Clinical Adjudication Research Articles

COMPARISON BETWEEN CARDIOVASCULAR TRIAL OUTCOMES AND TREATMENT EFFECTS USING CLINICAL ENDPOINT COMMITTEE ADJUDICATION VERSUS ROUTINE HEALTH DATA: A SYSTEMATIC REVIEW

Registry based randomized controlled trials (RRCTs) allow for outcome ascertainment using routine health data (RHD). While this method offers a potential solution to the rising cost and complexity of traditional RCTs, a rigorous quantitative analysis between sources of RHD has yet to be systematically performed to compare results to traditional outcome ascertainment by clinical endpoint committee (CEC) adjudication. Among cardiovascular (CV) trials, we set out to compare the incidence of CV events and estimated randomized treatment effects ascertained from RHD versus traditional clinical evaluation and adjudication.

P2780Accuracy of electronic healthcare records for diagnosis of stroke types in a large community-based prospective cohort study in China

Abstract Background and purpose Reliable assessment of the determinants of major pathological types of stroke is dependent on the accuracy of diagnosis in population-based studies. We evaluated the diagnostic accuracy of stroke types in a large community-based cohort study involving health records collected in China. Methods In 2004–08, >0.5 million adults aged 30–79 years were recruited from general populations of 10 diverse areas (5 urban, 5 rural) in China. During an approximate 7-year follow-up, 37,694 stroke cases had been reported by linkage to electronic health records from disease-specific and mortality registries and from national health insurance agencies. For all reported stroke cases, hospital medical records were retrieved systematically and relevant data extracted for subsequent adjudication by specialists using bespoke electronic platforms. Results Among all reported incident stroke cases, 80% were ischaemic stroke (IS), 17% were intracerebral haemorrhage (ICH), and 1% were subarachnoid haemorrhage (SAH). To date, medical records have been retrieved for 29,632 cases, with reports of stroke diagnosis verified by public health staff in 27,115 (92%) cases, of which 3,778 (14%) were secondary diagnoses. Evidence of neuroimaging was found in 92% of all verified stroke cases. Of 23,337 primary stroke cases sent for specialist adjudication, a diagnosis of pathological stroke type was confirmed in 19,718 cases with the positive predictive values being 82.4% (95% confidence interval [CI], 82.0% - 82.8%) for IS, 97.8% (97.6 - 97.9) for ICH, and 98.2% (98.1 - 98.3) for SAH. Overall, the strength of association of systolic blood pressure was over 6-fold greater for confirmed than non-confirmed stroke cases and was nearly 3-fold greater for confirmed ICH than IS cases. Diagnostic accuracy by stroke type Reported Retrieved Verified Adjudicated Confirmed Approximate 95% CI n n % n % n % n PPV, % IS 30,143 25,477 85 23,551 92 20,045 85 16,515 82.4 (82.0–82.8) ICH 6,484 3,486 54 3,079 88 2,919 95 2,854 97.8 (97.6–97.9) SAH 557 397 71 326 82 279 86 274 98.2 (98.1–98.3) Other 510 272 53 159 59 94 59 75 79.8 (79.4–80.2) IS indicates ischaemic stroke; ICH, intracerebral haemorrhage; SAH, subarachnoid haemorrhage; and PPV, positive predictive value. Conclusions The overall diagnostic accuracy of pathological stroke types obtained from hospital records in China is comparable to reports from Western populations. Despite advances in electronic healthcare records, reliable classification of stroke types requires clinical adjudication using additional relevant investigations. Acknowledgement/Funding Kadoorie Charitable Foundation, UK Wellcome Trust & National Natural Science Foundation and National Key Research and Development Program of China

Interactive Machine Learning for Laboratory Data Integration.

Laboratory data collected in the electronic health record as part of routine care can be used in secondary research. For example, the US Department of Veterans Affairs maintains a data warehouse covering over 20 million individuals and 6.6 billion lab tests. However, data aggregation in such a data warehouse can be difficult. In order to retrieve all or nearly all of one type of lab result with a high degree of precision, we perform clinical concept adjudication, which is the process of an expert determining which database records correspond to a target clinical concept. In this work, we develop an interactive machine learning tool to "extend the reach" of expert laboratory test adjudicators. Our tool provides access to automatic laboratory classification in a user-facing front end that covers all steps in an adjudication workflow, in order to lower barriers to collaboration, increase transparency of adjudication, and to promote efficiencies and data reuse.

