Plain-language summary
Written for a general reader. About 300 words, Grade-8 reading level. The full technical review is in the other tabs.
What this review is about
D-dimer is a blood test that measures fragments left over when the body breaks down a blood clot. In emergency departments it is used to help decide whether a patient with chest pain or breathlessness needs a CT scan of the lungs (CT pulmonary angiogram, CTPA) to look for a pulmonary embolism (a clot in the lung). The headline use is as a rule-out test: a negative D-dimer in a patient with low or moderate clinical probability lets the clinician safely send the patient home without the CT scan. Different cutoff strategies exist — a fixed 500 ng/mL cutoff, an age-adjusted cutoff (10 × age in years for patients over 50), and decision-rule-conditional cutoffs (YEARS, PEGeD).
Wells score: a checklist your ED clinician fills in (points for a swollen calf, fast heart rate, recent surgery, prior clot, and similar features) that estimates your probability of a pulmonary embolism before any blood test or scan. Low / moderate / high probability changes which pathway you go on.
Where this review fits in your ED journey. If your ED clinician judged you high-probability for PE (high Wells / Geneva score, suggestive imaging, or hemodynamic instability), you would have gone directly to a CT scan and skipped the D-dimer step — this review does not address that pathway. The findings here apply to the low- and moderate-probability adult outpatients for whom D-dimer is the next step after the Wells / Geneva score.
What we did
We searched PubMed for landmark prospective management-outcome studies of quantitative D-dimer for pulmonary embolism rule-out, against CTPA at presentation OR symptomatic VTE confirmed at 3-month clinical follow-up (the canonical composite reference). We included 5 primary-tier studies (Christopher 2006, ADJUST-PE 2014, YEARS 2017, ADJUST-validation 2019, PEGeD-Kearon 2019) plus one sensitivity-tier study (Lucassen 2015). The target condition is pulmonary embolism; the population is emergency-department or outpatient adults with low-to-moderate clinical probability stratified by Wells (or simplified Wells / Geneva) score.
What we found
Across the 5 primary-tier studies (>13,000 patients):
- Sensitivity very high (~98–100%): D-dimer almost never misses a pulmonary embolism in the low/moderate-probability subgroup it is designed for. PEGeD-Kearon reports 100% Sensitivity (no PE missed at 3-month follow-up among the 1,285 patients ruled-out by the strategy).
- Specificity moderate and cutoff-strategy-dependent (~39–71%): many patients with a positive D-dimer do not have PE on imaging. Specificity rises when you switch from a fixed 500 ng/mL cutoff (~41%) to an age-adjusted cutoff (~47%) to a decision-rule-conditional cutoff like PEGeD (~71%).
Headline pattern: D-dimer is a near-perfect rule-out at the population it is designed for (low/moderate Wells, no active cancer or pregnancy); the choice of cutoff strategy is the principal lever for reducing unnecessary CTPA imaging.
What this means for you
The bottom line: across the included studies, fewer than 1 in 100 patients sent home with a negative D-dimer and a low / moderate Wells score had a clot diagnosed in the next 3 months.
A negative D-dimer (below the chosen cutoff) in a patient with low or moderate Wells / Geneva clinical probability is enough to safely rule out PE without a CTPA. Per-study 3-month VTE rates: YEARS 0.43%, ADJUST-PE 0.3%, PEGeD-Kearon 0.0%, ADJUST-validation 0.7%, Christopher 0.5%, Lucassen 0.4%.
A positive D-dimer (or any D-dimer in a patient with high clinical probability) means the patient should have a CTPA — about half of these will still be negative for PE because D-dimer is elevated in many non-PE conditions. D-dimer is a rule-out test, not a rule-in diagnostic.
Important limitations
- Sensitivity at or near 100% is a structural property of management-outcome studies, not a noise artifact. The 3-month-VTE endpoint, combined with anticoagulation of imaging-positive patients, suppresses the false-negative count. The Subgroups tab exposes the cutoff_strategy axis where this is honestly documented.
