Explore the vast range of titles available.

The use of electronic health records (EHRs) holds the potential to enhance clinical trial activities. However, the identification of eligible patients within EHRs presents considerable challenges. We aimed to develop a CriteriaMapper system for phenotyping eligibility criteria, enabling the identification of patients from EHRs with clinical characteristics that match those criteria. We utilized clinical trial eligibility criteria and patient EHRs from the Mount Sinai Database. The CriteriaMapper system was developed to normalize the criteria using national standard terminologies and in-house databases, facilitating computability and queryability to bridge clinical trial criteria and EHRs. The system employed rule-based pattern recognition and manual annotation. Our system normalized 367 out of 640 unique eligibility criteria attributes, covering various medical conditions including non-small cell lung cancer, small cell lung cancer, prostate cancer, breast cancer, multiple myeloma, ulcerative colitis, Crohn’s disease, non-alcoholic steatohepatitis, and sickle cell anemia. About 174 criteria were encoded with standard terminologies and 193 were normalized using the in-house reference tables. The agreement between automated and manual normalization was high (Cohen’s Kappa = 0.82), and patient matching demonstrated a 0.94 F1 score. Our system has proven effective on EHRs from multiple institutions, showing broad applicability and promising improved clinical trial processes, leading to better patient selection, and enhanced clinical research outcomes.

Background Patients with inflammatory bowel diseases (IBD) sometimes require investigational medicinal therapy in a clinical trial. Before enrollment, patients must meet strict eligibility criteria, hampering recruitment rates. We investigated the rates, causes, and outcomes of screening failure (SF) in a tertiary IBD center. Methods We reviewed all IBD patients screened for sponsored multicenter phase 1-3 induction studies with available global SF rates between January 2008 and March 2021. We compared our SF rates with the global SF rates. Causes of SF were categorized into disease activity, hematology, chemistry, microbiology, protocol violation, and withdrawal of consent. Patient outcomes were categorized into rescreening for the same trial, screening for another trial, (re)introduction of commercially available therapy, surgery, or watchful waiting. Results During the study period, 642 local screenings were performed as part of 53 studies. We identified an overall SF rate of 17.1%, compared with 39.2% in the global study population (P < .00001). Causes of SF at our center included ineligible disease activity (36.4%), microbiology (25.5%), protocol violation (16.4%), withdrawal of consent (9.1%), chemistry (6.4%) and hematology (6.4%). Thirty SFs could have been avoided by prescreening that was more thorough. After SF, 34 patients were rescreened for the same trial, 17 screened for another trial, 38 initiated approved therapy, 9 were referred for surgery, and 12 did not receive further therapy. Conclusions A significant proportion of IBD patients consenting to clinical trials fail their screening. Main causes of SF are ineligible disease activity and abnormal finding on microbiology. Approximately one-fourth of SFs could have been avoided by prescreening that was more thorough.

Background Clinical trial protocols are the foundation for advancing medical sciences, however, the extraction of accurate and meaningful information from the original clinical trials is very challenging due to the complex and unstructured texts of such documents. Named entity recognition (NER) is a fundamental and necessary step to process and standardize the unstructured text in clinical trials using Natural Language Processing (NLP) techniques. Methods In this study we fine-tuned pre-trained language models to support the NER task on clinical trial eligibility criteria. We systematically investigated four pre-trained contextual embedding models for the biomedical domain (i.e., BioBERT, BlueBERT, PubMedBERT, and SciBERT) and two models for the open domains (BERT and SpanBERT), for NER tasks using three existing clinical trial eligibility criteria corpora. In addition, we also investigated the feasibility of data augmentation approaches and evaluated their performance. Results Our evaluation results using tenfold cross-validation show that domain-specific transformer models achieved better performance than the general transformer models, with the best performance obtained by the PubMedBERT model (F1-scores of 0.715, 0.836, and 0.622 for the three corpora respectively). The data augmentation results show that it is feasible to leverage additional corpora to improve NER performance. Conclusions Findings from this study not only demonstrate the importance of contextual embeddings trained from domain-specific corpora, but also shed lights on the benefits of leveraging multiple data sources for the challenging NER task in clinical trial eligibility criteria text.

The number of systematic reviews that aim to compare multiple interventions using network meta-analysis is increasing. In this study, we highlight aspects of a standard systematic review protocol that may need modification when multiple interventions are to be compared. We take the protocol format suggested by Cochrane for a standard systematic review as our reference and compare the considerations for a pairwise review with those required for a valid comparison of multiple interventions. We suggest new sections for protocols of systematic reviews including network meta-analyses with a focus on how to evaluate their assumptions. We provide example text from published protocols to exemplify the considerations. Standard systematic review protocols for pairwise meta-analyses need extensions to accommodate the increased complexity of network meta-analysis. Our suggested modifications are widely applicable to both Cochrane and non-Cochrane systematic reviews involving network meta-analyses.

