Clinical Trial Eligibility Criteria Research Articles

A Framework for Systematic Assessment of Clinical Trial Population Representativeness Using Electronic Health Records Data.

Clinical trials are the gold standard for generating robust medical evidence, but clinical trial results often raise generalizability concerns, which can be attributed to the lack of population representativeness. The electronic health records (EHRs) data are useful for estimating the population representativeness of clinical trial study population. This research aims to estimate the population representativeness of clinical trials systematically using EHR data during the early design stage. We present an end-to-end analytical framework for transforming free-text clinical trial eligibility criteria into executable database queries conformant with the Observational Medical Outcomes Partnership Common Data Model and for systematically quantifying the population representativeness for each clinical trial. We calculated the population representativeness of 782 novel coronavirus disease 2019 (COVID-19) trials and 3,827 type 2 diabetes mellitus (T2DM) trials in the United States respectively using this framework. With the use of overly restrictive eligibility criteria, 85.7% of the COVID-19 trials and 30.1% of T2DM trials had poor population representativeness. This research demonstrates the potential of using the EHR data to assess the clinical trials population representativeness, providing data-driven metrics to inform the selection and optimization of eligibility criteria.

Transformer-Based Named Entity Recognition for Parsing Clinical Trial Eligibility Criteria.

The rapid adoption of electronic health records (EHRs) systems has made clinical data available in electronic format for research and for many downstream applications. Electronic screening of potentially eligible patients using these clinical databases for clinical trials is a critical need to improve trial recruitment efficiency. Nevertheless, manually translating free-text eligibility criteria into database queries is labor intensive and inefficient. To facilitate automated screening, free-text eligibility criteria must be structured and coded into a computable format using controlled vocabularies. Named entity recognition (NER) is thus an important first step. In this study, we evaluate 4 state-of-the-art transformer-based NER models on two publicly available annotated corpora of eligibility criteria released by Columbia University (i.e., the Chia data) and Facebook Research (i.e.the FRD data). Four transformer-based models (i.e., BERT, ALBERT, RoBERTa, and ELECTRA) pretrained with general English domain corpora vs. those pretrained with PubMed citations, clinical notes from the MIMIC-III dataset and eligibility criteria extracted from all the clinical trials on ClinicalTrials.gov were compared. Experimental results show that RoBERTa pretrained with MIMIC-III clinical notes and eligibility criteria yielded the highest strict and relaxed F-scores in both the Chia data (i.e., 0.658/0.798) and the FRD data (i.e., 0.785/0.916). With promising NER results, further investigations on building a reliable natural language processing (NLP)-assisted pipeline for automated electronic screening are needed.

Extracapsular extension, pathologic node status, and adjuvant treatment in primary surgery patients with human papillomavirus-mediated oropharyngeal cancer: National hospital-based retrospective cohort analysis.

The significance of extracapsular extension (ECE) and adjuvant treatment paradigm in patients with surgically managed human papillomavirus-positive (HPV+) oropharyngeal cancer (OPC) is debated. National, hospital-based, retrospective cohort study of 2663 patients pN+ HPV+ OPC who underwent primary surgery. Patients with ECE had a 1.74-times risk of death (95% confidence interval [CI]: 1.26-2.40, p=0.001) compared to patients without ECE. Among patients with pN1, ECE-positive disease, risk of overall mortality was similar across treatment paradigms (surgery alone: ref; adjuvant radiation therapy [RT]: aHR: 0.81; 95% CI: 0.36-1.85; p=0.62; adjuvant CRT: aHR: 0.66; 95% CI: 0.34-1.32; p=0.24). Patients with pN2 ECE-positive disease treated with adjuvant RT alone exhibited similar risk of all-cause mortality (hazard ratio: 1.04, 95% CI: 0.24-4.47, p=0.96) compared to adjuvant chemoradiation (CRT). In patients with advanced, ECE-positive disease (e.g., pT3-T4pN2), adjuvant CRT did not reduce the risk of overall mortality relative to adjuvant RT. Although pathologic ECE negatively predicts for survival in patients with HPV+ OPC, our analyses support expansion of postoperative de-intensification clinical trial eligibility criteria in patients with ECE-positive disease.

