finding and appraising CPGs

clinical practice guidelines

help in decisions regarding dx, therapy, or related clinical circumstances

informed by a systematic review of literature

provide recommendations to optimize care

assess benefits and harms

change behavior

purpose of CPGs

make EBP efficient and realistic

make best available research evidence directly applicable to clinical practice

integrate research evidence w/ knowledge from clinical experts

consider research evidence w/ consideration for pt perspectives (evidence based vs evidence informed)

streamlines best available evidence into bits of translatable and clinically meaningful recommendations

SR vs CPG

analyze evidence based on combinations of original studies with limited clinical interpretations vs all of that + summarizing evidence, considering safety and make interpretable recs about how to care for pts

CPG mainly based on SR of evidence, helps interpret and provide easy translation guide

most relevant databases for CPGs

pubmed, pedro, apta

foreground ?s are broad/general

less is more: 2-3 keywords max (ex. body part, condition, intervention) + practice guideline filter as applicable

clinical prediction guides

are not the same as clinical practice guidelines

CPGs on pubmed

do not use clinical queries

search using main page + practice guideline filter

pedro CPGs

simple search (?)

APTA CPGs

under pt care tab: CPG→search key words

AGREE II cpg appraisal tool

=appraisal of guidelines for research & evaluation

structured tool used to assess quality, incl methodological rigor & transparency of CPGs

does NOT evaluate content or clinical accuracy of recs, but how well guideline was developed & reported

AGREE II appraising existing guidelines

judges if a guideline is trustworthy, evidence based/informed, should be used in practice

AGREE II developing new guidelines

a framework to ensure transparent methods, systematic evidence review, clear presentation, proper conflict of interest management

AGREE II structure

23 items across 6 domains

appraisers ask specific questions w/in each domain

AGREE II scoring

1= strongly disagree (poor/absent info)

7=strongly agree (full, high quality reporting)

AGREE II domains

scope and purpose

stakeholder involvement

rigor of development

clarity of presentation

applicability

editorial independence

AGREE II: scope and purpose

overall objectives specifically described

health questions specifically described

specificaly described population

AGREE II: stakeholder involvement

includes indivs from all professional groups

views preferences of target population have been sought

target users are defined

AGREE II: rigor of development

systematic methods for searching

criteria for evidence selection

strengths, limitations

formation methods are explicit

health benefits, side fx, risks considered

explicit link

external review

updating

AGREE II: clarity of presentation

specific (not ambiguous)

different options for management are clear

keys are easily identifiable

AGREE II: applicability

provide advice tools on how to apply

describes facilitators barriers to application

potential resource implications of applications

monitoring or auditing criteria

AGREE II: editorial independence

views of funding body have not influenced content

competing interests of guideline development group have been recorded & addressed

GRADE

grading of recommendations assessment, development and evaluation

method of guideline development focused on rating confidence consumers should have in evidence recommended

what is GRADE

approach for rating evidence in CPGs

allows for evidence to be upgraded/downgraded

clear separation btwn evidence & strength of recs

strives for transparency of judgments made during development of guideline recs

developed for guidelines related to interventions—application to dx and px ?s ongoign

often peer reviews

adopted by variety of int’l and interdisciplinary organizations

how to use GRADE

initial rating of certainty assigned (high/mod/low/very low): assigned based on study design (ex. RCT high, observational/nonrandomized low)

downgrading or upgrading of certainty based on several factors→final rating of certainty

done for EACH outcome

GRADE: downgrading evidence when

risk of bias present

inconsistent results across studies (unexplained heterogeneity) present

indirect evidence

imprecise results (wide CI)

publication bias present/suspected

GRADE: upgrade evidence when

large tx effect

dose response present

tx effect present when plausible factors working against tx are present

GRADE & SR/MAs

reporting guideline instructs authors to report assessment of certainty/confidence in body of evidence for each outcome

this approach is commonly used to determine certainty in a body of evidence (not the same as quality assessment or ROB)

very low certainty

little confidence in effect estimate

true effect is likely substantially different from effect estimate

low certainty

confidence in effect estimate limited

true effect may be substantially different from effect estimate

moderate certainty

moderate confidence in the effect estimate

true effect is likely close to the estimate, but possibly substantially different

high certainty

very confident that the true effect is close to the effect estimate

implications of a strong GRADE guideline recommendation

patients: most people want recommended course of action

clinicians: most pts should receive recommended course of action

policy makers: recommendation can be adopted as policy

implications of a weak (conditional) GRADE guideline recommendation

pts: some/many would want the recommended course of action

clinicians: different choices will be appropriate for different pts, course of action less clear

policy makers: policy making will require substantial debate & involvement of many stakeholders

shared decision making (clinician & patient) needed

AGREE II vs GRADE

critically appraise CPG vs. embedded in recs within CPG and some SRs

CPG + SR

bc of overgeneralization & need for updates, CPGs may need to be complemented with more recent SRs and RCTs for best evidence