Verification bias

O’Sullivan, Jack W; Banerjee, Amitava; Heneghan, Carl; Plüddemann, Annette

doi:10.1136/bmjebm-2018-110919

Cited by 35 publications

(40 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Partial verification bias might affect analyses of primary outcomes, 39 because reference data on the HIV status of deceased siblings are only available for a subset of the deaths reported during survey interviews. This usually leads to sensitivity estimates that are too high and specificity estimates that are too low 40 .…”

Section: Methodsmentioning

confidence: 99%

Collecting data on HIV-related mortality during household surveys: a randomized validation study in Malawi

Amoah

Brumfield

Crampin

et al. 2020

Preprint

View full text Add to dashboard Cite

Background: In many African countries, there are limited representative data on HIV/AIDS mortality. We tested whether such data could be collected during household surveys periodically conducted in most African countries. Methods: We added HIV questions to the module on adult and maternal mortality used in Demographic and Health Surveys. We conducted a validation study of the data generated by these questions in northern Malawi. We randomly assigned men and women aged 15-59 years old to a face-to-face interview or audio computer-assisted self-interviewing (ACASI). We compared survey reports of adult deaths to prospective data on mortality and HIV collected by the Karonga Health and Demographic Surveillance Site. We calculated the sensitivity and specificity of survey data in recording the HIV status of deceased siblings of respondents. We adjusted for partial verification bias using multiple imputations. Results: We interviewed 535 participants, who reported 885 deaths at ages 15 and older. The added HIV questions yielded largely complete data on the HIV status of deceased siblings of respondents, particularly those who died recently. The adjusted sensitivity of survey data on HIV status of the deceased was high in both study groups (0.78-0.82). There were few false positive reports of the HIV status of deceased siblings (specificity = 0.96-0.98). ACASI did not improve the accuracy of survey data, but it required more time to collect mortality reports. Asking the HIV questions only took 0.4 minute (approximately 25 seconds) per deceased sibling in face-to-face interviews. Conclusions: Adding HIV questions to mortality questionnaires used in household surveys yields accurate data on the HIV status of deceased adults. These new data could aid in tracking progress towards global HIV elimination targets.

show abstract

Section: Methodsmentioning

confidence: 99%

Collecting data on HIV-related mortality during household surveys: a randomized validation study in Malawi

Amoah

Brumfield

Crampin

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The data source was the Corporate Data Warehouse (CDW), a national repository of data from the VHA medical record system (VistA) and other VHA clinical and administrative systems 2 . Data were housed and analyzed within the Veterans Affairs Informatics and Computing Infrastructure (VINCI) 25 . This research was conducted in compliance with the Helsinki Declaration, with the approval of the University of Utah Institutional Review Board (IRB_00052363).…”

Section: Methodsmentioning

confidence: 99%

“…Inverse Probability Weighting (IPW) for estimating algorithm performance in the general population. The performance estimates in axSpA-enriched populations are subject to partial verification bias 25 . Verification bias occurs when the reference standard has not been carried out in all patients because of ethical or practical reasons 26 (i.e.…”

Section: Methodsmentioning

confidence: 99%

Identifying Patients With Axial Spondyloarthritis in Large Datasets: Expanding Possibilities for Observational Research

et al. 2020

View full text Add to dashboard Cite

Objective Observational research of axial spondyloarthritis (axSpA) is limited by a lack of methods for identifying diverse axSpA phenotypes in large datasets. AxSpA identification algorithms were previously designed to identify a broad spectrum of axSpA patients, including patients not identifiable with diagnosis codes. The study objective was to estimate the performance of axSpA identification methods in the general Veterans Affairs (VA) population. Methods A patient sample with known axSpA status (n=300) was established with chart review. For feasibility, this sample was enriched with Veterans with axSpA risk factors. Algorithm performance outcomes included sensitivities, positive predictive values (PPV), and F1 scores (an overall performance metric combining sensitivity and PPV). Performance was estimated with unweighted outcomes for the axSpA-enriched sample and inverse probability weighted (IPW) outcomes for the general VA population. These outcomes were also assessed for traditional identification methods using diagnosis codes for the ankylosing spondylitis (AS) subtype of axSpA. Results The mean age was 54.7 and 92% were male. Unweighted F1s (0.59-0.74) were higher than IPW F1s (0.48-0.65). The Full Algorithm had the best overall performance (F1IPW 0.65). The Early Algorithm was the most inclusive (sensitivityIPW 0.90, PPVIPW 0.38). The traditional method using ≥2 AS diagnosis codes from rheumatology had the highest PPV (PPVIPW 0.84, sensitivityIPW 0.34). Conclusion The axSpA identification methods demonstrated a range of performance attributes in the general VA population that may be appropriate for various types of studies. The novel identification algorithms may expand the scope of research, by enabling identification of more diverse axSpA populations.

show abstract

“…the "test" whether they would survive in case of transportation. 43 If all patients had been transported, it is possible that there would have been a few that would have survived.…”

Section: Diagnostic Performancementioning

confidence: 99%

Diagnostic performance of the basic and advanced life support termination of resuscitation rules: A systematic review and diagnostic meta-analysis

et al. 2020

View full text Add to dashboard Cite

Aim: To minimize termination of resuscitation (TOR) in potential survivors, the desired positive predictive value (PPV) for mortality and specificity of universal TOR-rules are !99%. In lack of a quantitative summary of the collective evidence, we performed a diagnostic meta-analysis to provide an overall estimate of the performance of the basic and advanced life support (BLS and ALS) termination rules. Data sources:We searched PubMed/EMBASE/Web-of-Science/CINAHL and Cochrane (until September 2019) for studies on either or both TORrules in non-traumatic, adult cardiac arrest. PRISMA-DTA-guidelines were followed.Results: There were 19 studies: 16 reported on the BLS-rule (205.073 patients, TOR-advice in 57%), 11 on the ALS-rule (161.850 patients, TOR-advice in 24%). Pooled specificities were 0.95 (0.89À0.98) and 0.98 (0.95À1.00) respectively, with a PPV of 0.99 (0.99À1.00) and 1.00 (0.99À1.00).Specificities were significantly lower in non-Western than Western regions: 0.84 (0.73À0.92) vs. 0.99 (0.97À0.99), p < 0.001 for the BLS rule. For the ALS-rule, specificities were 0.94 (0.87À0.97) vs. 1.00 (0.99À1.00), p < 0.001. For non-Western regions, 16 (BLS) or 6 (ALS) out of 100 potential survivors met the TOR-criteria. Meta-regression demonstrated decreasing performance in settings with lower rates of in-field shocks.Conclusions: Despite an overall high PPV, this meta-analysis highlights a clinically important variation in diagnostic performance of the BLS and ALS TOR-rules. Lower specificity and PPV were seen in non-Western regions, and populations with lower rates of in-field defibrillation. Improved insight in the varying diagnostic performance is highly needed, and local validation of the rules is warranted to prevent in-field termination of potential survivors.

show abstract

Verification bias

Cited by 35 publications

References 11 publications

Collecting data on HIV-related mortality during household surveys: a randomized validation study in Malawi

Collecting data on HIV-related mortality during household surveys: a randomized validation study in Malawi

Identifying Patients With Axial Spondyloarthritis in Large Datasets: Expanding Possibilities for Observational Research

Diagnostic performance of the basic and advanced life support termination of resuscitation rules: A systematic review and diagnostic meta-analysis

Contact Info

Product

Resources

About