Free scoping read. A written assessment of your manuscript in two business days, at no cost.

Start one
Contents

Academy · Research Data, Analytics & AI · Applying QUADAS-2

Assessing the Quality of Diagnostic Accuracy Studies Using the QUADAS-2 Tool

QUADAS-2 is applied in four explicit steps: specification of the review question, tailoring of the tool, construction of a flow diagram, and domain-by-domain judgment of risk of bias and applicability. Its developers recommend piloting the tool on a small number of included studies before applying it to the full set.

Pubrica AcademyWritten by a systematic review methodologist Reviewed bya biostatistics reviewer Published10 October 2026 Last checked10 October 2026 Reading time8 minutes Sources10 peer-reviewed

Assessing study quality with QUADAS-2 in a diagnostic accuracy systematic review involves four defined steps: specifying the review question with sufficient clarity to establish what constitutes the index test and reference standard; tailoring the tool’s signalling questions to the specific review topic; constructing a flow diagram for each included study showing how patients moved through recruitment and testing; and judging each of the four domains (patient selection, index test, reference standard, and flow and timing) for risk of bias and, for the first three domains, applicability concerns. Each step is specified by the tool’s developers, and omitting or compressing any of them is a documented and common source of inconsistent or unreliable risk-of-bias judgments.

Learning objectives

  • The four explicit steps specified by the developers of QUADAS-2 for applying the tool
  • The reasons for which tailoring the tool to the specific review question is a required step rather than optional customization
  • The process by which the final risk-of-bias and applicability judgments are reached and reported

Common misconceptions

  • “QUADAS-2 can be applied directly off the shelf, using its standard signalling questions unchanged, for any diagnostic accuracy review.” The developers of the tool specify tailoring as an explicit, required step: review teams are expected to review the signalling questions and, where necessary, add review-specific questions before applying QUADAS-2, because a generic set of questions written to cover diagnostic accuracy research broadly will not capture every bias relevant to a specific clinical question or test type.

Step 1: Specification of the review question

The first step specified by the developers of QUADAS-2 is to define the review question with sufficient precision to identify what constitutes the index test, the reference standard, the target condition, and the population and setting of interest, since every subsequent signalling question is answered relative to this specific definition1. This is important because the same primary study may be judged differently depending on how the review question is framed; for example, a study’s patient population may be judged appropriate for one review question (screening in primary care) and inappropriate for another (diagnosis in a specialist referral population), even though the primary study itself is unchanged1. Review teams are advised to record this specification explicitly before beginning assessment, both to maintain consistent judgments across assessors and across all included studies and to make the basis for applicability judgments transparent to readers of the completed review1.

Step 2: Tailoring of the tool to the review

QUADAS-2's own background documentation, maintained by the University of Bristol group that developed the tool, specifies tailoring as a required second step: review teams should review the standard signalling questions under each domain and add review-specific signalling questions where the generic set does not adequately capture a bias relevant to their particular review topic2. For example, a review of a test that requires substantial operator skill to perform or interpret might add a signalling question on operator experience or training under the index test domain, since this is a plausible, review-specific source of bias that the generic QUADAS-2 questions do not explicitly address2. The developers explicitly recommend piloting the tool by applying it first to a small sample of included studies and checking that assessors interpret and apply both the standard and any added questions consistently, before using it across the full set of included studies12.

Step 3: Construction of a flow diagram for each study

The third step is the construction of a flow diagram for each included study, showing the number of patients at each stage (eligible, included, tested with the index test, tested with the reference standard, and excluded, with reasons), closely resembling the flow diagram that STARD requires primary studies to report134. This step is necessary because the flow and timing domain cannot be judged reliably without understanding how patients moved through the study; it is at this stage that issues such as whether all enrolled patients received both the index test and the reference standard, and whether the interval between the two was appropriate for the target condition, become apparent1. Reviewers who omit this diagram and attempt to judge flow and timing directly from the narrative methods section of a study risk missing exclusions or gaps in testing that are easily overlooked without a structured, step-by-step accounting1.

