Who is the Audience?

• Healthcare professional or patient?
• information need
• education level
• reading level
• language spoken
• cultural considerations

Clinician's Role

• must see importance of interactions with patients and other health care providers which leads to opportunities for work collaboration and direct involvement in patient care 
• when presented with a clinical quest ion, anticipate additional questions and concerns
• listen and have great communication skills
• be aware of false perceptions that a clinical scenario/drug info question does not pertain to a patient specific question
• be aware of casual interactions with requestors and the lack of formal consultations and that they preclude the need for in depth analysis and extensive involvement with direct patient care

Desired Characteristics of

Clinical Question Response

• timely, current, accurate
• complete (concise), well referenced
• clear and logical
• free of bias
• applicable and appropriate
• answers important related questions

7 steps of Answering

Clinical Questions

• 1: secure demographics of requestor
• 2: obtain background information
• 3: determine ultimate question
• 4: develop strategy and conduct search
• 5: perform evaluation, analysis and synthesis
• 6: formulate and provide response
• 7: conduct follow-up and documentation 

Secure Demographics of Requestor

• Step 1 of Answering Clinical Questions
• who is asking the question? (name, location, phone, number/address/pager
• how do they want the response delivered?
• when do they need the response/answer?
• What resources has the requestor already utilized?

Obtain Background Information

• Step 2 of Answering Clinical Questions
• (most difficult step to complete: based on original question) 
• is the request patient specific or academic?
• if patient specific-patient's diagnosis?-Pertinent medical information?
• is the question being asked by true requestor or an intermediary or the patient? this may complicate the process

Determine Ultimate Question

• Step 3 of Answering Clinical Questions
• most often the original question is NOT the most effective way of finding the answer need/wanted
• if adequate background informatino is obtained, the ultimate question is usually found (be sure to focus on background)
• once ultimate question is found, confirm this with requestor to avoid conflict with response
• Use PP-ICO to create your question

PP-ICO

• used in Step 3 of Answering Clinical Questions (determine ultimate question)
• P-Problem
• P-Patient
• I-Intervention
• C-Comparison
• O-Outcome

Develop Strategy and Conduct Search

• Step 4 of Answering Clinical Questions
• if question relates to a specific patient, you must collect patient-specific data OR get appropriate background information from requestor
• to create patient-specific database that is complete, collect the following: patient factors(demographics, diagnosis, CC, HPI, past medical history, social history, ROS, medications, laboratory tests)
• disease factors: pathophysiology, signs and sympt, treatment, risk factors, complications, prognosis
• medication factors: pharm, adverse effects,monitoring, dosage, ect
• most questions require multiple references to get appropriate answer
• conduct search via tools learned earlier (pubmed, tertiary resources ect)

Filtered Information

• highest quality of information
• Systematic reviews (highest quality
• crically-appraised topics (evidence synthesase)
• critically-appraised individual articles(article synopses)

Unfiltered Information

• lower quality of evidence
• randomized controlled triale(RCTs) (highest)
• Cohort studies
• case-controlled studies (case series/reports)
• background information/expert opinion

Perform Evaluation, Analysis and Synthesis

• Step 5 of Answering Clinical Questions
• guess what? Evaluating clinical literature
• utilize skills learned to critique the validity, relevance and clinical applicability of the literature found in search
• PP-ICONS approach
• IF PP-ICO used to form your question is similar to PP-ICONS used to evaluate your articles, then you have answered your question

PP-ICONS Approach

• used in Step 5 of Answering Clinical Questions (perform evaluation, analysis and synthesis)
• Problem
• Patient or Population
• Intervention
• Comparison
• Outcomes
• Number of subjects
• Statistics 

Formulate and Provide Response 

• Step 6 of Answering Clinical Questions
• questioning strategy(ask logical questions): do i have a clear understanding? Do i know if corect question is being asked? do i know requestors expectations? Do i know the circumstances that generated the question? Do i have insight?
• prior to providing response, think about how information may be used-reference the background information
• inform the requestor of accurate and unbiased clinical literature

Conduct Follow Up and documentation

• Step 7 of Answering Clinical Questions
• completed after the verbal communication response is given to requestor
• follow-up is perceived as good clinical practice and enhances professional relations
• opportunity to readdress orginal request when newer literature/outcomes develop

Null Hypothesis (H₀)

• Group A=Group b
• determines whether any observed differences btw groups were caused by chance alone
• tests for significance
• reject->difference btw groups(unlikely to chance)
• accept -> no difference btw groups(diff due to chance)

Sample size needed for

80% power

• you would need 1,000 patients

Nominal Data

• two groups(or 2 drugs), no rank or order
• sex, patients at goal
• yes or no questions(die or didnt die)

Ordinal Data

• values can be ranked, no consistent magnitude of difference
• NIH stroke scale
• patient satisfaction scores

Continuous Data

• consistent magnitude of difference
• temperature, blood glucose
• most common in medical studies
• Interval(no abs zero) and ratio 

Central Tendency

• mean(used for coninuous data, sensitive to outlier)
• median(half of the values lie above, half of the values lie below; insensitive to outliers; used in ordinal and continuous data)
• mode (most common value observed)

Variability

• usually standard deviation, used normally distributed, continuous data
• ± 1 SD= 68% of the sample values are found
• ± 2 SD = 95% of the sample values are found
• can also use range, precentile or standard error of the mean (SEM)

Standard Error of the Mean

(SEM)

• used in variability (descriptive statistics)
• statistical trick
• makes the variability in the data appear smalled than it is
• actuallly an estimation of the error in the study's mean as compared to the population's mean
• not a measure of variability of data

