Kavita Kothari

Objectives: To develop, calibrate, and evaluate a machine learning (ML) classifier designed to reduce citation screening workload for COVID-19 research.Methods: Citations from the WHO COVID-19 Research Database were used to train a logistic regression classifier, which was then calibrated using another independent dataset. A lower threshold was set, below which, records could be excluded from manual screening, and an upper threshold, above which, records could be automatically included into the database. The classifier was then validated on eight COVID-19 reviews to assess workload reduction and recall.Findings: The WHO-Cochrane-EPPI classifier was calibrated for 99% recall, with an upper threshold of 70% and a lower threshold of 39%. During validation, it achieved 98.6% recall across eight reviews. A smaller set of three reviews estimated a 12.5% workload reduction, with 1,390 of 11,153 records below the lower threshold. For database use, workload reduction was estimated at 80%, with 8,955 out of 11,153 records being automatically screened.Conclusion: The WHO-Cochrane-UCL classifier significantly reduces manual screening workload with minimal risk of missing studies. It serves as a model for integrating machine learning into curated resources like the WHO COVID-19 Research Database, improving sustainability by reducing manual screening efforts. The classifier can also be adapted for systematic reviews on COVID-19 and other topics.

Background: Acute viral respiratory infections (AVRIs) rank among the most common causes of hospitalisation worldwide, imposing significant healthcare burdens and driving the development of pharmacological treatments. However, inconsistent outcome reporting across clinical trials limits evidence synthesis and its translation into clinical practice. A core outcome set (COS) for pharmacological treatments in hospitalised adults with AVRIs is essential to standardise trial outcomes and improve research comparability. Objective: To develop an internationally agreed COS for pharmacological treatments in hospitalised adults ≥18 years with acute viral respiratory infections (COSAVRI) through stakeholder agreement. Methods: This protocol follows a four-stage development process in accordance with Core Outcome Set Handbook guidelines. Stage 1 comprises a rapid scoping review of randomised controlled trials (2015–2025) to systematically catalogue patient-relevant outcomes reported in pharmacological AVRI treatment studies. Semi-automated screening and data extraction will employ machine learning and large language models, with human verification. Stage 2 involves an online Real-Time Delphi survey with international stakeholders, including healthcare professionals, researchers, patients/caregivers, and policymakers, to prioritise identified outcomes using a 9-point scale. Stage 3 consists of structured online consensus meetings utilising anonymous electronic voting to finalise the COS. Stage 4 focuses on dissemination and implementation through academic publications, conferences, and stakeholder engagement. Expected outcomes: COSAVRI will provide a standardised minimum set of outcomes for measuring and reporting in future pharmacological trials involving hospitalised adults with AVRIs. This initiative will enhance evidence synthesis, reduce research waste, support regulatory decision-making, and improve pandemic preparedness by facilitating the rapid deployment of harmonised outcomes in trial protocols.

Chlorine spraying was widely used during filovirus outbreaks, but concerns about occupational health risks led to a shift toward wiping. This systematic review aimed to evaluate the health risks associated with exposure to disinfectants among healthcare workers (HCWs), with a specific focus on chlorine-based products and spraying compared to alternative disinfectants and general disinfection tasks (GDTs). PubMed, Embase, and Scopus were searched from inception to March 2025. Eligible studies included observational or experimental research on HCWs exposed to chemical disinfectants. Two reviewers independently screened studies, assessed the risk of bias using a validated occupational health tool, and evaluated evidence certainty with the GRADE approach. Meta-analyses used fixed- and random-effects models; heterogeneity was assessed with I² statistics. Out of 7154 records, 29 studies were included. Most studies were cross-sectional with a high bias risk. Odds ratios (ORs) were calculated using non-exposed groups as reference. Significant associations with respiratory conditions were found for chlorine-based products (OR 1.71), glutaraldehyde (OR 1.44), spraying (OR 2.25), and GDTs (OR 2.20). Exposure to chlorine-based products, glutaraldehyde, spraying, and GDTs likely increases respiratory risk in HCWs, as supported by moderate-certainty evidence. These findings support prioritizing safer disinfectants and strengthening protective measures over banning specific application methods.

Background: Vaccination for coronavirus disease 2019 (COVID-19) has demonstrated reduced risk of hospitalisation and death against more recent variants of COVID-19. Some studies suggested improvements in patients with post-COVID-19 condition (PCC) following vaccination. We systematically explored available evidence on the effect of COVID-19 vaccines on the treatment of people with PCC. Methods: We conducted a rapid review of the literature following systematic approaches. We searched Medline (OVID), EMBASE (Elsevier), ClinicalTrials.gov, and the International Clinical Trials Registry Platform (ICTRP) for randomised trials, non-randomised trials, controlled before-after studies, and interrupted time-series studies of the effect of COVID-19 vaccines for treating people with PCC. Two independent review authors screened citations. Two review authors extracted data independently. We had planned to assess the risk of bias and use the GRADE approach (Grading of Recommendations, Assessment, Development, and Evaluation) to assess the certainty of evidence if there were completed studies. Results: We identified two ongoing randomised controlled trials. Both trials examine the effectiveness of therapeutic vaccines on PCC. The anticipated completion date of the CIMAvax-EGFA trial is January 2023, and the completion date of the COVID-19 mRNA vaccine trial is not stated. Conclusions: There is currently an absence of high‐quality evidence evaluating the effectiveness of COVID-19 vaccines for treating people with post-COVID-19 condition. The absence of published studies and only two ongoing trials highlight the need for additional studies on the effectiveness of vaccines for PCC. We recommend that researchers consider PCC as per the definition provided by the World Health Organization and use the available core outcome set for PCC in deciding which outcomes to measure and report in the trials. PROPSERO registration: CRD42022330821 (20/06/2022)