Safety and efficacy of inactivated varicella zoster virus vaccine in immunocompromised patients with malignancies: a two-arm, randomised, double-blind, phase 3 trial

Patients who are immunocompromised because of malignancy have an increased risk of herpes zoster and herpes zoster-related complications. We aimed to investigate the efficacy and safety of an inactivated varicella zoster virus (VZV) vaccine for herpes zoster prevention in patients with solid tumour or haematological malignancies. This phase 3, two-arm, randomised, double-blind, placebo-controlled, multicentre trial with an adaptive design was done in 329 centres across 40 countries. The trial included adult patients with solid tumour malignancies receiving chemotherapy and those with haematological malignancies, either receiving or not receiving chemotherapy. Patients were randomly assigned (1:1) to receive four doses of VZV vaccine inactivated by γ irradiation or placebo approximately 30 days apart. The patients, investigators, trial site staff, clinical adjudication committee, and sponsor's clinical and laboratory personnel were masked to the group assignment. The primary efficacy endpoint was herpes zoster incidence in patients with solid tumour malignancies receiving chemotherapy, which was assessed in the modified intention-to-treat population (defined as all randomly assigned patients who received at least one dose of inactivated VZV vaccine or placebo). The primary safety endpoint was serious adverse events up to 28 days after the fourth dose in patients with solid tumour malignancies receiving chemotherapy. Safety endpoints were assessed in all patients who received at least one dose of inactivated VZV vaccine or placebo and had follow-up data. This trial is registered (NCT01254630 and EudraCT 2010-023156-89). Between June 27, 2011, and April 11, 2017, 5286 patients were randomly assigned to receive VZV vaccine inactivated by γ irradiation (n=2637) or placebo (n=2649). The haematological malignancy arm was terminated early because of evidence of futility at a planned interim analysis; therefore, all prespecified haematological malignancy endpoints were deemed exploratory. In patients with solid tumour malignancies in the modified intention-to-treat population, confirmed herpes zoster occurred in 22 of 1328 (6·7 per 1000 person-years) VZV vaccine recipients and in 61 of 1350 (18·5 per 1000 person-years) placebo recipients. Estimated vaccine efficacy against herpes zoster in patients with solid tumour malignancies was 63·6% (97·5% CI 36·4 to 79·1), meeting the prespecified success criterion. In patients with solid tumour malignancies, serious adverse events were similar in frequency across treatment groups, occurring in 298 (22·5%) of 1322 patients who received the vaccine and in 283 (21·0%) of 1346 patients who received placebo (risk difference 1·5%, 95% CI -1·7 to 4·6). Vaccine-related serious adverse events were less than 1% in each treatment group. Vaccine-related injection-site reactions were more common in the vaccine group than in the placebo group. In the haematological malignancy group, VZV vaccine was well tolerated and estimated vaccine efficacy against herpes zoster was 16·8% (95% CI -17·8 to 41·3). The inactivated VZV vaccine was well tolerated and efficacious for herpes zoster prevention in patients with solid tumour malignancies receiving chemotherapy, but was not efficacious for herpes zoster prevention in patients with haematological malignancies. Merck & Co, Inc.

Results from a prospective multicenter feasibility study of Zenith p-Branch stent graft

ObjectiveTo evaluate the Cook Zenith p-Branch stent graft (William A. Cook Australia, Brisbane, Queensland, Australia) for the treatment of patients with asymptomatic juxtarenal or pararenal abdominal aortic aneurysms in a prospective, nonrandomized, multicenter, feasibility study conducted in the United States. MethodsThe off-the-shelf p-Branch endograft incorporates a celiac scallop, superior mesenteric artery (SMA) fenestration, and two conical-shaped pivot renal fenestrations. Patients were eligible if the aneurysm did not extend above the distal margin of the SMA and the anatomy of renal and visceral arteries were compatible with one of the two p-Branch configurations (configuration A, renal fenestrations at the same longitudinal position; configuration B, offset renal fenestration). ResultsA total of 30 patients (93% men; mean age, 73 years; mean aneurysm diameter, 64.7 ± 11.4 mm) were enrolled from January 2013 to June 2015. The mean follow-up period was 28.7 ± 12.5 months (as of October 2017). The technical success rate was 93% (28/30), with technical failure occurring in the first two study cases (device deliberately repositioned below the renal arteries because of difficulty in cannulating the renal arteries in one and an inability to place a renal stent in the other; both patients survived the procedure and withdrew from the study before the 12-month follow-up visit). No additional technical failures occurred after these first two cases, when an updated physician training and proctoring program was implemented. No 30-day mortality occurred. Of the remaining 28 patients, three died during the follow-up period; none within 30 days and none related to the device or procedure as determined by Clinical Events Committee adjudication. No treated aneurysm ruptures, conversion, or core laboratory-reported migration, or stent fracture were reported. The primary renal artery patency (interval to occlusion or reintervention for stenosis) rate was 88.6% ± 4.4% at 1 and 2 years. Two patients (7%) developed renal insufficiency secondary to a right renal artery stenosis in one patient and progression of chronic renal insufficiency in one patient. No patient required dialysis or developed mesenteric ischemia. Of the 28 patients, nine (32%) had undergone 13 secondary interventions for renal artery occlusion (n = 2), SMA occlusion (n = 1), renal or SMA stenosis (5 interventions in 4 patients), renal stent kinking (n = 1), lower extremity claudication (n = 2 in the same patient), type III endoleak between the p-Branch and renal stent (n = 1), and type II endoleak (n = 1). ConclusionsThese early and intermediate results, which incorporated physician learning curves, support the safety and feasibility of the off-the-shelf Zenith p-Branch device. Follow-up examinations through 5 years will continue to assess the long-term results.