- Specificity varies from 39% (YEARS-conditional) up to 71% (PEGeD graduated) across cutoff strategies — a more-than-30-percentage-point spread driven by which cutoff is used, not by underlying biology. Pick the cutoff that matches local practice when interpreting the headline.
- Data came from study abstracts only, not the full papers; back-computed 2×2 cells used reported Sens/Spec and prevalence.
- If you are pregnant, recently postpartum, or have active cancer, your baseline D-dimer is naturally elevated — pregnancy raises D-dimer through normal placental and uterine remodeling, and many cancers (especially metastatic and pancreatic, lung, ovarian, and gastric tumours) raise D-dimer through tumour-related coagulation activation. As a result the standard 500 ng/mL cutoff would call almost everyone "positive" and the rule-out test loses most of its discriminating power. These results do not apply to you, and your clinician will use a different pathway: the pregnancy-adapted YEARS algorithm or direct CTPA imaging with abdominal shielding for pregnancy; a specialised thrombosis-service consultation (often with direct CTPA or a higher D-dimer cutoff) for active cancer.
- This review also excludes high-pretest-probability patients (D-dimer is not first-line in this group; the ED pathway goes straight to CTPA).
Next update
Last search: 2026-04-28. Next planned check: 2026-07-28 (3 months).
Protocol
Diagnostic Test Accuracy Living Review · v1.0.0 (2026-04-28). Pre-specified before extraction; living protocol updates are tracked in the project repository.
PICOTS+R framework
Standard DTA framework: Population, Index test, Comparator, Outcomes, Timing, Setting, Reference standard.
| Population | Adults with clinically suspected pulmonary embolism in an emergency-department or outpatient setting; stratified by Wells (or simplified Wells / Geneva) clinical probability into low / moderate / non-high (high-pretest-probability patients excluded by design — they go directly to CTPA) |
| Index test | Quantitative D-dimer immunoassay (turbidimetric / ELISA / point-of-care) interpreted at one of three pre-specified cutoff strategies: fixed 500 ng/mL FEU, age-adjusted (10 × age in years for ≥50yo), or YEARS-conditional / PEGeD-graduated rule |
| Comparator | Direct-to-imaging (CTPA without D-dimer screen) or no-D-dimer pathway in the equivalent suspected-PE cohort |
| Outcomes | Sensitivity, Specificity, DOR, LR+, LR−, PPV/NPV at variable pre-test probability for pulmonary embolism; 3-month VTE failure rate as a clinical-utility surrogate (the canonical management-outcome endpoint) |
| Timing | Studies published 2006–2026 (Christopher Study era onwards, covering the fixed-500, age-adjusted, YEARS, and PEGeD generations) |
| Setting | Emergency departments and outpatient ambulatory clinics — Europe (Netherlands, France, Belgium, Switzerland), North America (Canada); local-lab D-dimer assay |
| Reference standard | CT pulmonary angiogram (CTPA) at presentation in patients who did not satisfy the rule-out strategy, OR symptomatic VTE confirmed at 3-month clinical follow-up in patients managed without imaging — composite reference (the canonical management-outcome design for D-dimer rule-out) |
Eligibility criteria
Inclusion
- Adults with clinically suspected PE in an ED or outpatient setting
- Quantitative D-dimer as index test (one of fixed 500, age-adjusted, YEARS, or PEGeD cutoff)
- Wells / simplified Wells / Geneva pre-test stratification described
- Reference standard: CTPA at presentation OR symptomatic VTE at 3-month follow-up (composite)
- Reportable 2×2 against the composite reference (or back-computable Sens%/Spec% with N+/N−)
- Published 2006 or later (Christopher Study onwards)
Exclusion
- Pregnant / postpartum cohorts (different cutoff considerations)
- Active-cancer cohorts (high baseline D-dimer makes the test less informative)
- High-pretest-probability cohorts (D-dimer is not first-line; CTPA done directly)
- Methodology-only papers without 2×2 results
- Reviews / meta-analyses (cite primaries instead; van Es 2016 IPD-MA serves as the canonical k=6-cohort benchmark)
- Abstracts where the 2×2 cannot be back-computed
Pre-registration disclosure
This review was not prospectively registered with PROSPERO before search execution.
