BACKGROUND. Viral load (VL) surrogate endpoints transformed development of HIV and hepatitis C therapeutics. Surrogate endpoints for CMV-related morbidity and mortality could advance development of antiviral treatments. Although observational data support using CMV VL as a trial endpoint, randomized controlled trials (RCTs) demonstrating direct associations between virological markers and clinical endpoints are lacking. METHODS. We performed CMV DNA PCR on frozen serum samples from the only placebo-controlled RCT of ganciclovir for early treatment of CMV after hematopoietic cell transplantation (HCT). We used established criteria to assess VL kinetics as surrogates for CMV disease or death by weeks 8, 24, and 48 after randomization and quantified antiviral effects captured by each marker. We used ensemble-based machine learning to assess the predictive ability of VL kinetics and performed this analysis on a ganciclovir prophylaxis RCT for validation. RESULTS. VL suppression with ganciclovir reduced cumulative incidence of CMV disease and death for 20 years after HCT. Mean VL, peak VL, and change in VL during the first 5 weeks of treatment fulfilled the Prentice definition for surrogacy, capturing more than 95% of ganciclovir’s effect, and yielded highly sensitive and specific predictions by week 48. In the prophylaxis trial, the viral shedding rate satisfied the Prentice definition for CMV disease by week 24. CONCLUSIONS. Our results support using CMV VL kinetics as surrogates for CMV disease, provide a framework for developing CMV preventative and therapeutic agents, and support reductions in VL as the mechanism through which antivirals reduce CMV disease. FUNDING. Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc.
29 March 2024 in International Journal of Infectious Diseases
To evaluate the clinical and economic outcomes in adults hospitalized with invasive pneumococcal disease (IPD) and noninvasive all-cause pneumonia (ACP) overall and by antimicrobial resistance (AMR) status. Hospitalized adults from the BD Insights Research Database with an ICD10 code for IPD, noninvasive ACP or a positive Streptococcus pneumoniae culture/urine antigen test were included. Descriptive statistics and multivariable analyses were used to evaluate outcomes (in-hospital mortality, length of stay [LOS], cost per admission, and hospital margin [costs - payments]). The study included 88,182 adult patients at 90 US hospitals (October 2015-February 2020). Most (98.6%) had noninvasive ACP and 40.2% were <65 years old. Of 1450 culture-positive patients, 37.7% had an isolate resistant to ≥1 antibiotic class. Observed mortality, median LOS, cost per admission, and hospital margins were 8.3%, 6 days, $9791, and $11, respectively. Risk factors for mortality included ≥50 years of age, higher risk of pneumococcal disease (based on chronic or immunocompromising conditions), and intensive care unit admission. Patients with IPD had similar mortality rates and hospital margins compared with noninvasive ACP, but greater costs per admission and LOS. IPD and noninvasive ACP are associated with substantial clinical and economic burden across all adult age groups. Expanded pneumococcal vaccination programs may help reduce disease burden and decrease hospital costs.
Background: Systematic literature review (SLR), a robust method to identify and summarize evidence from published sources, is considered to be a complex, time-consuming, labor-intensive, and expensive task. Objective: This study aimed to present a solution based on natural language processing (NLP) that accelerates and streamlines the SLR process for observational studies using real-world data. Methods: We followed an agile software development and iterative software engineering methodology to build a customized intelligent end-to-end living NLP-assisted solution for observational SLR tasks. Multiple machine learning–based NLP algorithms were adopted to automate article screening and data element extraction processes. The NLP prediction results can be further reviewed and verified by domain experts, following the human-in-the-loop design. The system integrates explainable articificial intelligence to provide evidence for NLP algorithms and add transparency to extracted literature data elements. The system was developed based on 3 existing SLR projects of observational studies, including the epidemiology studies of human papillomavirus–associated diseases, the disease burden of pneumococcal diseases, and cost-effectiveness studies on pneumococcal vaccines. Results: Our Intelligent SLR Platform covers major SLR steps, including study protocol setting, literature retrieval, abstract screening, full-text screening, data element extraction from full-text articles, results summary, and data visualization. The NLP algorithms achieved accuracy scores of 0.86-0.90 on article screening tasks (framed as text classification tasks) and macroaverage F1 scores of 0.57-0.89 on data element extraction tasks (framed as named entity recognition tasks). Conclusions: Cutting-edge NLP algorithms expedite SLR for observational studies, thus allowing scientists to have more time to focus on the quality of data and the synthesis of evidence in observational studies. Aligning the living SLR concept, the system has the potential to update literature data and enable scientists to easily stay current with the literature related to observational studies prospectively and continuously.