A host gene expression approach for identifying triggers of asthma exacerbations.

RationaleAsthma exacerbations often occur due to infectious triggers, but determining whether infection is present and whether it is bacterial or viral remains clinically challenging. A diagnostic strategy that clarifies these uncertainties could enable personalized asthma treatment and mitigate antibiotic overuse.ObjectivesTo explore the performance of validated peripheral blood gene expression signatures in discriminating bacterial, viral, and noninfectious triggers in subjects with asthma exacerbations.MethodsSubjects with suspected asthma exacerbations of various etiologies were retrospectively selected for peripheral blood gene expression analysis from a pool of subjects previously enrolled in emergency departments with acute respiratory illness. RT-PCR quantified 87 gene targets, selected from microarray-based studies, followed by logistic regression modeling to define bacterial, viral, or noninfectious class. The model-predicted class was compared to clinical adjudication and procalcitonin.ResultsOf 46 subjects enrolled, 7 were clinically adjudicated as bacterial, 18 as viral, and 21 as noninfectious. Model prediction was congruent with clinical adjudication in 15/18 viral and 13/21 noninfectious cases, but only 1/7 bacterial cases. None of the adjudicated bacterial cases had confirmatory microbiology; the precise etiology in this group was uncertain. Procalcitonin classified only one subject in the cohort as bacterial. 47.8% of subjects received antibiotics.ConclusionsOur model classified asthma exacerbations by the underlying bacterial, viral, and noninfectious host response. Compared to clinical adjudication, the majority of discordances occurred in the bacterial group, due to either imperfect adjudication or model misclassification. Bacterial infection was identified infrequently by all classification schemes, but nearly half of subjects were prescribed antibiotics. A gene expression-based approach may offer useful diagnostic information in this population and guide appropriate antibiotic use.

Comparison of subsequent injury categorisation (SIC) models and their application in a sporting population

BackgroundThe original subsequent injury categorisation (SIC-1.0) model aimed to classify relationships between chronological injury sequences to provide insight into the complexity and causation of subsequent injury occurrence. An updated model has recently been published. Comparison of the data coded according to the original and revised subsequent injury categorisation (SIC-1.0 and SIC-2.0) models has yet been formally compared.MethodsMedical attention injury data was prospectively collected for 42 elite water polo players over an 8 month surveillance period. The SIC-1.0 and SIC-2.0 models were retrospectively applied to the injury data. The injury categorisation from the two models was compared using descriptive statistics.ResultsSeventy-four injuries were sustained by the 42 players (median = 2, range = 0–5), of which 32 injuries (43.2%) occurred subsequent to a previous injury. The majority of subsequent injuries were coded as occurring at a different site and being of a different nature, while also being considered clinically unrelated to the previous injury (SIC-1.0 category 10 = 57.9%; SIC-2.0 clinical category 16 = 54.4%). Application of the SIC-2.0 model resulted in a greater distribution of category allocation compared to the SIC-1.0 model that reflects a greater precision in the SIC-2.0 model.ConclusionsSubsequent injury categorisation of sport injury data can be undertaken using either the original (SIC-1.0) or the revised (SIC-2.0) model to obtain similar results. However, the SIC-2.0 model offers the ability to identify a larger number of mutually exclusive categories, while not relying on clinical adjudication for category allocation. The increased precision of SIC-2.0 is advantageous for clinical application and consideration of injury relationships.