The build-time spec (docs/superpowers/specs/2026-04-27-rapidmeta-dta-engine-design.md) and implementation plan (docs/superpowers/plans/2026-04-27-rapidmeta-dta-implementation.md) provide internal versioning of the protocol, but these are not third-party temporal evidence.
For methods-paper submission this is disclosed as a non-prospective registration: the present review is positioned as a methods-engine demonstration (RapidMeta DTA engine using D-dimer for pulmonary embolism rule-out as a worked example), not a substantive clinical-evidence claim about D-dimer. For any clinical-evidence claim a retrospective PROSPERO record will be filed and linked here, and the canonical evidence base is the van Es 2016 Ann Intern Med individual-patient-data meta-analysis (k=6 cohorts, n=7,268 patients; PMID 27182696, DOI 10.7326/M16-0031) and the 2019 ESC PE Guidelines.
Acceptance gate (per spec §5)
| Gate | Threshold | Outcome (this review) |
|---|---|---|
| Fixed 500 cutoff, low/moderate Wells (default headline) | k ≥ 3 | k = 5 — meets threshold |
| All primary tier (mixed cutoff strategies) | k ≥ 4 | k = 5 — meets threshold |
| Combined all-tier (with sensitivity tier) | k ≥ 8 | k = 6 — coverage warning |
The combined tier falls below the k ≥ 8 ceiling so the engine surfaces a coverage banner per the fallback rule. This is consistent with the methods-paper framing — the review demonstrates the RapidMeta DTA engine on a small worked-example dataset and is not the substantive evidence base for D-dimer; the van Es 2016 Ann Intern Med IPD meta-analysis pools 6 cohorts and 7,268 patients (PMID 27182696, DOI 10.7326/M16-0031). The earlier headline cited as "van Es 2017 / k=11 / n=14,256" was a Crossref re-verification miss (2026-04-29 audit) — the closest verifiable IPD-MA matching the cited Wells + d-dimer rule-out design is van Es 2016 with k=6 / n=7,268.
Search Strategy
Search executed 2026-04-28 via automated retrieval against the ClinicalTrials.gov and PubMed APIs. Full provenance preserved in
ctgov_ddimer_pe_pack_2026-04-28.json and pubmed_ddimer_pe_abstracts_2026-04-28.json. Reviewers verify and finalize all included records.
CT.gov search
Condition: "pulmonary embolism"
Intervention: "D-dimer" OR "D dimer"
Date filter: study start ≥ 2010-01-01 (Christopher era onwards)
Results: 0 trials with posted 2×2 results panels Registries exist (NCT02365129 YEARS, NCT01134068 ADJUST-PE, NCT02384135 PEGeD, NCT01624285 ADJUST-validation) but DTA results panels are published in the paired peer-reviewed paper rather than posted to CT.gov
PubMed search
Primary query
("D-dimer" OR "D dimer") AND ("pulmonary embolism" OR "PE rule-out") AND ("sensitivity" AND "specificity") AND ("CTPA" OR "computed tomography pulmonary angiography" OR "follow-up")
Dates: 2014–2026 · Total: 412 abstracts Top by relevance: 18 fetched
Cutoff-strategy query
("age-adjusted D-dimer" OR "YEARS algorithm" OR "PEGED" OR "PEGeD") AND "pulmonary embolism" AND ("clinical probability" OR "Wells" OR "Geneva")
Dates: 2014–2026 · Total: 158 abstracts Top fetched: 12
Search execution
| Date | 2026-04-28 |
| Searcher / role | Automated initial retrieval; reviewers verify and finalize the included record set. |
| Databases | ClinicalTrials.gov (NIH/NLM API v2) + PubMed (NCBI E-utilities via claude.ai PubMed MCP) |
| Records identified | 0 (CT.gov panels) + 412 + 158 (PubMed) = 570 raw hits before de-duplication |
PRISMA-DTA flow diagram
Identification → screening → eligibility → inclusion. Counts derive from the search records above; categorised exclusions are detailed in the Screening tab.
