We use a single Google Analytics cookie to count anonymous visits and improve the textbook. No advertising, no profile building. You can decline and the site still works.
Part 1/Chapter 1/5-min read
Evidence, Guidelines, Registries, and Living-Textbook Method
How to read vascular evidence at the bedside: which trial, registry, or guideline statement is strong enough to change management for this patient, this lesion, and this anatomy. The chapter frames evidence appraisal, guideline trustworthiness, and registry use so that recommendations stay tied to current source support.
Listen to this chapter6 min · AI audio edition · two hostsAI narration
Choose a chapter overview or a specialist discussion. Each recording follows this published edition.
General medical education, not patient-specific advice.
Choose an audio edition
0:00 / 5:54
Evidence, Guidelines, Registries, and Living-Textbook Method · MedicalTextbook.ai
1.1Evidence appraisal and methodological standards
Evidence-based vascular care integrates the best current evidence with patient anatomy, operative risk, life expectancy, symptom burden, care goals, and local procedural performance .
GuidelinesEvidence appraisal tools by study design
Evidence type
Primary appraisal focus
Prescribed tool
Citation
Parallel-group randomized trial
Reporting completeness and bias domains
CONSORT 2010, RoB 2
Observational intervention study
Reporting transparency and confounding
STROBE, ROBINS-I
Diagnostic accuracy study
Reporting completeness and patient-selection bias
STARD 2015, QUADAS-2
Prognostic or prediction model using regression or machine learning
Reporting completeness; development quality and applicability; evaluation risk of bias and applicability
TRIPOD+AI (reporting), PROBAST+AI (appraisal)
Systematic review and meta-analysis
Search rigour and synthesis credibility
PRISMA 2020, AMSTAR 2
Parallel-group randomized trial
Primary appraisal focus
Reporting completeness and bias domains
Prescribed tool
CONSORT 2010, RoB 2
Citation
Observational intervention study
Primary appraisal focus
Reporting transparency and confounding
Prescribed tool
STROBE, ROBINS-I
Citation
Diagnostic accuracy study
Primary appraisal focus
Reporting completeness and patient-selection bias
Prescribed tool
STARD 2015, QUADAS-2
Citation
Prognostic or prediction model using regression or machine learning
Primary appraisal focus
Reporting completeness; development quality and applicability; evaluation risk of bias and applicability
Prescribed tool
TRIPOD+AI (reporting), PROBAST+AI (appraisal)
Citation
Systematic review and meta-analysis
Primary appraisal focus
Search rigour and synthesis credibility
Prescribed tool
PRISMA 2020, AMSTAR 2
Citation
Randomized trials are evaluated for reporting completeness using CONSORT 2010 and for methodological bias using RoB 2 . Observational evidence, including cohort, case-control, and registry analyses, is evaluated for reporting transparency using STROBE and for risk of bias using ROBINS-I . Diagnostic accuracy studies underpinning vascular imaging and physiologic tests are evaluated for reporting completeness using STARD 2015 and for methodological quality and applicability using QUADAS-2 . Systematic reviews and meta-analyses are evaluated for reporting completeness using PRISMA 2020 and for methodological credibility using AMSTAR 2 .
1.2Guideline development and recommendation strength
Guidelines translate evidence into structured practice recommendations. The GRADE methodology provides the standard vocabulary for rating evidence certainty and guidance strength . GRADE rates certainty of evidence as high, moderate, low, or very low. Randomized-trial evidence starts at high certainty and observational evidence at low certainty; certainty is rated down for risk of bias, inconsistency, indirectness, imprecision, and publication bias, and rated up for a large effect magnitude, a dose-response gradient, or plausible residual confounding that would bias against the observed effect. The Evidence to Decision framework incorporates clinical priority, expected effects, evidence certainty, patient values, resources, cost-effectiveness, equity, acceptability, and feasibility .
A strong recommendation applies to most patients and supports default practice unless contraindicated by patient-specific factors. A conditional recommendation requires explicit individual selection based on benefit-harm trade-offs and patient preferences . ACC/AHA guidelines grade every recommendation by Class of Recommendation and Level of Evidence . Class I means benefit far outweighs risk (is recommended), Class IIa that benefit outweighs risk (is reasonable), Class IIb that benefit equals or marginally exceeds risk (may be considered), and Class III captures either no benefit or harm. Level A rests on high-quality evidence from more than one randomized trial or meta-analyses of high-quality trials, B-R on moderate-quality evidence from one or more randomized trials, B-NR on nonrandomized or observational studies, C-LD on limited data, and C-EO on expert opinion.
Guideline trustworthiness and methodology are appraised using AGREE II, Guidelines 2.0, and NASEM standards, which assess scope, stakeholder involvement, development rigour, conflict management, and external review . Recommendation strength reflects the balance of benefits and harms rather than citation volume alone .