Rationale and design of a prospective substudy of clinical endpoint adjudication processes within an investigator-reported randomised controlled trial in patients with coronary artery disease: the GLOBAL LEADERS Adjudication Sub-StudY (GLASSY)

IntroductionThe GLOBAL LEADERS is an open-label, pragmatic and superiority randomised controlled trial designed to challenge the current treatment paradigm of dual antiplatelet therapy (DAPT) for 12 months followed by aspirin...

Identifying dementia outcomes in UK Biobank: a validation study of primary care, hospital admissions and mortality data

Prospective, population-based studies that recruit participants in mid-life are valuable resources for dementia research. Follow-up in these studies is often through linkage to routinely-collected healthcare datasets. We investigated the accuracy of these datasets for dementia case ascertainment in a validation study using data from UK Biobank—an open access, population-based study of > 500,000 adults aged 40–69 years at recruitment in 2006–2010. From 17,198 UK Biobank participants recruited in Edinburgh, we identified those with ≥ 1 dementia code in their linked primary care, hospital admissions or mortality data and compared their coded diagnoses to clinical expert adjudication of their full-text medical record. We calculated the positive predictive value (PPV, the proportion of cases identified that were true positives) for all-cause dementia, Alzheimer’s disease and vascular dementia for each dataset alone and in combination, and explored algorithmic code combinations to improve PPV. Among 120 participants, PPVs for all-cause dementia were 86.8%, 87.3% and 80.0% for primary care, hospital admissions and mortality data respectively and 82.5% across all datasets. We identified three algorithms that balanced a high PPV with reasonable case ascertainment. For Alzheimer’s disease, PPVs were 74.1% for primary care, 68.2% for hospital admissions, 50.0% for mortality data and 71.4% in combination. PPV for vascular dementia was 43.8% across all sources. UK routinely-collected healthcare data can be used to identify all-cause dementia in prospective studies. PPVs for Alzheimer’s disease and vascular dementia are lower. Further research is required to explore the geographic generalisability of these findings.

Uppsala Clinical Research Center—development of a platform to promote national and international clinical science

Uppsala Clinical Research Center (UCR) is a non-profit organization that provides service for clinical research aiming for development and improvement of health care in Sweden and worldwide. UCR was started in 2001 with the ambition to shift the focus of clinical research from new medications or devices launched by the industry to problem-based research on issues identified in clinical reality, for example through the national quality registries. In order to accomplish these goals, UCR has established services in: 1) clinical trials of new and old methods in health care; 2) quality development of the health care system supported by internet-based national quality registries; 3) biostatistics, epidemiology, and data management; 4) biobanking of biological materials (Uppsala Biobank); 5) high-throughput biochemical analyses (UCR laboratory); and 6) academic leadership by the members of the UCR research faculty. The UCR clinical trials group provides services for investigator-driven projects in all areas of health care, for global mega-trials on new pharmaceutical treatments and devices, for biobanking including biomarker and genetics analyses, and for clinical events adjudication in national as well as global mega-trials. During the last few years, UCR has been a pioneer in establishing the registry-based randomized clinical trial (R-RCT), which today is an international model on how to perform cost-effective pragmatic randomized trials in the real-world environment. In 2002, UCR started the first national competence center for national quality registries, which pioneered the development of the current internet-based technologies for registering, reporting, and supporting continuous systematic improvement of health care. UCR is currently harboring around 20 national quality registries in all areas of health care. Today, UCR is the leading European center for registry-based quality development and evaluation of new medical treatments in cardiovascular care and has started to support other European countries in implementing the UCR registry platform in order to improve quality of care in the European Union.

2012. FilmArray® Measurement of Host Response Signatures RapidLy Discriminates Viral, Bacterial, and Non-infectious Etiologies of Illness