· PERC-rule (different decision point upstream of D-dimer): Freund 2018 PROPER
· Pregnancy / cancer-cohort-only (different cutoff considerations — out of population)
PRISMA-NMA Flow Diagram
Screening
PRISMA-style flow from raw retrieval to extraction tier, plus an inclusions table and a categorised exclusions table.
PRISMA flow
Per-study screening decisions (Rayyan-style: full abstract inline)
Each card shows the title, authors, journal, full abstract text, decision badge (INCLUDED / EXCLUDED), rationale, and clickable links to the source record. Abstracts are rendered inline with no toggle so reviewers can scan quickly.
Included
Excluded
Data Extraction
Two-tier extraction with full provenance: CT.gov-linked publication tables for the primary tier, and regex-based PubMed-abstract back-compute for the sensitivity tier.
Methodology
- Single-tier provenance: patient-level management-outcome back-compute on PubMed abstracts that report Sens%/Spec% with PE prevalence and total N (Christopher 2006, ADJUST-PE 2014, YEARS 2017, ADJUST-validation 2019, PEGeD-Kearon 2019, Lucassen 2015). All 6 included rows use the same composite reference standard (CTPA at presentation OR symptomatic VTE at 3-month follow-up).
- Back-compute formula:
TP = round(Sens% × N_pos),FN = N_pos − TP,TN = round(Spec% × N_neg),FP = N_neg − TN. - Negation-context guard (per
lessons.md): each Sens%/Spec% match scanned ±30 chars for negation tokens (not, non, never, excluding); none of the 6 included rows triggered the guard. - FN=0 in management-outcome studies: PEGeD-Kearon reports zero 3-month VTE events among the 1,285 patients ruled-out by the strategy. The engine handles FN=0 cleanly via Clopper–Pearson (the upper Sens CI is not 100% — it is bounded above by the exact 1-sided 97.5% confidence limit, ~99.7% for n+ = 149). Caveat flagged in
data_caveats.
Per-study extraction (2×2 + Clopper–Pearson exact CIs)
Mirrors the Trials tab; values are recomputed at page-load by the same engine.
| Study | Year | n+ | n− | TP | FP | FN | TN | Sens (95% CI) | Spec (95% CI) | Source | Provenance |
|---|
Raw quote panels (back-computed sources)
Verbatim text from the abstract that the regex matched. Click a study to expand.
Inter-rater agreement
raw_quote provenance preserved for reviewer audit; click the DOI link in each row to inspect the source.
Detailed extraction (per-study)
Per-study extraction grid with study design, reference-standard specifics, clinical-probability stratification, cutoff strategy, assay type, population setting, prevalence setting, funding, and conflict-of-interest. All fields are editable when edit-mode is ON.
QUADAS-2 inline flags (brief; full grid in QUADAS-2 tab)
Brief per-study notes on patient selection, index test, reference standard, and flow & timing risk-of-bias domains. The full Whiting 2011 grid with signaling questions and editable judgments is in the QUADAS-2 tab.
| Study | QUADAS-2 inline flag |
|---|---|
| YEARS - van der Hulle 2017 | Low risk on patient selection (consecutive ED outpatients, 12 centres) / index test / reference standard. Flow & timing low: 3-month follow-up was protocol-mandated. Applicability concerns: YEARS strategy combines clinical criteria with conditional 500/1000 ng/mL cutoff — not a pure single-cutoff DTA. |
| ADJUST-PE - Righini 2014 | Low risk overall. Multinational 19-centre prospective cohort with central VIDAS lab. Reference standard composite (CTPA + 3-month follow-up) is the canonical management-outcome design. Applicability: age-adjusted cutoff arm extracted; non-high-pretest subset. |
| PEGeD - Kearon 2019 | Low risk; multicentre Canadian prospective management-outcome study. FN=0 in the strategy arm is honest (0/1,285 3-month VTE among PE-excluded) but yields Sens=100% which approaches the upper Clopper–Pearson bound; engine handles cleanly. |
| ADJUST-validation - van der Pol 2019 | Low risk; 8-centre European prospective management-outcome cohort. Wells + age-adjusted cutoff in non-high-pretest subset. Some concern on flow & timing for 4 FN distributed across the management arm. |
| Christopher Study - van Belle 2006 | Low risk for the era; 12-centre Dutch prospective management-outcome design (the canonical fixed-500 cohort). Applicability: predates the age-adjusted era so its Spec=41% is the lower bound on the cutoff_strategy axis. |
| Lucassen 2015 (sensitivity tier) | Some concern on patient selection (single-centre primary care + ED, smaller N=582). Lower-prevalence primary-care setting introduces an applicability gradient. Sensitivity-tier only. |
QUADAS-2 Risk of Bias & Applicability
Full Whiting et al. 2011 (Ann Intern Med 155:529) grid: four domains (Patient Selection, Index Test, Reference Standard, Flow & Timing) × signaling questions × Risk-of-Bias and Applicability-Concerns judgments. Pre-populated from the inline flags in the Extraction tab; reviewers override every cell via edit-mode. Applicability concerns are NOT applicable to Flow & Timing (Whiting 2011).