1.3Registries and routinely collected data
Registries capture real-world practice, uncommon presentations, device use, and outcomes in anatomic subgroups typically excluded from randomized trials. Observational databases remain vulnerable to selection bias, unmeasured confounding, changing definitions, missing data, and variation in follow-up. Studies using electronic health records, claims data, and disease registries are reported according to the RECORD extension to STROBE . Registry analyses used to infer causal intervention effects require explicit bias assessment with ROBINS-I .
1.4Artificial intelligence and prediction models
Prediction scores, risk calculators, and diagnostic classifiers require transparent reporting, external validation, and calibration to the local target population. TRIPOD+AI supersedes the TRIPOD 2015 reporting checklist and provides harmonised reporting guidance for studies developing or evaluating clinical prediction models using regression or machine-learning methods . Reporting completeness is distinct from methodological appraisal: PROBAST+AI assesses quality and applicability of model development, and risk of bias and applicability of model-performance evaluation . Randomized trials of AI-supported interventions are assessed using the CONSORT-AI extension .
1.5Clinical integration and living updates
Clinical integration matches the patient-specific uncertainty to the corresponding evidence tool. Follow-up plans and surveillance tests rely on diagnostic accuracy and prediction-model calibration proven in similar patient populations. Evidence application is fundamentally constrained by false-positive findings, low prior probability, inadequate statistical power, bias, and multiple testing . Avoidable research waste stems from poor question prioritization, weak methodology, and incomplete reporting .
Living systematic reviews continually incorporate newly available evidence. They are particularly relevant when a question matters to decisions, current certainty is limited and emerging research could alter conclusions. A living review needs a planned surveillance schedule and explicit update processes; continual surveillance does not mean every new study changes practice .
1.6Areas of controversy
The translation of observational registry data into causal treatment effects remains methodologically controversial, despite formal frameworks such as ROBINS-I and RECORD . The criteria for adopting artificial-intelligence models into routine vascular decision-making lack universal consensus, specifically regarding the required degree of external validation versus local calibration . Living evidence synthesis requires planned reassessment as evidence accumulates; review updates and changes to practice recommendations are distinct decisions .
Further reading
Methods and guidance on conducting, reporting, publishing, and appraising living systematic reviews: a scoping review. (2023). Methods for conducting, reporting and appraising living systematic reviews; methodological rather than disease-specific chapter.
References
1.
Evidence based medicine: what it is and what it isn't. 1996.
PubMed-indexed article1996
Evidence based medicine: what it is and what it isn't. 1996. doi:10.1136/bmj.312.7023.71.
Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): the TRIPOD statement.
PubMed-indexed article2015
Collins GS, Reitsma JB, Altman DG, Moons KG. Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): the TRIPOD statement. Ann Intern Med. 2015;162(1):55-63. doi:10.7326/M14-0697. PMID:25560714.
AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. 2017.
PubMed-indexed article2017
AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. 2017. doi:10.1136/bmj.j4008.
TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. 2024.
PubMed-indexed article2024
TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. 2024. doi:10.1136/bmj-2023-078378.
Further Evolution of the ACC/AHA Clinical Practice Guideline Recommendation Classification System. 2016.
PubMed-indexed article2016
Halperin JL, Levine GN, Al-Khatib SM, et al. Further Evolution of the ACC/AHA Clinical Practice Guideline Recommendation Classification System: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation. 2016;133(14):1426-1428. doi:10.1161/CIR.0000000000000312.
Methods and guidance on conducting, reporting, publishing, and appraising living systematic reviews: a scoping review. 2023.
PubMed-indexed article2023
Methods and guidance on conducting, reporting, publishing, and appraising living systematic reviews: a scoping review. 2023. doi:10.1186/s13643-023-02396-x.
PROBAST+AI: an updated quality, risk of bias, and applicability assessment tool for prediction models using regression or artificial intelligence methods.
PubMed-indexed article2025
Moons KGM, Damen JAA, Kaul T, Hooft L, Andaur Navarro C, Dhiman P, Beam AL, Van Calster B, Celi LA, Denaxas S, Denniston AK, Ghassemi M, Heinze G, Kengne AP, Maier-Hein L, Liu X, Logullo P, McCradden MD, Liu N, Oakden-Rayner L, Singh K, Ting DS, Wynants L, Yang B, Reitsma JB, Riley RD, Collins GS, van Smeden M. PROBAST+AI: an updated quality, risk of bias, and applicability assessment tool for prediction models using regression or artificial intelligence methods.. BMJ (Clinical research ed.). 2025;388:e082505.
Questions, clarifications and conversations with fellow readers and the editorial team. Comments belong to the discussion; they do not change the chapter.
You can send feedback without signing in. If you are signed in, its status appears in your free subscription. Keep patient information and contact details out of your comment.
Educational use only
AI assists this editorial workflow. Published updates are human-reviewed before publication.
Not intended to diagnose, monitor, predict, prognose, treat, or alleviate disease.
Verify clinically relevant information against primary sources and current guidelines.