Background Target trial emulations often face challenges in accurately identifying the intended target population within real-world data (RWD) because eligibility criteria cannot always be directly mapped to available variables. This scoping review aimed to systematically characterize the proportion of eligibility criteria from oncology target trials that are successfully mapped to their emulated counterparts and to describe the approaches used for this mapping. Methods We searched MEDLINE for peer-reviewed studies published between January 1, 2016, and April 23, 2025, that were designed to emulate oncology drug intervention trials using RWD. For each target trial emulation, we recorded design elements, database types, eligibility criteria, and methodological details. We identified how each target trial eligibility criterion was mapped to available RWD. For each applicable criterion, we assessed whether it was explicitly represented and conceptually consistent with the target trial; criteria meeting both were classified as mapped. Eligibility criteria were categorized into 11 domains representing key demographic, clinical, and safety considerations. Results Our search identified 200 studies, of which 47 reported the results of 74 individual target trial emulations. Of these 74, 42 (56.8%) were emulations of actual clinical trials and 32 (43.2%) were emulations of hypothetical target trials. Most target trial emulations used registry data (30, 40.5%) or electronic health records (29, 39.2%) as their primary data sources. Emulations of actual clinical trials specified a median of 29 (interquartile range [IQR] 25.0–34.0) eligibility criteria and mapped a median of 35.5% (IQR 27.3%-40.6%) of these criteria per emulation, whereas emulations of hypothetical trials specified a median of 5.5 (IQR 4.3–12.8) eligibility criteria and mapped a median of 91.7% (IQR 70%-100%) of these criteria per emulation. Among 1,222 individual criteria, demographic criteria were most frequently mapped (46, 93.9%), while safety and concomitant risk–related criteria were least frequently mapped (7, 3.7%). The most common mapping method was direct mapping using administrative coding systems (37, 72.5%); no studies reported using common data models or computational methods. Conclusions This scoping review of oncology target trial emulations found substantial heterogeneity in how eligibility criteria were mapped, with potential implications for emulation results. These findings highlight the need for greater transparency and for methods that extend beyond reliance on structured administrative codes.

The objective of the study was to examine the characteristics of randomized controlled trials (RCTs) evaluating physiotherapy interventions for low back pain (LBP) that specified a language-grounded eligibility criterion and the proportion of people being excluded consequently. This is a meta-epidemiological study of RCTs evaluating at least one type of physiotherapy intervention for treatment or prevention of LBP. Records were retrieved from Physiotherapy Evidence Database (PEDro), LILACS, and SciELO from inception to May 2021. We retrieved metadata of each record from PEDro and extracted from included studies: country of recruitment, language-grounded eligibility criterion, and the number of consequent exclusions (if specified). This study included 2,555 trials. A language-grounded eligibility criterion was specified in 463 trials (18.1%); the proportion was higher in trials conducted in North America and Europe, published after 2000, investigating cognitive and behavioral interventions, and including large sample size. Of these 463 trials, 75 trials (16.2%) reported a total number of 2,152 people being excluded due to lack of language proficiency, equivalent to 12.5% of randomized participants. Nearly one in five physiotherapy clinical trials on LBP excludes people based on language proficiency, compromising the evidence to manage LBP in minority populations.

Background Breast cancer (BC) is the most common cancer type in women. The purpose of this study was to assess the eligibility criteria in recent clinical trials in BC, especially those that can limit the enrollment of older patients as well as those with comorbidities and poor performance status. Methods Data on clinical trials in BC were extracted from ClinicalTrials.gov. Co-primary outcomes were proportions of trials with different types of the eligibility criteria. Associations between trial characteristics and the presence of certain types of these criteria (binary variable) were determined with univariate and multivariate logistic regression. Results Our analysis included 522 trials of systemic anticancer treatments started between 2020 and 2022. Upper age limits, strict exclusion criteria pertaining to comorbidities, and those referring to inadequate performance status of the patient were used in 204 (39%), 404 (77%), and 360 (69%) trials, respectively. Overall, 493 trials (94%) had at least one of these criteria. The odds of the presence of each type of the exclusion criteria were significantly associated with investigational site location and trial phase. We also showed that the odds of the upper age limits and the exclusion criteria involving the performance status were significantly higher in the cohort of recent trials compared with cohort of 309 trials started between 2010 and 2012 (39% vs 19% and 69% vs 46%, respectively; p  < 0.001 for univariate and multivariate analysis in both comparisons). The proportion of trials with strict exclusion criteria was comparable between the two cohorts ( p  > 0.05). Only three of recent trials (1%) enrolled solely patients aged 65 or 70 and older. Conclusions Many recent clinical trials in BC exclude large groups of patients, especially older adults, individuals with different comorbidities, and those with poor performance status. Careful modification of some of the eligibility criteria in these trials should be considered to allow investigators to assess the benefits and harms of investigational treatments in participants with characteristics typically encountered in clinical practice.