Eligibility criteria and clinical trials: An FDA perspective

IntroductionRandomized clinical trials are regarded as the gold standard for evaluating the effectiveness and safety of interventions. The US Food and Drug Administration (FDA) has made efforts to promote inclusive eligibility criteria. The objective of this study is to identify common trends in eligibility criteria and identify patterns of exclusion criteria among different diseases. MethodsThe authors evaluated the inclusion and exclusion criteria of 38 pivotal clinical trials representing 38 novel drugs approved between 2014 and 2017. Additionally, the authors reviewed the demographic characteristics of participants enrolled across 38 trials. ResultsEighty two percent of trials in our study excluded participants based on hepatic related criteria and 79% trials excluded participants based on renal related criteria and specific infectious diseases. For trials in conditions that affected both men and women, there were no exclusions based on gender and race. More than 90% of trials excluded pregnant, lactating women and women not on adequate contraception. Of the 36,644 patients enrolled in trials for which both men and women were eligible, 62% were men. ConclusionsThe most frequent exclusion criteria were pregnancy, lactation/breastfeeding, renal and hepatic abnormalities, and specific infectious diseases. The preponderance of men in our study likely indicates that factors other than exclusion criteria affect enrollment. The lower representation of women may have been influenced by two large cardiovascular trials that included 75% men. Our study marks an important step in the ongoing efforts of the Agency to increase inclusivity in clinical trials by understanding common clinical trial eligibility criteria patterns.

Remission, treatment failure, and relapse in pediatric ALL: an international consensus of the Ponte-di-Legno Consortium

Comparison of treatment strategies in de novo pediatric acute lymphoblastic leukemia (ALL) requires standardized measures of efficacy. Key parameters that define disease-related events, including complete remission (CR), treatment failure (TF; not achieving CR), and relapse (loss of CR) require an updated consensus incorporating modern diagnostics. We collected the definitions of CR, TF, and relapse from recent and current pediatric clinical trials for the treatment of ALL, including the key components of response evaluation (timing, anatomic sites, detection methods, and thresholds) and found significant heterogeneity, most notably in the definition of TF. Representatives of the major international ALL clinical trial groups convened to establish consensus definitions. CR should be defined at a time point no earlier than at the end of induction and should include the reduction of blasts below a specific threshold in bone marrow and extramedullary sites, incorporating minimal residual disease (MRD) techniques for marrow evaluations. TF should be defined as failure to achieve CR by a prespecified time point in therapy. Relapse can only be defined in patients who have achieved CR and must include a specific threshold of leukemic cells in the bone marrow confirmed by MRD, the detection of central nervous system leukemia, or documentation of extramedullary disease. Definitions of TF and relapse should harmonize with eligibility criteria for clinical trials in relapsed/refractory ALL. These consensus definitions will enhance the ability to compare outcomes across pediatric ALL trials and facilitate development of future international collaborative trials.

Automated classification of clinical trial eligibility criteria text based on ensemble learning and metric learning

BackgroundEligibility criteria are the primary strategy for screening the target participants of a clinical trial. Automated classification of clinical trial eligibility criteria text by using machine learning methods improves recruitment efficiency to reduce the cost of clinical research. However, existing methods suffer from poor classification performance due to the complexity and imbalance of eligibility criteria text data.MethodsAn ensemble learning-based model with metric learning is proposed for eligibility criteria classification. The model integrates a set of pre-trained models including Bidirectional Encoder Representations from Transformers (BERT), A Robustly Optimized BERT Pretraining Approach (RoBERTa), XLNet, Pre-training Text Encoders as Discriminators Rather Than Generators (ELECTRA), and Enhanced Representation through Knowledge Integration (ERNIE). Focal Loss is used as a loss function to address the data imbalance problem. Metric learning is employed to train the embedding of each base model for feature distinguish. Soft Voting is applied to achieve final classification of the ensemble model. The dataset is from the standard evaluation task 3 of 5th China Health Information Processing Conference containing 38,341 eligibility criteria text in 44 categories.ResultsOur ensemble method had an accuracy of 0.8497, a precision of 0.8229, and a recall of 0.8216 on the dataset. The macro F1-score was 0.8169, outperforming state-of-the-art baseline methods by 0.84% improvement on average. In addition, the performance improvement had a p-value of 2.152e-07 with a standard t-test, indicating that our model achieved a significant improvement.ConclusionsA model for classifying eligibility criteria text of clinical trials based on multi-model ensemble learning and metric learning was proposed. The experiments demonstrated that the classification performance was improved by our ensemble model significantly. In addition, metric learning was able to improve word embedding representation and the focal loss reduced the impact of data imbalance to model performance.