BackgroundThe inability to reliably discriminate bacterial, viral, and non-infectious illness has led to an epidemic of antibiotic overuse and rising rates of antimicrobial resistance. Host gene expression provides a powerful approach to distinguish infection etiologies and guide appropriate therapy. However, existing platforms for transcriptomic analysis are not amenable for clinical application. The FilmArray platform is a sample-to-answer multiplex RT-PCR system that automates transcriptomic analysis including sample preparation and data analysis. Here we report the validation of a host gene expression FilmArray test to discriminate bacterial, viral, and non-infectious etiologies of illness.MethodsResearch use only (RUO) FilmArray pouches were manufactured with 45 host response assays previously shown to discriminate bacterial, viral, and non-infectious illness. These pouches were tested on whole blood samples from 226 patients with acute respiratory illness (ARI). Using clinical adjudication as the reference standard, there were 52 bacterial, 100 viral, and 75 non-infectious cases. Quantification cycles were recorded for each assay and normalized to an internal control. A logistic regression model generated probabilities of each condition, which were used to classify subjects.ResultsBeginning with 100 µL of blood, the FilmArray host response panel provided results in ~45 minutes from sample to answer. Overall accuracy for bacterial ARI relative to clinical adjudication was 85% with an area under the receiver operating characteristic curve (AUC) of 0.92. Accuracy and AUC for viral infection were 85% and 0.91, respectively. Ill patients without infection were correctly identified 86% of the time with an AUC of 0.88.ConclusionThese results show that the FilmArray system can rapidLy measure host gene expression to accurately discriminate bacterial, viral, and non-infectious illness. The development of such a system creates a new option to mitigate inappropriate antibiotic use. It also presents opportunities to use host gene expression as a diagnostic modality for a variety of disease states.Disclosures E. L. Tsalik, Host Response, Inc.: Founder, Equity. J. Montgomery, BioFire Diagnostics: Employee, Salary. J. Nawrocki, BioFire Diagnostics, LLC.: Employee, Salary. M. Deneris, BioFire Diagnostics, LLC.: Employee, Salary. C. Gritzen, BioFire Diagnostics, LLC.: Employee, Salary. J. Jones, BioFire Diagnostics, LLC.: Employee, Salary. R. Crisp, BioFire Diagnostics, LLC: Employee, Salary. G. S. Ginsburg, Host Response Inc: Board Member, Founder, Scientific Advisor and Shareholder, Stock (currently worth < $100). A. Hemmert, BioFire Diagnostics, LLC.: Employee and Investigator, Salary. C. W. Woods, Host Response, Inc.: Founder, Equity.

2355. Clinically vs. Serologically Identified Varicella: A Hidden Infection Burden. Lessons of 10-Year Follow-up in Varicella Endemic Countries

BackgroundA randomized, controlled, long-term varicella vaccine efficacy study (NCT00226499) was conducted in several European countries with no universal varicella vaccination. The control group allows studying the immunological response profiles to natural varicella exposure and disease in children from 1 year of age onwards.MethodsOf 5,803 subjects enrolled in the study, 744 were vaccinated with 2 doses of measles-mumps-rubella vaccine as active control. Follow-up lasted for up to 10 years. Varicella case ascertainment was done by combining PCR testing of skin lesions, epidemiological context and blinded clinical case adjudication by independent experts. A confirmed case required a positive PCR result or a positive epidemiological context with a positive case adjudication.ResultsIn the control group, a total of 352 confirmed varicella episodes were captured; of which, 339 had an available serum sample taken before and after the episode at any of the year 1 or bi-yearly visits scheduled per protocol. All subjects were seronegative for anti-VZV antibodies before vaccination. The immunological profile showed that 96% of subjects with varicella episodes experienced an 8-fold increase of the anti-VZV titers when comparing the sample available after the episode vs. the one before the episode. The data indicate that the GMC levels were similar to that induced by 2-dose immunization with varicella vaccines, and persisted over many years after the varicella episode without indication of waning (figure). In subjects without any reported varicella episode, 8-fold increases of the anti-VZV titers between 2 successive blood samples were measured in up to 34% of consecutive pairs of serum samples. ConclusionTo our knowledge, this is the first long-term analysis of the immunological history of anti-varicella immunological profile in young children exposed to a high varicella force of infection. About one-third of subjects not vaccinated against varicella developed an anti-VZV immune response although no varicella disease was reported. Sub-clinical varicella may occur more frequently than anticipated. The total incidence of varicella infections might be under-estimated by syndromic surveys only. Funding: GlaxoSmithKline Biologicals SA.Disclosures P. Gillard, GSK: Employee and Shareholder, Salary. M. Povey, GSK: Employee, Salary. S. Carryn, GSK: Employee and Shareholder, Restricted shares, and Salary.

When do we need clinical endpoint adjudication in clinical trials?

Clinical endpoint adjudication (CEA) is a standardized process for assessment of safety and efficacy of pharmacologic or device therapies in clinical trials. CEA plays a key role in many large clinical trials with the aim of achieving consistency and accuracy of the study results, by applying independent and blinded evaluation of suspected clinical events reported by investigators. However, due to high costs there are different opinions regarding the use of central adjudication versus more simplified strategies or site-based assessments and whether the final results differ significantly. There is a lack of scientific evaluation of different adjudication strategies, and more knowledge is needed on the optimal adjudication process and how to achieve the best cost-effectiveness. New methodologies using national registry data and artificial intelligence may challenge the traditional adjudication strategy and could potentially reduce cost considerably with a similar result. Further research and evidence in this field of clinical trials methodology are essential.