Background: Adult pneumococcal disease (PD) represents a significant clinical and economic burden in the United States. Individuals with immunocompromising conditions and other chronic medical conditions, as well as those ≥65 years of age, have an increased risk of acute PD and its long-term complications. The aim of the current study was to describe the health care resource use and direct health care costs associated with invasive PD (IPD) and non-bacteremic pneumococcal pneumonia (NBPP) among adults in the United States, stratified by age group and health-based risk level.Methods: This was a retrospective study of administrative claims from the Merative™ MarketScan® Commercial Database from 2017 to 2019. The study population comprised individuals ≥18 years of age with ≥1 episode of IPD (with hospitalization) or NBPP (with or without hospitalization) during the study period. The study outcomes were the PD-associated health care resource use (outpatient visits and length of any hospitalizations, in days) and direct health care costs per episode.Results: The average health care resource use and direct costs associated with PD were significantly higher for IPD (mean [95% CI] overall cost $49,481 [$45,803–53,159] per episode; N = 949 affected individuals) than for NBPP with hospitalization ($27,330 [$23,807–30,852] per episode; N = 389) and NBPP without hospitalization ($1,090 [$927–1,252] per episode; N = 1,951). For IPD and for NBPP without hospitalization, the direct costs of treatment were significantly higher among groups with immunocompromising or other relevant comorbidities. The costs associated with NBPP without hospitalization were also significantly higher in the oldest age group (≥65 years).Conclusion: Targeting PD prevention efforts to high-risk groups based on age and/or health risk level could decrease the clinical and economic burden of adult PD in the US.
The 23-valent pneumococcal polysaccharide vaccine (PPSV23) targets 23 common serotypes and is recommended for use in adults in various countries to protect against pneumococcal infection. Test-negative design (TND) studies aim to include cases and controls from the same healthcare facilities; however, design choices or limitations associated with conducting real-world research can affect the study results. Here, we highlight how some methodological limitations may have affected results and conclusions of a published study described by Chandler et al.
Background: Automated systematic literature review (SLR) may reduce the workload and errors associated with manual review, enabling faster, up-to-date reviews even with increasing publication volumes. Large language models (LLMs) have demonstrated strong capabilities in understanding unstructured languages. However, few studies have explored the potential of a comprehensive LLM platform to streamline the entire SLR process from article screening to data extraction. Objective: This study aimed to investigate the feasibility of applying an LLM-based system to assist with SLR development. Methods: We developed the Intelligent Systematic Literature Review (ISLaR 2.0) platform, powered by an LLM, and applied it to a use case of the economic burden of pneumococcal disease (PD) literature. First, we established the inclusion and exclusion criteria for the SLR. Second, we defined data elements related to economic burden and domain knowledge, along with guidelines for applying these definitions. Finally, we used the criteria and data element specifications to develop LLM prompts for screening and data extraction. For data extraction, we identified relevant study characteristics and economic burden outcomes. We evaluated ISLaR 2.0’s performance against a gold standard of 50 expert-curated PD articles, using standard metrics (accuracy, precision, recall, and F1 -score). We also conducted a qualitative analysis to describe errors made by the system. Results: ISLaR 2.0 performed well in abstract and full-text screening ( F1 -scores of 86.27 for abstract screening and 87.18 for full-text screening) and data extraction from text ( F1 -scores of 92.83 for study details and 79.76 for economic burden outcomes). The F1 -score for data extraction of tabular economic burden outcome data was 94.83. The qualitative analysis revealed 2 main challenges in extracting economic burden details: misclassification of cost categories and failure to extract relevant information. Conclusions: ISLaR 2.0 enabled efficient execution of an SLR regarding the economic burden of PD. The platform allowed users to flexibly define and modify criteria and data elements, supporting its use across a broad range of health research topics.