Uncovering the Blind Spot of Clinical Trials: First Report of Baseline Characteristics of Newly Diagnosed (ND) and Relapsed/Refractory (RR) Multiple Myeloma (MM) Patients (Pts) in Insight-MM, a Global, Prospective, Observational Study

Uncovering the Blind Spot of Clinical Trials: First Report of Baseline Characteristics of Newly Diagnosed (ND) and Relapsed/Refractory (RR) Multiple Myeloma (MM) Patients (Pts) in Insight-MM, a Global, Prospective, Observational Study

Model Heterogeneity in Predicting Outcomes of Patients with Myelodysplastic Syndromes (MDS)

BackgroundThe outcome of MDS patients (pts) is heterogeneous, with some pts living for years, while others die within a few months of diagnosis. Several prognostic models have been developed to risk stratify pts with MDS including: the International Prognostic Scoring System (IPSS), Revised IPSS (IPSS-R), World Health Organization classification-based Prognostic Scoring System (WPSS), and MD Anderson Prognostic Scoring System (MDPSS). When applying these models to clinical practice, considerable heterogeneity in their prediction of outcomes is observed. Understanding such heterogeneity can impact disease expectations and treatment recommendations, as these can differ considerably based on the model used.MethodsClinical data from pts who were diagnosed with MDS according to 2008 WHO criteria and treated at our institution between 1/2000 and 1/2017 were analyzed. All prognostic models were calculated at the time of MDS diagnosis. Overall survival (OS) was calculated from the time of diagnosis to time of last follow up or death. Descriptive statistics were used to study the changes in risk categories across models.ResultsOf 687 pts included in the final analysis, the median age at diagnosis was 68 (11-99) years with a male (63%) predominance. Cytogenetic analysis per IPSS-R: very good 19 (3%), good 441(64%), intermediate (INT) 117 (17%), poor 47 (7%), and very poor 63 (9%). Risk categories by IPSS included: 204 (30%) with low risk, 319 (46%) intermediate-1 (INT-1), 124 (18%) intermediate-2 (INT-2), and 40 (6%) high risk. Risk categories by IPSS-R included: 114 pts (17%) with very low risk, 264 (38%) low, 128 (19%) intermediate, 118 (17%) high, and 63 (9%) very high.Significant heterogeneity was observed when comparing all models to IPSS-R risk categories. Among 378 pts with lower-risk disease (Very Low / Low) per IPSS-R, 3 pts (1%) were re-classified INT-2 per IPSS, 73 (19%) as INT and 28 (7%) as high/very high per WPSS, and 11 (29%) as higher-risk per MDAPSS, Table 1. Among 128 pts with intermediate risk by IPSS-R, 102 (20%) were reclassified as lower-risk per IPSS, 44 (34%) as lower-risk and 47 (37%) as higher-risk per WPSS, and 65 (50%) as lower-risk per MDAPSS, Table 1. Among 181 pts with high/very high-risk by IPSS-R, 44 (25%) were reclassified as lower-risk by IPSS, 54 (30%) per WPSS, and 58 (32%) per MDAPSS, Table1. Among 523 pts with lower-risk (low / INT-1) by IPSS, 102 (20%) were reclassified as INT risk and 46 (9%) as higher-risk per IPSS-R, 113 (22%) as INT and 62 (12%) as higher-risk per WPSS, and 186 (36%) as higher risk per MDAPSS, Table 2. Among 164 pts with higher-risk (INT-2/High) per IPSS, 26 (16%) were reclassified as INT per IPSS-R, 24 (15%) per WPSS, and 53 (32%) per MDAPSS, Table 2.The predicted OS per IPSS-R was overestimated when compared to the actual OS (evaluable in 429 pts) for pts in lower-risk categories. The difference between the median predicted IPSS-R OS and median actual OS was 70.6 months (m) for very low risk, 54 m for low, 9.5 m for INT, 6.7 m for high, and -2.4 m for very high risk (pts lived longer than their predicted OS). The IPSS, on the other hand, underestimated the OS, as the difference between the median predicted OS and median actual OS was -23.3 m for low risk, -16.5 m for INT1, -11.8 m for INT2, and -11.1 m for high risk.ConclusionsSignificant intra-patient (difference of outcome in the same pt based on the model used) and intra-group (difference in outcome among pts who have similar risk category by a given model) variability is observed when applying commonly used models in MDS. More importantly, IPSS-R overestimate OS for lower-risk disease and IPSS underestimate OS across all risk categories. This heterogeneity can impact survival expectations, treatment strategies, and clinical trial eligibility criteria based on the model used. A personalized prediction model that can precisely forecast the actual outcome of a pt is needed. [Display omitted] DisclosuresGerds:Incyte: Consultancy; CTI BioPharma: Consultancy. Advani:Pfizer: Consultancy; Takeda/ Millenium: Research Funding. Maciejewski:Alexion Pharmaceuticals, Inc.: Consultancy, Membership on an entity's Board of Directors or advisory committees, Other: Speaker Fees; Apellis Pharmaceuticals: Consultancy; Ra Pharma: Consultancy. Sekeres:Celgene: Membership on an entity's Board of Directors or advisory committees.