Type I Error

• claiming there is a difference when one does not truly exist
• falsely rejecting the null hypothesis
• alpha-the amount of risk of a type i error that we can stomach(determined by ahead of time)
• customarily set at 0.05 or 5% 

P-Value

• chance of a type I error occuring with our results
• determined by statistical testing
• if less than our alpha, a significant difference exists

Type II Error

• saying there is no significant difference when realy there is
• falsely accepting the null hypothesis
• probability of making a type II error is β (normally set between .1 and .2)
• power= 1 - β

Power Analysis

• must be calculated before 
• relies on certain assumptions to determine the number of patients needed PER GROUP(1k in placebo and 1k in drug) to detect a difference: event rate, population mean, population standard deviation
• smaller differences require larger study populations
• have to have both to happen (enough patients and right percentage in order to meet power)

Confidence Intervals

• can describe both significance and magnitude of difference (p values only do significance)
• made up of a point estimate for the result and a range of potential results
• if it includes 0, can be interpreted as p>0.05
• the larger the confidence interval, the larger the risk of type II error
• for absolute values, interval should not cross 0, for ratios, interval should not cross one

Absolute Risk

• first you calculate event rate
• absolute risk reduction (or increase) you subtract event rates
• 2/6033 and 53/6031 = 0.03 and 0.9
• .9-.03 = .87% absolute risk increase

Relative Risk

• divide the absolute risk of the two groups (.9%/.03 = 30(30x increased risk)
• ARR/risk in control group (0.87%/0.03 = 2900% increased risk
• be weary of relative risk, artificially inflates effectiveness and adverse events (try to stick with absolute results as much as possible)

Number Needed to Treat or Harm

• calculated by dividing 1 by absolute risk reduction (or increase)
• for our SEARCH Trial example: 1 ÷ 0.0087 = 114.94
• because we cant have a fraction of a person, round up to 115
• only calculated for nominal data, statistically significant results, and the time frame of the study

Intention to Treat

• all data is included in the analysis, regardless of wheter the patient completed the study
• makes data more conservative
• often described as more "real world"
• versus per protocol

Per Protocol

• only patients that followed protocol perfectly included in analysis
• can give a better idea of the effect of an agent
• versus intention to treat

Why Do we need EBM?

• expanding literature base of research
• limited amount of time and human mental capacity
• clinical questions arise 3.2 times for every 10 patients seen, but clinicians only seek an answer 34% of the time
• clinicians on average take 2 minutes to find an answer for 1 question

Hunting Tools

(information tools)

• "hunting tools" - quick answers to specific questions
• cochrane database
• pub med clinical queries
• USPSTF 
• National Guidelines Clearinghouse

Foraging Tools

• "foraging" -quick periodic updates, areas of interests
• American Family Physician (POEMS)
• Journal of Family Practice (POEMS)
• British Medical Journal Clinical Guidelines
• Journal Watch
• POEMS (Patient oriented evidence that matters)

Usefulness of Medical Information

usefulness = (relevance X Validity)/WORK

History of Jounral Clubs

• 1835-54 refered to by sir james paget of clubs that met in rooms over a bakers shop near st bartholomew hospital
• 1875 oslers journal club in montreal credited as 1st journal club in north america

Journal Clubs in Clinical Practice

• accreditation board mandates that all residency programs teach and asses various competencies
• practice-based learning is one of these competencies. journal clubs can be used as a tool to teach and assess this competency in residency programs

PP-ICONS in practice

• a way to screen for validity and relevance of an article
• most often times, can screen the abstract of the article for this approach
• allows a quick, effective screening tool when evaluating literature

Evaluating -Relevance and Validity

• POEMs (patient oriented evidence that matters vs DOE (disease-oriented evidence)
• 3 screening questions - would this info answer my question, are the subjects representative of mine?, would this require me to change my current practice?  

Evidence Based Medicine

• empiric testing
• controlled trials
• good clinical practice
• balanced with shared decision making

Shared Decision making

• patient centered
• empowerment
• relationship building
• PCMH
• balanced with evidence based medicine

POEM

• patient oriented evidence that matters
• adresses a clinical issue commonly encountered
• patient centered outcome
• has potential to change practice

DOE

• disease oriented evidence
• disease centered outcomes
• can often be premature

Self management goals

• improve outcomes, and patient attitudes
• set goal(could be small)
• make a plan
• get to work
• stick to it
• reach goal

Steps for Effective Communication

• understand expectations
• build partnerships
• provide the evidence (RISK) 
• make recommendation
• ensure understanding
• answer any questions 

Critically Appraised Topic (CATs)

• evidence syntheses
• essential evidence plusnational guideline clearinghouse
• bandolier
• trip database(focused on ebm materials)
• everything from prevention, to treatment, to prognosis

Critically Appraised Individual Articles

• POEMs (patient oriented information that matters)
• evidence updates
• ACP journal club
• bandolier  

Levels of evidence

• used to give the reader a quick answer for the quality of the resource/guideline
• multiple versions available
• can differ for different types of questions: treatment; prognosis; diagnosis; etc

When do i need to go from a filtered

Resource to the original resource?

• making major changes to an individual patients care
• my patient does not match the clinical guidelines
• assignments/studying for boards

Information Mastery 

• patient is the center of all care decisions
• keeping up with new information without being overwhelmed
• choosing the best quality evidence for the question being asked 

FIRE Approach

• F-formulate a question
• I-information search
• R-Review of information and critical appraisal
• E-Employ the results in your clinical practice 

FRISBEE

• validity of research
• F-Follow-up
• R-Randomization
• I-Intention to treat analysis
• S-Similar baseline
• B-Binding
• E-Equal Treatment
• E-Equivalence to your patient