Background: Vaccination for coronavirus disease 2019 (COVID-19) has demonstrated reduced risk of hospitalisation and death against more recent variants of COVID-19. Some studies suggested improvements in patients with post-COVID-19 condition (PCC) following vaccination. We systematically explored available evidence on the effect of COVID-19 vaccines for the treatment of people with PCC. Methods: We conducted a rapid review of the literature following systematic approaches. We searched Medline (OVID), EMBASE (Elsevier), ClinicalTrials.gov, and the International Clinical Trials Registry Platform (ICTRP) for randomised trials, non-randomised trials, controlled before-after studies, and interrupted time-series studies of the effect of COVID-19 vaccines for treating people with PCC. Two independent review authors screened citations. Two review authors extracted data independently. We had planned to assess the risk of bias and use the GRADE approach (Grading of Recommendations, Assessment, Development, and Evaluation) to assess the certainty of evidence if there were completed studies. Results: We identified two ongoing randomised controlled trials. Both trials examine the effectiveness of therapeutic vaccines on PCC. The anticipated completion date of the CIMAvax-EGFA trial is January 2023, and the completion date of the COVID-19 mRNA vaccine trial is not stated. Conclusions: There is currently an absence of high‐quality evidence evaluating the effectiveness of COVID-19 vaccines for treating people with post-COVID-19 condition. The absence of published studies and only two ongoing trials highlight the need for additional studies on the effectiveness of vaccines for PCC. We recommend that researchers consider PCC as per the definition provided by the World Health Organization and use the available core outcome set for PCC in deciding which outcomes to measure and report in the trials. PROPSERO registration: CRD42022330821 (20/06/2022)

Introduction This research project addresses the lack of screening tools for the early detection of high-risk individuals for long-term care, through four individual studies. Study 1 investigates the predictive ability of the ‘Kihon Check List’, study 2 the ‘Cognitive Function instrument’ and EuroQol-5 Dimension (EQ-5D) and study 3 the ‘Cognitive Function instrument’ and EQ-5D as well as the ‘Frail Kenshin’ health check-up, for incident long-term care certification over a follow-up period of up to 4 years. This is the first large prospective study to evaluate the predictive ability of these tools for the outcome measure long-term care certification. The last subsection of this project study four aims to explore a mixed methods intervention for delaying the need for long-term care. This section is purely exploratory, looking for clues for further studies. Methods and analysis Baseline data have been collected through local government programs, as well as through postal self-reported questionnaires. The primary outcome variable for all studies is long-term care certification data. Statistical analysis will be carried out using Kaplan-Meier, Multiple Cox regression as well as logistic regression. Conclusion This project hopes to identify tools effective in predicting long-term care need. This will enable identification of citizens that are of higher risk for long-term care in the near future. This subset of high-risk individuals can in the future be addressed for extra support/intervention. Ethics and dissemination All studies have been approved by respective institutional ethical committees and the WHO ethical committee ERC.0002899. In addition, all studies conform to the provisions of the Declaration of Helsinki and are conducted in accordance with Japan’s ‘Ethical Guidelines for Medical and Health Research Involving Human Subjects’. All findings will be disseminated at conferences and published in peer-reviewed journals. Trial registration number UMIN000023283.

Aim: The present population‐based study investigated the predictive ability of the Kihon Checklist (a self‐reported frailty questionnaire) and the cognitive domain therein for incident long‐term care need certification. This is the first large population‐based study to investigate an association between the Kihon Checklist and the outcome measure, long‐term care need certification.Methods: The study population consisted of community‐dwelling citizens aged >65 years who responded to the Kihon Checklist in Kobe City. The Kihon Checklist is a simple 26‐item questionnaire to identify frail citizens, including three items (Q18‐20) on subjective cognitive function (the cognitive domain).Results: A total of 182 099 citizens were included for statistical analysis. The overall incidence of long‐term care need certification was 1.6%, 3.5% and 5.4% at 1, 2 and 3 years. Associations were found between the Kihon Checklist questions and long‐term care need certification. Furthermore, each unfavorable answer on the cognitive domain was associated with the risk for long‐term care need (HR 1.493 for Q18, 1.285 for Q19, 1.321 for Q20, all P < 0.0001), independent of age, sex and other items on the Kihon Checklist. Also, after 3 years, as the number of unfavorable answers to the cognitive domain increased from 0 to 1, 2 and 3, the incidence of long‐term care need certification progressively increased from 3.5% to 6.4%, 12.6% and 29.6%.Conclusions: The Kihon Checklist, especially the cognitive domain therein, appears to be predictive of long‐term care need in community‐dwelling citizens, suggesting the potential utility of the questionnaire for early detection of high‐risk individuals. Geriatr Gerontol Int 2019; 19: 598–603.