Real-World and Clinical Trial Data in Relapsed/Refractory Multiple Myeloma (RRMM): Evaluating Treatment Duration and Comparing Effectiveness and Efficacy

Real-World and Clinical Trial Data in Relapsed/Refractory Multiple Myeloma (RRMM): Evaluating Treatment Duration and Comparing Effectiveness and Efficacy

Are clinical trial eligibility criteria representative of older patients with lung cancer? A population-based data linkage study

ObjectivesOlder adults constitute the majority of patients with lung cancer. However, they are under-represented in clinical trials as eligibility criteria often restrict enrolment based on comorbidities that are common with aging. We aimed to describe comorbidities relating to trial exclusion criteria in older adults with lung cancer, determine the proportion that would typically be excluded from trials, and examine the impact on treatment uptake. Materials and MethodsWe conducted a population-based study of people aged ≥65 years diagnosed with metastatic lung cancer using linked data for clients of the Australian Government Department of Veterans' Affairs (2005–2015). We defined trial-typical patients based on the absence of comorbidities related to the following: inadequate organ (cardiac, renal, hepatic, marrow) function; cognitive dysfunction; poor performance status (PS); prior malignancy within 5 years. We report systemic therapy uptake within 3 months of diagnosis. ResultsOur study included 677 patients (median age 84). Over half (53.4%) were not trial-typical, with the most common reasons being poor PS (37.5%), cardiac disease (19.2%), and prior cancer (12.9%). Eighty-two (12.1%) received systemic therapy. Patients with poor PS, cardiac disease, and dementia had lower treatment uptake rates. However, there was no significant difference in treatment uptake between trial-typical and non-trial-typical patients (13.4 vs 11.0%). ConclusionMore than half of older adults with advanced lung cancer would be typically excluded from trial participation. Future clinical trials of older adults need to consider broader eligibility criteria to better reflect this population to gain the best evidence for their care.

Potential Role of Clinical Trial Eligibility Criteria in Electronic Phenotyping.