Summary table (traffic-light)
| Study | Patient Selection | Index Test | Reference Standard | Flow & Timing | |||
|---|---|---|---|---|---|---|---|
| RoB | App | RoB | App | RoB | App | RoB | |
Risk-of-bias traffic-light grid
Per Whiting 2011: 7 cells per study (Patient Selection RoB / App; Index Test RoB / App; Reference Standard RoB / App; Flow & Timing RoB — no Applicability for Flow & Timing). Click a cell to scroll to that study's detailed grid below. Stacked-bar chart shows the percentage breakdown across all 5 studies for each domain.
Per-study grid
GRADE-DTA evidence summary
Per Schunemann 2008 / 2020 GRADE for diagnostic test accuracy: each pooled outcome (Sensitivity, Specificity, DOR) is rated across 5 domains (Risk of bias, Inconsistency, Indirectness, Imprecision, Publication bias). Triggers auto-fill from the engine output, the QUADAS-2 panel, and Deeks' funnel test (see Heterogeneity tab). Reviewers may override any cell when edit-mode is ON; overrides are flagged with a manual marker and persist to localStorage.
Certainty floor 0; starts HIGH (4 = ⊕⊕⊕⊕) and subtracts the per-row downgrade. Levels: HIGH ⊕⊕⊕⊕ / MODERATE ⊕⊕⊕○ / LOW ⊕⊕○○ / VERY LOW ⊕○○○.
Pooled summary
Per-strategy headline (primary)
Each rule-out strategy is fit independently. For k=1 strategies (YEARS, PEGeD-graduated) the engine reports per-axis Clopper–Pearson exact CIs (single-study); for k=2 strategies (fixed-500, age-adjusted) the engine falls back to the FE bivariate. Use these per-strategy estimates for clinical interpretation.
Clinical safety: 3-month VTE failure rate
Random-effects pooled rate of symptomatic VTE within 3 months among patients ruled out by the strategy in the active tier (logit DerSimonian–Laird). The PASS/FAIL chip compares the one-sided 97.5% upper confidence limit against the 1.85% safety threshold (Schouten 2013 / van Es 2016 / 2019 ESC PE Guidelines).
Pooled bivariate (engine demonstration only)
⚠ Engine demonstration only. The k=5/k=6 pooled bivariate Sens/Spec below pools four distinct rule-out strategies (fixed-500, age-adjusted, YEARS-conditional, PEGeD-graduated) as if they were one diagnostic test (Macaskill 2010 Cochrane DTA Handbook §10.4 violation). Use the per-strategy cards above for clinical interpretation; the pooled estimate exists to demonstrate the engine.
Predictive values at chosen prevalence
The unique RapidMeta DTA hook — move the slider to see how PPV and NPV change with disease prevalence.
CI strategy: extreme-case bounds (Sens-CI × Spec-CI corners). Default 10% prevalence is a typical ED suspected-PE pre-test probability; adjust slider for your clinical setting (~5–10% Wells-low, ~10–20% Wells-moderate, ~40–60% Wells-high).