Overly restrictive clinical trial eligibility criteria can reduce generalizability, slow enrollment, and disproportionately exclude historically underrepresented populations. The eligibility criteria for 196 Alzheimer’s Disease and Related Dementias (AD/ADRD) trials funded by the National Institute on Aging were analyzed to identify common criteria and their potential to disproportionately exclude participants by race/ethnicity. The trials were categorized by type (48 Phase I/II pharmacological, 7 Phase III/IV pharmacological, 128 non-pharmacological, 7 diagnostic, and 6 neuropsychiatric) and target population (51 AD/ADRD, 58 Mild Cognitive Impairment, 25 at-risk, and 62 cognitively normal). Eligibility criteria were coded into the following categories: Medical, Neurologic, Psychiatric, and Procedural. A literature search was conducted to describe the prevalence of disparities for eligibility criteria for African Americans/Black (AA/B), Hispanic/Latino (H/L), American Indian/Alaska Native (AI/AN) and Native Hawaiian/Pacific Islander (NH/PI) populations. The trials had a median of 15 criteria. The most frequent criterion were age cutoffs (87% of trials), specified neurologic (65%), and psychiatric disorders (61%). Underrepresented groups could be disproportionately excluded by 16 eligibility categories; 42% of trials specified English-speakers only in their criteria. Most trials (82%) contain poorly operationalized criteria (i.e., criteria not well defined that can have multiple interpretations/means of implementation) and criteria that may reduce racial/ethnic enrollment diversity.

Identifying appropriate clinical trials for cancer patients with specific gene mutations remains a significant challenge, largely due to limitations in current search tools like ClinicalTrials.gov, which at times return irrelevant or misleading results. This diagnostic accuracy study investigates the efficacy of 2 large language models (LLMs), GPT-4.0 and Gemini 2.0, in evaluating the eligibility of patients with specific cancer-related gene mutations for clinical trials. The study prompts GPT 4.0 and Gemini 2.0 with trial details from ClinicalTrials.gov and a particular cancer mutation. We then assess model performance against physician-curated benchmarks across 6 gene mutations (ALK, BRAF, EGFR, ERBB2, KIT, and KRAS). The results demonstrate good 1-scores for both LLMs-averaging 64% for GPT-4.0 and 70% for Gemini 2.0-highlighting their potential to streamline clinical trial matching. Furthermore, decision trees provided interpretability by identifying key textual indicators that LLMs use. This work demonstrates the feasibility of using proprietary LLMs such as GPT 4.0 and Gemini 2.0 \"off the shelf\" with both limited LLM fine-tuning and limited patient information to evaluate clinical trial eligibility.

The CheckMate 649 trial demonstrated the clinical benefit of nivolumab plus chemotherapy in patients with human epidermal growth factor receptor 2 (HER2)-negative advanced gastric cancer. However, the background discrepancy between clinical practice and randomized controlled trials may impact the therapeutic strategy and prognosis of patients. This study aimed to assess the clinical significance of eligibility criteria determined by the CheckMate 649 trial and identify prognostic factors in clinical practice. A total of 160 patients with HER2-negative metastatic gastric cancer who underwent chemotherapy were retrospectively enrolled and classified into two groups based on eligibility criteria. Among the 160 patients, 76 (47.5%) and 84 (52.5%) were included in the eligible and ineligible groups, respectively. The ineligible group had a significantly lower induction of second- and third-line chemotherapy than the eligible group ( P  = 0.02 and P  = 0.02, respectively). The median overall survival in the eligible and ineligible groups was 22.9 and 10.5 months, respectively, with the ineligible group having a significantly poorer prognosis than the eligible group ( P  < 0.01). Responders to first-line chemotherapy had a better prognosis than non-responders in both groups ( P  = 0.01 and P  < 0.01, respectively). Multivariate analyses identified disease control as an independent prognostic factor in both groups ( P  < 0.01 and P  < 0.01, respectively). Patients with a poor general condition who fulfilled the ineligibility criteria as determined using randomized controlled trials are included in clinical practice, and these criteria are related to tumor response and prognosis.