2,719 distinctive phenotyping variables from 176 electronic phenotypes were compared with 57,150 distinctive clinical trial eligibility criteria concepts to assess the phenotype knowledge overlap between them. We observed a high percentage (69.5%) of eMERGE phenotype features and a lower percentage (47.6%) of OHDSI phenotype features matched to clinical trial eligibility criteria, possibly due to the relative emphasis on specificity for eMERGE phenotypes and the relative emphasis on sensitivity for OHDSI phenotypes. The study results show the potential of reusing clinical trial eligibility criteria for phenotyping feature selection and moderate benefits of using them for local cohort query implementation.

Participatory Design of a Clinical Trial Eligibility Criteria Simplification Method.

Clinical trial eligibility criteria are important for selecting the right participants for clinical trials. However, they are often complex and not computable. This paper presents the participatory design of a human-computer collaboration method for criteria simplification that includes natural language processing followed by user-centered eligibility criteria simplification. A case study on the ARCADIA trial shows how criteria were simplified for structured database querying by clinical researchers and identifies rules for criteria simplification and concept normalization.

Phenotyping of clinical trial eligibility text from cancer studies into computable criteria in electronic health records.

6592 Background: Clinical trial phenotyping is the process of extracting clinical features and patient characteristics from eligibility criteria. Phenotyping is a crucial step that precedes automated cohort identification from patient electronic health records (EHRs) against trial criteria. We establish a clinical trial phenotyping pipeline to transform clinical trial eligibility criteria into computable criteria and enable high throughput cohort selection in EHRs. Methods: Formalized clinical trial criteria attributes were acquired from a natural-language processing (NLP)-assisted approach. We implemented a clinical trial phenotyping pipeline that included three components: First, a rule-based knowledge engineering component was introduced to annotate the trial attributes into a computable and customizable granularity from EHRs. The second component involved normalizing annotated attributes using standard terminologies and pre-defined reference tables. Third, a knowledge base of computable criteria attributes was built to match patients to clinical trials. We evaluated the pipeline performance by independent manual review. The inter-rater agreement of the annotation was measured on a random sample of the knowledge base. The accuracy of the pipeline was evaluated on a subset of randomly selected matched patients for a subset of randomly selected attributes. Results: Our pipeline phenotyped 2954 clinical trials from five cancer types including Non-Small Cell Lung Cancer, Small Cell Lung Cancer, Prostate Cancer, Breast Cancer, and Multiple Myeloma. We built a knowledge base of 256 computable attributes that included comorbidities, comorbidity-related treatment, previous lines of therapy, laboratory tests, and performance such as ECOG and Karnofsky score. Among 256 attributes, 132 attributes were encoded using standard terminologies and 124 attributes were normalized to customized concepts. The inter-rater agreement of the annotation measured by Cohen’s Kappa coefficient was 0.83. We applied the knowledge base to our EHRs and efficiently identified 33258 potential subjects for cancer clinical trials. Our evaluation on the patient matching indicated the F1 score was 0.94. Conclusions: We established a clinical trial phenotyping pipeline and built a knowledge base of computable criteria attributes that enabled efficient screening of EHRs for patients meeting clinical trial eligibility criteria, providing an automated way to efficiently and accurately identify clinical trial cohorts. The application of this knowledge base to patient matching from EHR data across different institutes demonstrates its generalization capability. Taken together, this knowledge base will be particularly valuable in computer-assisted clinical trial subject selection and clinical trial protocol design in cancer studies based on real-world evidence.

Broadening Eligibility Criteria in Cancer Clinical Trials

Broadening Eligibility Criteria in Cancer Clinical Trials

Clinical trial representativeness and treatment intensity in a real-world sample of women with early-stage breast cancer.

