Internal Validity

• Methodological rigor.
• Causality: confidence in causal relationships between variables.
• Dependent variable: the variable researchers either need to understand or be able to explain why it varies
• Independent variable: the variable researchers believe may produce some variation in the dependant variable - need to be confident that x is responsible for variation in y and not some other extraneous variable

External Validity

• The degree in the which the studys findings can be generalized
• Whether causal relationships can be generalized to different measures, persons, settings and times
• Ecological validity: do findings reflect real-life settings?
• Efficacy studies: focus on internal validity through controlled methodology.
• Effectiveness studies: follow efficacy studies in less controlled situations.

Bias

• Systematic error in a study.
• Not measuring what you intended to measure.
• Sample or selection bias
• Measurement or detection bias
• Intervention or performance bias

Limitations of Quantitative Research

• The Hawthorne effect: Participants behave differently when they know they are being scrutinized.

Reliability and Validity

• Are we measuring the right thing and are we measuring it consistently
• Reliability = the extent to which the outcomes are consistent when the experiment is repeated more than once
• Validity = the extent to which the instruments that are used in the experiment measure exactly what you want them to measure

Measurement Reliability

• The greater the measurement error the lower the reliability
• Consistency - correlation statistic that represents the strength of the association between two measurements
• Agreement - level of agreement between two sets of observations
• Reliability of the test as a whole (test-retest reliability)
• Reliability of the tester (intra-tester reliability)
• Reliability of different testers (inter-tester reliability)

Error

• Systematic Error - usually predictable and therefore directional
• Random Error - unpredictable and is scattered about the true value

P-vale

• p<0.05 we conclude that a finding is statistically significant
• p>0.05 we conclude that no statistical significance was detected

Mixed Methods Research

• Collection or analysis of both quantitative and qualitative data in a single study.
• Data collected concurrently or sequentially.
• Integration of data at one or more stages.
• Provides a more complete picture.
• Reasons for using mixed methods:
  • To complement the data: Sometimes the research is seeking elaboration, enhancement, illustration and clarification by mixing the results of one method with those of another
  • To develop the data: When the researcher using the results of the first method to develop or inform research using the other method * To understand paradoxes or contradictions emerging from one type of data: When the researcher follows up on findings that need further explanation
  • To expand the data: When the researcher seeks to extend the breadth and range of a study by using different methods for different components of the study

Types of Mixed Methods Designs

• Convergent Parallel Design:
  • Obtain different but complementary data on a topic
  • Might include cross-sectional data in a survey, focus groups data and in-depth interviews all occurring at the same time, with the results of each then brought together
• Explanatory Sequential Design:
  • Two-phased study
  • First stage involves collection of quant data that specifically addresses study's questions
  • Second stage involves collection of qual data to help explain or build on initial quant results
• Exploratory Sequential Design:
  • Begins with and prioritises qualitative data
  • Results of the qual stage are used to develop the quant stage
  • Premise that exploration of an issue or concept is required first
• Embedded Design:
  • One dataset alone cannot fully explain existence of patterns in the data
  • Offset these limitations by adding different types of data that offer a secondary role

Categorical Variables

• Made of distinct categories
• Binary Variables - have two possibilities e.g. yes/no, true/false
• Nominal Variables - use labels that cannot be orders e.g. colour, religion, postcode

Continuous Variables

• Gives us a score for each person and can take on any value on the measurement scale of interest
• Interval Variables - measured using any scale that has equal distances between the intervals on the scale (equality of units) e.g. temperature, sea level
  • Has no absolute zero
• Ratio Variables - same property as interval variables but also have a zero point e.g. age, height
  • Can be rescaled into nominal, ordinal and ratio scales

Descriptives of data

• Normal distribution - the bell curve
• Descriptive statistics are used to summarise or condense a group of scores
• Include measure of:
  • Central tendency scores
    • Mean = mathematical average of the scores
    • Median = middle point on an ordered distribution
    • Mode = most frequently occurring score
  • Variability scores - standard deviation, variance, range of scores

Inferential Statistics

• Ratio
• Proportion
• Rates
  • Incidence Rate - number of new cases divided by total population at risk of developing the disorder
  • Prevalence Rate - number of existing cases divided by the total population at risk of developing the disorder
• Mean
• Median
• Mode

Statistical Hypotheses

• Null Hypothesis
• Alternative Hypothesis
• Null Hypothesis Significance Testing

P-Value

• A smaller SE = more precise estimate of the population mean
• MOE = how much you can expect the sample result to differ from the true population value. Range of uncertainty around your estimate
• Larger sample size decreases MOE
• *Higher confidence level increases MOE
• More variability in data increases MOE
• Confidence Interval (CI)

Ethics in Research and Practice

• Human Research and Ethics Committees (HRECs)
• Applications must include information about:
  • The proposed research design and procedures
  • The prospective human participants
  • Recruitment and informed consent processes
  • Possible impact of research on participants and others
  • Control and storage of data
• Informed Consent
• Researchers should be attuned to whether participants are limited in their ability to fully understand what their participation might involve
• Data guards against participant identification with the use of pseudonyms and keeping identifying information in a separate place from data