Step 4: Domain-level judgments of risk of bias and applicability

The final step is to answer the signalling questions (standard and any tailored questions) for each domain and to use those answers to reach an explicit risk-of-bias judgment (low, high or unclear) for all four domains (patient selection, index test, reference standard, and flow and timing), together with an applicability judgment for the first three domains, recording a justification for each judgment alongside the rating1. The developers of QUADAS-2 recommend that this be done by at least two independent assessors, with disagreements resolved through discussion or by a third assessor, and the completed judgments are then typically presented as a graphical summary in the finished review (often a traffic-light figure showing the ratings of every included study across all domains)1. The Cochrane Handbook for Diagnostic Test Accuracy reviews further recommends using these completed assessments analytically, for example by conducting sensitivity analyses that exclude studies at high risk of bias to determine whether a review’s conclusions depend on their inclusion, rather than treating the risk-of-bias table as a standalone reporting exercise disconnected from the review’s findings3.

Application of the four domains: a worked example

A representative, illustrative example clarifies how the four domains relate to the concrete questions an assessor answers. For patient selection, the assessor asks whether a consecutive or random sample of eligible patients was enrolled and whether a case-control design was avoided, since the foundational empirical work of Lijmer and colleagues showed that case-control designs can substantially inflate estimated accuracy relative to a representative, consecutively enrolled sample5. For the index test domain, the assessor asks whether the test results were interpreted without knowledge of the reference standard results, and whether any threshold used to classify results as positive or negative was pre-specified rather than chosen after inspection of the data, since a threshold selected post hoc to maximize apparent accuracy is a well-documented source of bias6. For the reference standard domain, the assessor asks whether the reference standard is likely to classify the target condition correctly, and whether its interpretation was blind to the index test results. For flow and timing, the assessor uses the Step 3 flow diagram directly to check whether an appropriate interval elapsed between the index test and the reference standard, whether all patients received the same reference standard, and whether all enrolled patients were included in the final analysis1. The applicability judgment for each of the first three domains then addresses a related but distinct question: not whether bias is likely, but whether the patients, test execution or reference standard used in the specific study correspond closely enough to the review’s specified question for the study’s results to generalize to the population and setting of interest to the review17. Working through all four domains in this manner, question by question rather than proceeding directly to an overall impression, is what the structured signalling-question format of the tool is specifically designed to enforce. Once all four domains have been judged for every included study, the PRISMA-DTA statement specifies how this risk-of-bias information should inform the review’s reporting: authors are expected to summarize risk-of-bias findings across the whole set of included studies, not only to report them study by study, and to describe how any studies rated at high risk of bias affected the review’s overall interpretation8. Deeks’ overview of systematic reviews of diagnostic test evaluations presents this final reporting step as methodologically as important as the domain judgments themselves, because a risk-of-bias table that is produced but never linked to the review’s stated conclusions has not fulfilled the purpose for which the assessment exists9. The guidance of the Cochrane diagnostic test accuracy methods group on conducting these reviews reinforces the sequencing principle described throughout this article: each of the four specified steps builds on the one before it, in that a clearly specified review question makes tailoring possible, a properly tailored tool makes the flow diagram meaningful, and a complete flow diagram makes a confident, well-justified domain judgment possible; omitting or compressing any single step therefore weakens all that follows10. A review team that works through all four steps in order, documenting each as it proceeds rather than reconstructing the process retrospectively for the methods section, produces an assessment that is both more reliable to apply and considerably easier for a reader, or a subsequent guideline panel, to audit and trust.

Key points

Key points

Appropriate application of QUADAS-2 involves four explicit steps: specification of the review question, tailoring of the tool with review-specific signalling questions, construction of a flow diagram for each study, and judgment of all four domains for risk of bias and applicability by two independent assessors. The completed assessment should be used analytically rather than solely as a reporting formality.

Practical application

Practical application

  1. The review’s exact definitions of the index test, reference standard, target condition and population should be recorded before assessment begins. This practice promotes consistent judgments and makes the reasoning about applicability transparent to readers.
  2. The tool, including any tailored questions, should first be piloted on a small sample of included studies. Piloting verifies that independent assessors interpret and apply the questions consistently before assessment is extended to the full set of included studies.
  3. The completed risk-of-bias table should be used analytically. A sensitivity analysis excluding studies at high risk of bias indicates whether the review’s conclusions depend on their inclusion, rather than the table merely being reported.

Frequently asked questions

Specify the review question, tailor the tool with review-specific signalling questions, construct a flow diagram for each included study, and judge risk of bias and applicability domain by domain.

No; the tool’s background documentation specifies tailoring as a required step, and reviewers are expected to add review-specific questions where the standard set does not adequately capture a bias relevant to their topic.

The flow and timing domain cannot be judged reliably without a structured accounting of how patients moved through recruitment, testing and exclusion; constructing the diagram reveals gaps and exclusions that are easily missed when reading narrative methods sections.