6584 Background: Early stage breast cancer (EBC) treatment is used in women of all ages, races, and health states. However, as clinical trials often do not represent real-world populations, the extent to which evidence-based treatments are prescribed to populations not well represented in these trials is not known. This study evaluated treatment intensity for patients traditionally well represented, underrepresented, and unrepresented in clinical trials. Methods: This retrospective cohort study used the nationwide de-identified electronic health record derived Flatiron Health database for patients diagnosed with EBC between 2011-2020. We categorized treatments as either high- (AC-TH [doxorubicin, cyclophosphamide followed by paclitaxel or docetaxel, trastuzumab]; ACT [paclitaxel or docetaxel, doxorubicin, cyclophosphamide]; TCH [paclitaxel or docetaxel, carboplatin, trastuzumab]; TCHP [paclitaxel or docetaxel, carboplatin, trastuzumab, pertuzumab]) or low-intensity (AC [doxorubicin, cyclophosphamide]; TC [paclitaxel or docetaxel, cyclophosphamide]; TH [paclitaxel or docetaxel, trastuzumab]). Unrepresented patients often have one or more comorbidities and/or prior cancer; underrepresented patients are typically Black, Indigenous, people of color, or of age extremes ( < 45, 70+); well represented patients are White and between the ages of 45-69. Odds ratios (OR), predicted proportions, and 95% confidence intervals (CI) from a two-level (patients nested in practice) hierarchical logistic regression model evaluated associations between receipt of high-intensity chemotherapy and patient characteristics of clinical trial representation (age, race/ethnicity, presence of comorbidity). Results: Our study included 970 patients with EBC with 13%, 45%, and 41% characterized as unrepresented, underrepresented, and well represented in clinical trials, respectively. In the adjusted model, those aged ≥ 70 vs 45-69 had lower odds of receiving a high-intensity treatment (OR 0.40, 95% CI 0.26-0.60), while those aged < 45 vs 45-69 had higher odds of receiving high-intensity treatment (OR 1.82, 95% CI 1.10-3.01). The predicted proportion of patients receiving a high-intensity treatment was 87% (95% CI: 80%-92%) for patients aged < 45, 79% (95% CI: 74%-84%) for patients aged 45-69, and 60% (95% CI: 50%-70%) for patients aged ≥ 70. Neither race/ethnicity nor comorbidity status were associated with odds of receiving high-intensity chemotherapy. Conclusions: Over half of the EBC population is not well represented in clinical trials. Age was associated with differential treatment intensity, despite a lack of evidence that these differences are appropriate. Widening clinical trial eligibility criteria is one way to better understand survival outcomes, identify potential toxicities, and ultimately make evidence-based treatment decisions using a more diverse sample.

Outcomes of patients with solid tumour malignancies treated with first-line immuno-oncology agents who do not meet eligibility criteria for clinical trials

BackgroundImmuno-oncology (IO)–based therapies have been approved based on randomised clinical trials, yet a significant proportion of real-world patients are not represented in these trials. We sought to compare the outcomes of trial-ineligible vs. -eligible patients with advanced solid tumours treated with first-line (1L) IO therapy. Patients and methodsUsing the International Metastatic Renal Cell Carcinoma (RCC) Database Consortium and the Alberta Immunotherapy Database, patients with advanced RCC, non–small-cell lung cancer (NSCLC) or melanoma treated with 1L PD-(L)1 inhibition–based therapy were included. Trial eligibility was retrospectively determined as per commonly used exclusion criteria. The outcomes of interest were overall survival (OS), overall response rate (ORR), treatment duration (TD) and time to next treatment (TTNT). ResultsA total of 395 of 1241 (32%) patients were deemed trial-ineligible. The main reasons for ineligibility based on preselected exclusion criteria were Karnofsky performance status <70%/Eastern Cooperative Oncology Group performance status >1 (40%, 158 of 395), brain metastases (32%, 126 of 395), haemoglobin < 9 g/dL (16%, 63 of 395) and estimated glomerular filtration rate <40 mL/min (15%, 61 of 395).Between the ineligible vs. eligible groups, the median OS, ORR, median TD and median TTNT were 10.2 vs. 39.7 months (p < 0.01), 36% vs. 47% (p < 0.01), 2.7 vs. 6.9 months (p < 0.01) and 6.0 vs. 16.8 months (p < 0.01), respectively. Subgroup analyses showed statistically significant inferior OS, TD and TTNT for trial-ineligible vs. -eligible patients across all tumour types. Adjusted hazard ratios for death in RCC, NSCLC and melanoma were 1.84 (95% confidence interval [CI] 1.22–2.77), 2.21 (95% CI 1.58–3.11) and 1.82 (95% CI 1.21–2.74), respectively.. ConclusionsThirty-two percent of real-world patients treated with contemporary 1L IO-based therapies were ineligible for clinical trials. Although one-third of the trial-ineligible patients responded to treatment, the overall trial-ineligible population had inferior outcomes than trial-eligible patients. These data may guide patient counselling and temper expectations of benefit.