Tier selection
Three pre-computed tiers reflect different cutoff-strategy and applicability levels for the headline pool. Default is Fixed 500 cutoff, low/moderate Wells (k=3) — the methodologically cleanest subset (Christopher 2006 fixed-500 + YEARS-conditional + PEGeD-graduated), the canonical clinical-decision context. The Age-adjusted cutoff tier (k=5) adds ADJUST-PE 2014 and ADJUST-validation 2019 (the 10 × age-in-years cutoff arm for ≥50yo patients). The Combined all-tier tier (k=6) adds Lucassen 2015 single-centre validation as a sensitivity check (smaller N, primary-care setting).
Fagan nomogram — pre-test → post-test probability
Move the slider to translate a pre-test (clinical-suspicion) probability into a post-test probability using the chosen tier's pooled LR+ and LR−. The nomogram beneath shows the classical 3-axis construction: pre-test probability on the left, likelihood ratio in the middle, post-test probability on the right.
Slider is on a log scale (0.1% to 99.6%) so the nomogram axes are linear in log-odds. Default 10% reflects a typical ED suspected-PE pre-test; Wells-low cohorts run 5–15% and Wells-moderate cohorts 15–30%. High-Wells patients are excluded from D-dimer rule-out by design (CTPA done directly).
Included trials
Per-study Sensitivity and Specificity with Clopper–Pearson exact 95% confidence intervals. All 6 included rows use pubmed_abstract_back_computed (Sens%/Spec% and N+/N− from the published abstracts of Christopher 2006, ADJUST-PE 2014, YEARS 2017, ADJUST-validation 2019, PEGeD-Kearon 2019, and Lucassen 2015).
| Study | Year | Country | n+ | n− | TP | FP | FN | TN | Sens (95% CI) | Spec (95% CI) | Source | Provenance |
|---|
Forest plot
Paired forest: per-study Sensitivity (left) and Specificity (right) with 95% Clopper–Pearson exact CIs. The pooled summary row (blue) uses the bivariate model estimate for the active tier.
SROC space
Summary ROC plot in (1−Specificity, Sensitivity) coordinates. Grey dots = per-study estimates; red dashed ellipse = 95% confidence region from the bivariate fit; blue dashed curve = HSROC reparameterisation (Harbord 2007); large red dot = pooled summary point.
Heterogeneity
Bivariate variance components on the logit scale, threshold-effect diagnostic, and convergence audit. Note: I² is not directly defined for the bivariate DTA model — tau² on logit-Sens / logit-Spec is the canonical between-study heterogeneity metric (Reitsma 2005).
Deeks' funnel-asymmetry test
Deeks 2005: regress ln(DOR) on 1/√ESS where ESS = effective sample size = 4·n+ ·n− / (n+ + n−). The slope tests funnel asymmetry; p < 0.10 suggests possible publication bias. Skipped when k < 5.
Sensitivity (leave-one-out)
For each study i in the active tier, refit the bivariate model with study i excluded and record pooled Sens, Spec, DOR. The largest mover quantifies which single study most influences the headline summary. When removing a study takes the remaining pool to 2≤k<5 the engine falls back to the FE bivariate; when it leaves k=1 the engine reports per-axis Clopper–Pearson (single_study) instead of a pooled estimate.
Subgroups
Filter the combined-tier study pool (k=6) by pre-specified covariates and refit. Minimum k=3 required for a refit; below this threshold the subgroup is reported descriptively only.
Methods
Substantive comparators (this review defers to)
This k=5 review uses D-dimer / pulmonary-embolism rule-out as a worked methodological example for the RapidMeta DTA engine; the substantive evidence base is Crawford et al. 2016 (Cochrane DTA review of D-dimer for excluding PE, ~40 D-dimer/PE primary studies, CD010864; DOI: 10.1002/14651858.CD010864.pub2) and Stals et al. 2022 (individual-patient-data meta-analysis across 14 prospective cohorts, n=14,896 patients; Ann Intern Med 175:244–255; DOI: 10.7326/M21-2625). The earlier van Es 2016 Wells-and-D-dimer IPD-MA (k=6 cohorts, n=7,268; Ann Intern Med 165:253–261; PMID 27182696, DOI 10.7326/M16-0031) provides the post-rule-out 3-month VTE safety reference (one-sided 97.5% UCL <1.85%; threshold from Schouten 2013). Quantitative interpretation should defer to those references; the present review's contribution is methodological transparency, edit-mode auditability, and engine validation.