At least two independent assessors are recommended, with a documented process (discussion or a third assessor) for resolving disagreements.

In addition to reporting it, the Cochrane Handbook for Diagnostic Test Accuracy reviews recommends using it analytically, for example by conducting sensitivity analyses that exclude studies at high risk of bias to determine whether the review’s conclusions hold without them.

References

  1. Whiting PF, Rutjes AWS, Westwood ME, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Annals of Internal Medicine. 2011;155(8):529-536. read 10 October 2026. View on Annals of Internal Medicine
  2. University of Bristol. QUADAS-2 Background Document and Guidance. read 10 October 2026. View on University of Bristol
  3. Reitsma JB, Rutjes AWS, Whiting P, Yang B, Leeflang MM, Deeks JJ. Chapter 9: Assessing methodological quality. In: Deeks JJ, Bossuyt PM, Leeflang MM, Takwoingi Y, eds. Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy, version 2.0. Cochrane, 2023. read 10 October 2026. View the Cochrane DTA Handbook
  4. Bossuyt PM, Reitsma JB, Bruns DE, et al. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. BMJ. 2015;351:h5527. read 10 October 2026. View on BMJ
  5. Lijmer JG, Mol BW, Heisterkamp S, et al. Empirical evidence of design-related bias in studies of diagnostic tests. JAMA. 1999;282(11):1061-1066. read 10 October 2026. View on JAMA
  6. Leeflang MMG, Moons KGM, Reitsma JB, Zwinderman AH. Bias in sensitivity and specificity caused by data-driven selection of optimal cutoff values: mechanisms, magnitude, and solutions. Clinical Chemistry. 2008;54(4):729-737. read 10 October 2026. View on Clinical Chemistry
  7. Mallett S, Halligan S, Thompson M, Collins GS, Altman DG. Interpreting diagnostic accuracy studies for patient care. BMJ. 2012;345:e3999. read 10 October 2026. View on BMJ
  8. McInnes MDF, Moher D, Thombs BD, et al. Preferred Reporting Items for a Systematic Review and Meta-analysis of Diagnostic Test Accuracy Studies: The PRISMA-DTA Statement. JAMA. 2018;319(4):388-396. read 10 October 2026. View on JAMA
  9. Deeks JJ. Systematic reviews of evaluations of diagnostic and screening tests. BMJ. 2001;323(7305):157-162. read 10 October 2026. View on BMJ
  10. Leeflang MMG, Deeks JJ, Takwoingi Y, Macaskill P. Cochrane diagnostic test accuracy reviews. Systematic Reviews. 2013;2:82. read 10 October 2026. View on Systematic Reviews

Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.

How to cite this article

Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.

Pubrica Academy. (2026, October 10). Assessing the Quality of Diagnostic Accuracy Studies Using the QUADAS-2 Tool. Pubrica. https://pubrica.com/academy/meta-analysis/how-to-assess-the-quality-of-the-studies-using-the-quadas-2-tool/

This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.

QUADAS-2 Application Worksheet

These steps should be followed in order when applying QUADAS-2 to a diagnostic accuracy systematic review.

[ ] Step 1: Written explicit definitions of index test, reference standard, target condition, population, and setting
[ ] Step 2: Reviewed standard signalling questions and added any review-specific ones; piloted on a small sample
[ ] Step 3: Built a flow diagram for each included study
[ ] Step 4: Answered signalling questions and recorded domain-level risk-of-bias and applicability judgments with justification, using two independent assessors
[ ] Produced a graphical summary of all included studies' judgments
[ ] Planned a sensitivity analysis excluding high-risk-of-bias studies

Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.

How Pubrica can help

Pubrica's systematic review team applies QUADAS-2 through all four specified steps, including tailoring, flow diagrams, and dual independent assessment, and documents the process for journal submission.

Find something else

Search the Pubrica Academy

Access is free and requires no account. Enter a topic, or open the full library to filter by category, sub-topic or resource type.

Verification that every step of QUADAS-2 has been followed

A scoping assessment is free and takes two business days. It provides a written assessment of what the work requires and which specialist would undertake it.

Get a free quote

Regulations

Check your project against this rule

  1. 1Your work
  2. 2Your details
  3. 3Send
What do you need?

Material you send is held under signed confidentiality, with access limited to the assigned team and encrypted transfer. It is not used to train anything and is not passed to third-party tools. Ask for the agreement before you send anything you would rather cover first.