Building an OMOP common data model-compliant annotated corpus for COVID-19 clinical trials

Clinical trials are essential for generating reliable medical evidence, but often suffer from expensive and delayed patient recruitment because the unstructured eligibility criteria description prevents automatic query generation for eligibility screening. In response to the COVID-19 pandemic, many trials have been created but their information is not computable. We included 700 COVID-19 trials available at the point of study and developed a semi-automatic approach to generate an annotated corpus for COVID-19 clinical trial eligibility criteria called COVIC. A hierarchical annotation schema based on the OMOP Common Data Model was developed to accommodate four levels of annotation granularity: i.e., study cohort, eligibility criteria, named entity and standard concept. In COVIC, 39 trials with more than one study cohorts were identified and labelled with an identifier for each cohort. 1,943 criteria for non-clinical characteristics such as “informed consent”, “exclusivity of participation” were annotated. 9767 criteria were represented by 18,161 entities in 8 domains, 7,743 attributes of 7 attribute types and 16,443 relationships of 11 relationship types. 17,171 entities were mapped to standard medical concepts and 1,009 attributes were normalized into computable representations. COVIC can serve as a corpus indexed by semantic tags for COVID-19 trial search and analytics, and a benchmark for machine learning based criteria extraction.

P166 Multimorbidity in SLE, a barrier to clinical trial eligibility: results from the BILAG-BR

Abstract Background/Aims Stringent inclusion and exclusion criteria are employed in SLE clinical trials. Organ dysfunction and co-morbidities are common exclusion criteria which may affect how representative trials are of real-world SLE populations. We aimed to apply published trial eligibility criteria to patients with SLE in a large national register. Methods A literature review of all major published double-blinded randomised phase III trials in non-renal SLE was performed. Common inclusion and exclusion criteria were applied to all patients recruited to the BILAG-Biologics Register (BILAG-BR), a large UK-wide register of SLE patients. Data on comorbidities for all patients registered was collected. The mean (SD) number of co-morbidities was calculated. Patients were then classified as being eligible or ineligible. Groups were compared initially using a chi-squared or Wilcoxon rank-sum test and logistic regression model was used to test the age and sex adjusted association between trial eligibility and comorbidities. Results Common inclusion and exclusion criteria were identified from 12 published trials. When applied to the 837 patients recruited to BILAG-BR, 562 (67%) patients would not be eligible for inclusion in these trials. Ineligible patients had a shorter disease duration (2.9 vs. 5.1 years, p &amp;lt; 0.01), but were similar in age (P = 1.0), sex (P = 0.7) and ethnicity (p = 0.5) to those who were eligible. Of eligible patients, 128 (53%) had 1 or more comorbidities compared with 340 (60%) who were ineligible (p = 0.05). The mean (SD) number of comorbidities was 0.9 (1.2) vs 1.2 (1.3) for eligible and ineligible patients respectively. After adjusting for age and sex, inclusion in clinical trials was associated with fewer comorbidities (OR 0.81, 95% CI 0.70, 0.94, p &amp;lt; 0.01). Conclusion Patients with multi-morbidity are more likely to be ineligible for SLE clinical trials. Evidence from real world studies and registers are therefore needed to fully understand the safety and effectiveness of new therapies. Our data also underscores the need to develop more pragmatic eligibility criteria for clinical trials. Disclosure S. Dyball: None. S. Collinson: None. E. Sutton: None. E. McCarthy: None. B. Parker: Consultancies; GSK, AstraZenica, UCB, Abbvie, Pfizer, BMS, Celltrion. Grants/research support; GSK, Sanofi Genzyme. I. Bruce: Consultancies; GSK, Medimmune, AstraZenica, Eli Lilly, Merck Serono, UCB, ILTOO. Member of speakers’ bureau; AstraZeneca, Medimmune, GSK, UCB. Grants/research support; GSK, Genzyme Sanofi, UCB.

