Evidence & Research Questions

What is Evidence?

Information that is systematically obtained and critically appraised.

- Can be used to inform decisions.

Scientifically, evidence is:

Empirical information that modifies (increases or decreases) confidence in the truth of a claim, hypothesis, or decision.

Characteristics of Evidence

1. addresses the specific clinical question the PT is trying to answer

2. the subjects studied have characteristics similar to the patient whom the PT has a clinical question

3. the context of the evidence and/or the technique of interest are consistent with contemporary health care

4. the evidence was published in a peer-reviewed medium (paper, electronic)

sources of evidence - primary

Original research reports

Found in:

• Peer reviewed journals

• Proceedings from professional meetings

• Theses and dissertations

• Websites

sources of evidence - secondary

Summary reviews of works of other researchers

Found in:

• Systematic and narrative reviews in peer-reviewed journals

• Textbooks

• Practice guidelines

• Websites

Two Distinct Contexts of Clinically-Relevant Questions

FINER

PICO

FINER

• Decides if a question should become a study.

• Evaluates research value

• Used mainly by researchers

• Asks “Should we study this?”

PICO

• Guides question formulation to ensure it can be answered by evidence.

• Structures clinical questions

• Used daily by clinicians and students

• Asks “What does the evidence say?”

Elements of a Good Research Question

Answerable, Important, Relevant, Specific

OR

Feasible Interesting Novel Ethical Relevant

Types of Questions - Background

Usually about general knowledge.

- Help understand the condition, the population, or the concepts.

- “What is this?” or “How does this work?”

Examples:

- What is tendinopathy?

- What are common causes of stroke?

Types of Questions - Foreground

- Usually about clinical decisions.

- Help with the selection of what to do for a specific patient.

- “What should I do for this person?”

Examples:

- In young adults with Achilles tendinopathy, does eccentric exercise improve pain more than rest?

- In patients post-stroke, does task-specific training improve SSC gait speed compared to usual care?

* Foreground questions are typically structured (often PICO).

Elements of a good question - PICO

Person, problem, or population

issue

comparison

outcome

PICO - “P” element

person, problem, or population (i.e., relevant personal and/or clinical details such as age, gender, diagnosis, acuity, severity, and/or preferences)

PICO - “I” element

issue (i.e., diagnostic test, clinical measure, prognostic factor, intervention, clinical prediction rule, outcome, self-report outcome measure of interest, patient or client perspective);

PICO - “C” element

comparison (i.e., a comparison test, measure, predictive factor, intervention, clinical prediction rule, outcome, or self-report outcome measure);

PICO - “O” element

outcome (i.e., the consequence[s] of applying or including the issue of interest).

Simple: Which is the most accurate special test (I) to use to rule in a rotator cuff tear (O) in a 65-year-old male with shoulder pain (P)?

Comparative: Is the Neer’s Test (I) more accurate than the Hawkin’s Test (C) for ruling in a rotator cuff tear (O) in a 65-year-old male with shoulder pain (P)?

Example: Diagnostic Test Question

Simple: Is a manual muscle test (I) a reliable and valid measure of shoulder girdle strength (O) in a 65-year-old male with shoulder pain (P)?

Comparative: Is a manual muscle test (I) as reliable and valid as a hand held dynamometer (C) for measuring shoulder girdle strength (O) in a 65-year-old male with shoulder pain (P)?

Example: Clinical Measure Question

Simple: Does hand dominance (I) predict return to function (O) in a 65-year-old male with shoulder pain (P)?

Comparative: Is hand dominance (I) a more accurate predictor of return to function (O) than duration of symptoms (C) in a 65-year-old male with shoulder pain (P)?

Example: Prognosis Question

Simple: Is joint mobilization (I) effective in restoring functional use of the UE (O) in a 65-year-old male with shoulder pain (P)?

Comparative: Is joint mobilization plus exercise (I) more effective than exercise alone (C) in restoring functional use of the UE (O) in a 65-year-old male with shoulder pain (P)?

Example: Intervention Question

Simple: Can a clinical prediction rule (I) predict the functional outcome (O) in a 65-year-old male with shoulder pain whose first physical therapy consultation occurred 6 weeks prior (P)?

Comparative: Is a clinical prediction rule (I) more accurate for predicting the functional outcome (O) at 6 weeks (I) or 6 months (C) in a 65-year-old male who seeks an initial physical therapy consultation for shoulder pain?

Example: Clinical Prediction Rule Question

Simple: Will a 65-year-old male with shoulder pain (P) achieve a meaningful change in outcome (O) with modalities (I)?

Comparative: Will a 65-year-old male with shoulder pain (P) achieve a meaningful change in outcome (O) with modalities (I) or with therapeutic exercise (C)?

Example: Outcomes Question

Hypothesis

In evidence-based practice, a hypothesis is:

A testable prediction about the relationship between variables, evaluated through systematic evidence that increases or decreases confidence in the claim.

NULL HYPOTHESIS (Ho):

• Statistical hypothesis

• There will be NO STATISTICALLY SIGNIFICANT DIFFERENCE (or relationship) between groups in the study

• Any differences (relationships) observed are due to chance

Study Objective: Intervention to improve dynamic balance

Experiment: Children with cerebral palsy randomized to lower extremity strengthening program or standard physical therapy care

Ho: There will be no (statistically significant) difference in the control of dynamic balance (via TUG test) between children who engage in the leg strengthening program versus children performing standard care.

Null Hypothesis Example

Research (Alternate) Hypothesis (Ha)

Investigator expects to find a difference (or a relationship)

Once committed to this expectation (usually based on literature), hypothesis must be maintained!

Ha: Improved control of dynamic balance will be observed (via statistically significant faster TUG test completion) among children who engaged in the leg strengthening program compared with children who receive standard care.

Research Hypothesis Example

Finding the Evidence - General Plan of Attack

• Prioritize type of study that can give you the most information

• Determine which database will be most useful or efficient for your search

• Identify search terms and their synonyms to use in the search

Finding the Evidence - Push Approach

Objective: information capture

User: passive

Technology: always active

Best for:

• continuing knowledge

• urgent or time dependent information

• creating awareness of what can be pulled (notification)

Main advantage: brings information issues to the user’s attention

Main disadvantage: can overload users since they do not control it

Finding the Evidence - Pull Approach

Objective: navigation

User: active

Technology: active only when used

Best for:

• one time knowledge needed

• conducting research

• detailed information

Main advantage: allows users to access information at the point of need

Main disadvantage: required time on the part of the user (to both master the technology and to deploy it)

Which articles to choose?

• Relevance to the question posed

• Consistency with contemporary practice

• Ranking on evidence hierarchy / Peer-reviewed

• Disclosure of funding sources or conflicts of interest