Using the Kinetic Estimating Glomerular Filtration Rate Equation for Estimating Glomerular Filtration Rate and Detecting Acute Kidney Injury: A Pilot Study

Background: Estimating the glomerular filtration rate (eGFR) when the creatinine (Cr) is rapidly changing, as in acute kidney injury (AKI), has been a challenge. The Kinetic Estimated Glomerular Filtration Rate (KeGFR) formula by S. Chen estimates the GFR in the acute state by factoring the time interval between rising Cr values and the volume of distribution (V_D). It provides the clinician with an eGFR value for each non-steady state Cr value. We applied the KeGFR formula to detect AKI in an adult non-ICU inpatient setting. We then compared KeGFR with the current standard Acute Kidney Injury Network (AKIN criteria) and Risk Injury Failure Loss End-Stage Kidney Disease (RIFLE criteria) and new criteria (Waikar-Bonventre, Delta check) for AKI detection. Methods: A total of 250 consecutive adult patients admitted to the Medical wards were screened. Patient episodes with a change in Cr of > 4.3% (Biological Variation) were included in the study (n = 80). The KeGFR equation was applied to this cohort after calculating the V_D individually after estimating the initial GFR by using MDRD equation. A fall in KeGFR of 25% or more was considered as AKI. The AKIN, Waikar-Bonventre, RIFLE, and Delta Check criteria were also applied to this cohort and compared with the KeGFR criterion. Clinical adjudication was performed when there was discordance between AKI episodes detected by AKIN and KeGFR criteria. Results: There were 50 episodes of AKI by AKIN classification and 31 episodes by KeGFR criterion. All but 1 (30) episode detected by KeGFR criterion fulfilled the AKI definition by AKIN. AKIN diagnosed an additional 20 episodes. However, all of these had an elevated Cr level on admission; thus, requiring the incorporation of baseline Cr by AKIN which is not part of the KeGFR formula. Five of these 20 patients were deemed as not having AKI by clinical adjudication. All patients with in-hospital AKI and ongoing AKI were detected by both the criteria. The KeGFR criteria detected almost all (24/25) episodes of AKI identified by the Waikar-Bonventre method. The latter method detected 77% (24/31) of the AKI episodes identified by KeGFR. Conclusion: The KeGFR equation can be readily applied to estimate GFR in the non-steady state. A KeGFR-based criterion successfully detected ongoing and in-hospital AKI in this study. Community acquired AKI that did not progress after admission was not detected by the KeGFR criterion.

Use of Flutemetamol F 18–Labeled Positron Emission Tomography and Other Biomarkers to Assess Risk of Clinical Progression in Patients With Amnestic Mild Cognitive Impairment

ImportancePatients with amnestic mild cognitive impairment (aMCI) may progress to clinical Alzheimer disease (AD), remain stable, or revert to normal. Earlier progression to AD among patients who were β-amyloid positive vs those who were β-amyloid negative has been previously observed. Current research now accepts that a combination of biomarkers could provide greater refinement in the assessment of risk for clinical progression.ObjectiveTo evaluate the ability of flutemetamol F 18 and other biomarkers to assess the risk of progression from aMCI to probable AD.Design, Setting, and ParticipantsIn this multicenter cohort study, from November 11, 2009, to January 16, 2014, patients with aMCI underwent positron emission tomography (PET) at baseline followed by local clinical assessments every 6 months for up to 3 years. Patients with aMCI (365 screened; 232 were eligible) were recruited from 28 clinical centers in Europe and the United States. Physicians remained strictly blinded to the results of PET, and the standard of truth was an independent clinical adjudication committee that confirmed or refuted local assessments. Flutemetamol F 18–labeled PET scans were read centrally as either negative or positive by 5 blinded readers with no knowledge of clinical status. Statistical analysis was conducted from February 19, 2014, to January 26, 2018.InterventionsFlutemetamol F 18–labeled PET at baseline followed by up to 6 clinical visits every 6 months, as well as magnetic resonance imaging and multiple cognitive measures.Main Outcomes and MeasuresTime from PET to probable AD or last follow-up was plotted as a Kaplan-Meier survival curve; PET scan results, age, hippocampal volume, and aMCI stage were entered into Cox proportional hazards logistic regression analyses to identify variables associated with progression to probable AD.ResultsOf 232 patients with aMCI (118 women and 114 men; mean [SD] age, 71.1 [8.6] years), 98 (42.2%) had positive results detected on PET scan. By 36 months, the rates of progression to probable AD were 36.2% overall (81 of 224 patients), 53.6% (52 of 97) for patients with positive results detected on PET scan, and 22.8% (29 of 127) for patients with negative results detected on PET scan. Hazard ratios for association with progression were 2.51 (95% CI, 1.57-3.99; P < .001) for a positive β-amyloid scan alone (primary outcome measure), 5.60 (95% CI, 3.14-9.98; P < .001) with additional low hippocampal volume, and 8.45 (95% CI, 4.40-16.24; P < .001) when poorer cognitive status was added to the model.Conclusions and RelevanceA combination of positive results of flutemetamol F 18–labeled PET, low hippocampal volume, and cognitive status corresponded with a high probability of risk of progression from aMCI to probable AD within 36 months.