Analysis of Not Structurable Oncological Study Eligibility Criteria for Improved Patient-Trial Matching.

Background Higher enrolment rates of cancer patients into clinical trials are necessary to increase cancer survival. As a prerequisite, an improved semiautomated matching of patient characteristics with clinical trial eligibility criteria is needed. This is based on the computer interpretability, i.e., structurability of eligibility criteria texts. To increase structurability, the common content, phrasing, and structuring problems of oncological eligibility criteria need to be better understood. Objectives We aimed to identify oncological eligibility criteria that were not possible to be structured by our manual approach and categorize them by the underlying structuring problem. Our results shall contribute to improved criteria phrasing in the future as a prerequisite for increased structurability. Methods The inclusion and exclusion criteria of 159 oncological studies from the Clinical Trial Information System of the National Center for Tumor Diseases Heidelberg were manually structured and grouped into content-related subcategories. Criteria identified as not structurable were analyzed further and manually categorized by the underlying structuring problem. Results The structuring of criteria resulted in 4,742 smallest meaningful components (SMCs) distributed across seven main categories (Diagnosis, Therapy, Laboratory, Study, Findings, Demographics, and Lifestyle, Others). A proportion of 645 SMCs (13.60%) was not possible to be structured due to content- and structure-related issues. Of these, a subset of 415 SMCs (64.34%) was considered not remediable, as supplementary medical knowledge would have been needed or the linkage among the sentence components was too complex. The main category “Diagnosis and Study” contained these two subcategories to the largest parts and thus were the least structurable. In the inclusion criteria, reasons for lacking structurability varied, while missing supplementary medical knowledge was the largest factor within the exclusion criteria. Conclusion Our results suggest that further improvement of eligibility criterion phrasing only marginally contributes to increased structurability. Instead, physician-based confirmation of the matching results and the exclusion of factors harming the patient or biasing the study is needed.

Eyes that Do Not Meet the Eligibility Criteria of Clinical Trials on Age-Related Macular Degeneration: Proportion of the Real-World Patient Population and Reasons for Exclusion.

Background To evaluate the proportion of eyes that do not meet the eligibility criteria of clinical trials on neovascular age-related macular degeneration (AMD) and the reasons for exclusion. Methods This retrospective, observational study included 512 eyes of 463 patients diagnosed with treatment-naïve neovascular AMD. The proportion of eyes that did not meet the eligibility criteria of the Vascular Endothelial Growth Factor Trap-Eye: Investigation of Efficacy and Safety in Wet AMD (VIEW) studies were evaluated. The two most common reasons for exclusion were also evaluated in each subtype of neovascular AMD (typical neovascular AMD, polypoidal choroidal vasculopathy (PCV), and type 3 neovascularization). Results Among the 512 eyes, 229 (44.7%) did not meet the eligibility criteria. In all the included eyes, the most common reasons for exclusion were good or poor visual acuity (169 eyes, 33.0%), followed by the presence of subretinal hemorrhage (47 eyes, 9.5%). Moreover, good or poor visual acuity was the most common reason for exclusion in all three subtypes of neovascular AMD. The second most common reason was a fovea-involving scar or fibrosis in typical neovascular AMD, subretinal hemorrhage in PCV, and other vascular diseases affecting the retina in type 3 neovascularization. Conclusions Among the included cases, 44.7% did not meet the eligibility criteria for VIEW study, suggesting that the conclusion derived from clinical trials may not directly reflect the real-world outcomes. Additionally, the reasons for ineligibility differed among the different subtypes of neovascular AMD.