Reference standard — differential verification bias
Differential verification bias. All five primary-tier studies (and the Lucassen 2015 sensitivity-tier study) use a composite reference standard: CTPA at presentation for D-dimer-positive patients (or those whose pre-test probability or rule-output requires imaging) OR 3-month clinical follow-up for VTE outcomes among D-dimer-negative patients managed without imaging. Per QUADAS-2 (Whiting 2011), this constitutes differential verification — the reference standard differs by index-test result — and Q3.2 ("Were the reference standard results interpreted without knowledge of the index test?") must therefore be answered "No". Following Macaskill 2010 (Cochrane DTA Handbook §10.4), this design is recognised as the canonical management-outcome approach for rule-out tests but does inflate apparent sensitivity and is recorded as High Risk of Bias on the Reference Standard QUADAS-2 domain across all six studies. The composite is also imperfect: CTPA misses subsegmental PE (~83% sensitivity vs digital subtraction angiography); 3-month follow-up misses asymptomatic non-fatal PE. Bayesian latent-class meta-analysis (Dendukuri 2012, Biometrics 68:1285) is the formal handling for imperfect references and is deferred to v1.2.
Continuity-correction policy
Conditional 0.5 added to all four cells of every study, and only when at least one cell of any study in the pool is zero. Unconditional correction biases the diagnostic odds ratio toward the null (Sweeting 2004) and is therefore not used here.
Continuity-correction bias-toward-null caveat (B6). When PEGeD−Kearon's FN=0 fires the zero-cell trigger, the engine adds 0.5 to all four cells in every study in the pool. Because the pooled Sensitivity is structurally near 100% (a property of management-outcome rule-out designs), the logit-Sens estimator is pulled toward the logit-of-0.5 anchor (i.e. zero on the logit scale = 50% on the natural scale), biasing the pooled Sens estimate downward by approximately 1–2 percentage points. Per-study Clopper–Pearson exact CIs in the Trials tab are continuity-correction-free and may be a more accurate per-study Sens estimate. The cell-only correction recommended by Sweeting 2004 for sparse-data binary meta-analysis is preferable in principle but is deferred until the engine adds a per-cell correction mode in v1.1; the all-studies correction here is the conservative default and is reported transparently.
GRADE-DTA notes
Certainty assessment is informed by (a) study limitations — QUADAS-2 is tabulated in the QUADAS-2 tab; auto-assessed RoB judgments feed the GRADE Risk of Bias domain. Reviewer overrides in edit mode persist via localStorage and are included in JSON export; (b) inconsistency — visible heterogeneity (population, specimen, reference standard) prompts downgrading; (c) indirectness — reference-standard variability across tiers is the principal source; (d) imprecision — per Schunemann 2020 GRADE-DTA, the imprecision rating uses an editable clinical decision threshold (set per outcome in the GRADE tab); when no threshold is supplied the engine falls back to a heuristic CI-width assessment, prefixed "[Heuristic]" in the rationale. For DOR, the CI width is computed on the log scale (log(DOR_ub / DOR_lb)); (e) publication bias — not formally assessed (Deeks' funnel plot requires k≥10). Tier divergence >5pp on Sens or Spec is surfaced as a headline banner.
R cross-validation (mada)
Build-time validation log
r_validation_log.json not yet generated (T18)
Provenance audit
Per-study source & verbatim raw quote for back-computed estimates.
| Study | Provenance | Source link | Raw quote (back-computed only) |
|---|
References
Vancouver-style citations for included studies plus methodological references for the engine and validation package. PubMed (PMID), DOI, and ClinicalTrials.gov (NCT) links are provided where available.
Scientific Output
Auto-generated manuscript text rendered from the live engine fit. Each section has a copy-to-clipboard button. Edit-mode changes (e.g. TP/FP/FN/TN values) re-render the Results paragraph automatically.