Administrative data underestimate acute ischemic stroke events and thrombolysis treatments: Data from a multicenter validation survey in Italy.

BackgroundInforming health systems and monitoring hospital performances using administrative data sets, mainly hospital discharge data coded according to International-Classification-Diseases-9edition-Clinical-Modifiers (ICD9-CM), is now commonplace in several countries, but the reliability of diagnostic coding of acute ischemic stroke in the routine practice is uncertain. This study aimed at estimating accuracy of ICD9-CM codes for the identification of acute ischemic stroke and the use of thrombolysis treatment comparing hospital discharge data with medical record review in all the six hospitals of the Florence Area, Italy, through 2015.MethodsWe reviewed the medical records of all the 3915 potential acute stroke events during 2015 across the six hospitals of the Florence Area, Italy. We then estimated sensitivity and Positive Predictive Value of ICD9-CM code-groups 433*1, 434*1 and thrombolysis code 99.10 against medical record review with clinical adjudication. For each false-positive case we obtained the actual diagnosis. For each false-negative case we obtained the primary and secondary ICD9-CM diagnoses.ResultsThe medical record review identified 1273 acute ischemic stroke events. The hospital discharge records identified 898 among those (true-positive cases),but missed 375 events (false-negative cases), and identified 104 events that were not eventually confirmed as acute ischemic events (false-positive cases). Code-group specific Positive Predictive Value was 85.7% (95%CI,74.6–93.3) for 433*1 and 89.9% (95%CI, 87.8–91.7) for 434*1 codes. Thrombolysis treatment, as identified by ICD9-CM code 99.10, was only documented in 6.0% of acute ischemic stroke events, but was 13.6% in medical record review.ConclusionsHospital discharge data were found to be fairly specific but insensitive in the reporting of acute ischemic stroke and thrombolysis, providing misleading indications about both quantity and quality of acute ischemic stroke hospital care. Efforts to improve coding accuracy should precede the use of hospital discharge data to measure hospital performances in acute ischemic stroke care.

Clinical Trial End Points for Hemodialysis Vascular Access: Background, Rationale, and Definitions.

Hemodialysis vascular access is the “lifeline” for patients on hemodialysis. Unfortunately, because of its poor patency and significant complication rate, it is also the “Achilles heel” of hemodialysis. For example, the current unassisted arteriovenous fistula (AVF) maturation rate is only

Inferring pregnancy episodes and outcomes within a network of observational databases.

Administrative claims and electronic health records are valuable resources for evaluating pharmaceutical effects during pregnancy. However, direct measures of gestational age are generally not available. Establishing a reliable approach to infer the duration and outcome of a pregnancy could improve pharmacovigilance activities. We developed and applied an algorithm to define pregnancy episodes in four observational databases: three US-based claims databases: Truven MarketScan® Commercial Claims and Encounters (CCAE), Truven MarketScan® Multi-state Medicaid (MDCD), and the Optum ClinFormatics® (Optum) database and one non-US database, the United Kingdom (UK) based Clinical Practice Research Datalink (CPRD). Pregnancy outcomes were classified as live births, stillbirths, abortions and ectopic pregnancies. Start dates were estimated using a derived hierarchy of available pregnancy markers, including records such as last menstrual period and nuchal ultrasound dates. Validation included clinical adjudication of 700 electronic Optum and CPRD pregnancy episode profiles to assess the operating characteristics of the algorithm, and a comparison of the algorithm’s Optum pregnancy start estimates to starts based on dates of assisted conception procedures. Distributions of pregnancy outcome types were similar across all four data sources and pregnancy episode lengths found were as expected for all outcomes, excepting term lengths in episodes that used amenorrhea and urine pregnancy tests for start estimation. Validation survey results found highest agreement between reviewer chosen and algorithm operating characteristics for questions assessing pregnancy status and accuracy of outcome category with 99–100% agreement for Optum and CPRD. Outcome date agreement within seven days in either direction ranged from 95–100%, while start date agreement within seven days in either direction ranged from 90–97%. In Optum validation sensitivity analysis, a total of 73% of algorithm estimated starts for live births were in agreement with fertility procedure estimated starts within two weeks in either direction; ectopic pregnancy 77%, stillbirth 47%, and abortion 36%. An algorithm to infer live birth and ectopic pregnancy episodes and outcomes can be applied to multiple observational databases with acceptable accuracy for further epidemiologic research. Less accuracy was found for start date estimations in stillbirth and abortion outcomes in our sensitivity analysis, which may be expected given the nature of the outcomes.

A Case Study of the Incremental Utility for Disease Identification of Natural Language Processing in Electronic Medical Records

Information exists as unstructured medical text in healthcare databases. Such information is not routinely considered in safety surveillance but typically relies solely on structured (coded) data. Natural language processing (NLP) may allow the capture of concepts from unstructured data and thus enhance safety surveillance capability. We sought to assess the added contribution of unstructured data extracted from medical text by NLP for detecting acute liver dysfunction (ALD) in patients with inflammatory bowel disease (IBD). Using a previously developed rule, we evaluated structured and unstructured NLP-extracted terms from a commercially available electronic medical record (EMR) system. The rule was intended to identify ALD diagnosis and timing of onset and was the result of three iterations of rule development using 150 ALD candidate cases. We evaluated the performance of the rule with or without NLP among all candidate cases and among 50 new cases with clinical adjudication. NLP terms were necessary for the diagnosis of 9% of cases and for ruling out 3% of false-positive cases. Inclusion of NLP terms led to an identification of an additional 9% of ALD-onset dates, with consequent earlier recognition in 5%. NLP-derived terms in one large commercially available EMR system modestly improved the sensitivity and specificity in the identification of ALD and identified earlier onset.

The SEP-SEQ Trial: Clinical Validation of the Karius Plasma Next-Generation Sequencing Test for Pathogen Detection in Sepsis

BackgroundSepsis is a leading cause of death and can be caused by a wide range of potential pathogens. In up to 40% of cases, a causative pathogen is never identified. There is a need for improved diagnostic tests that can accurately identify the breadth of potential pathogens to inform effective antimicrobial therapy.MethodsWe enrolled a prospective cohort of patients presenting to the hospital with signs and symptoms of sepsis. Plasma samples were collected for NGS testing at time of initial blood culture. Extracted plasma cell-free DNA was sequenced, human sequences removed and remaining reads aligned against a pathogen database consisting of viruses, bacteria, and eukaryotic pathogens. Relative abundance was estimated; pathogens present at high statistical significance were identified. NGS results were compared with a composite reference standard of all microbiology testing performed within 7 days of admission and clinical diagnosis.ResultsOf 286 patients enrolled, plasma NGS identified potential pathogens in 60.1% (172 of 286) of septic subjects including DNA viruses, bacteria (including fastidious/unculturable bacteria like Mycobacterium tuberculosis), and fungi. In contrast, 15.7% (45 of 286) subjects had a positive initial blood culture and 38.1% (109 of 286) had a potential infectious etiology identified using a composite microbiology laboratory standard. The NGS plasma assay had a positive agreement of 86.7% (39 of 45) and 79.5% (78 of 98) compared with initial blood culture (after excluding contaminants) and the composite laboratory reference standard, respectively. After clinical adjudication, 81.4% (140 of 172) of the positive plasma NGS results were deemed to be consistent with the septic event. Of the remaining 32 subjects, 15 had NGS results that were plausible causes of sepsis but clinical were insufficient to confirm this.ConclusionsWith a single blood draw, the Karius plasma NGS assay identified a broad range of pathogens in septic patients three times more often than blood culture and more often than all microbiology tests combined. This plasma NGS test can identify a viruses, bacteria, and eukaryotic pathogens which can provide valuable information to help clinicians better target antimicrobial therapy for patients with sepsis.Disclosures H. Seng, Karius, Inc.: Employee, Salary. D. Hollemon, Karius, Inc.: Employee, Salary. D. Hong, Karius, Inc.: Employee, Salary. T. Blauwkamp, Karius, Inc.: Employee, Salary. M. Kertesz, Karius, Inc.: Employee, Salary. S. Yang, Karius, Inc.: Research Contractor, Research grant.