To develop and evaluate an explainable machine learning framework enhanced with synthetic data generation to predict unplanned 30-day hospital readmissions among patients with chronic obstructive pulmonary disease (COPD), heart failure (HF) and type 2 diabetes mellitus (T2DM), and to identify key clinical and social predictors of readmission.

A retrospective cohort study using electronic health record data incorporating both structured variables and information extracted from unstructured clinical notes. Synthetic data were generated using advanced resampling and deep learning-based techniques to address outcome imbalance and improve model training.

Intensive care unit and general ward admissions at a single tertiary academic medical centre included in the MIMIC-IV (Medical Information Mart for Intensive Care IV) database.

Adult patients (≥18 years) were admitted with a primary diagnosis of COPD (n=14 050), HF (n=7097) or T2DM (n=12 735) between 2008 and 2019, with complete 30-day follow-up and no in-hospital mortality during the index admission.

The primary outcome was unplanned all-cause hospital readmission within 30-days of discharge. Predictors were drawn from six domains, including demographics, comorbidities, clinical acuity, therapies, behavioural factors and care continuity. Predictive performance was evaluated using multiple machine learning methods and fivefold cross-validation, with model interpretability assessed using established goal and local explanation approaches.

Ensemble-based machine learning models demonstrated the strongest predictive performance across all three disease cohorts. Key predictors of readmission included higher illness severity, greater comorbidity burden, medication non-adherence, gaps in preventive care and limited social support. Models incorporating synthetic data augmentation showed improved discrimination compared with models trained on original data alone.

An explainable synthetic-data driven framework incorporating clinical, behavioural and social data can support prediction of 30-day readmissions among patients with common chronic conditions using routinely available electronic health record data.

To determine the personal, National Health Service and wider societal resource use in relation to caring responsibilities for carers of people living with non-memory led dementias (NMLDs); and to design a resource use measure (RUM) that can be delivered in the Better Living with Non-memory-led Dementia (BELIDE) randomised controlled trial, part of the Rare Dementia (RD) - TALK research programme.

The first stage was to identify and review any existing RUMs that could be used or adapted to the trial population and setting. If no measures were identified, the second stage was initial informal discussions with healthcare professionals (HCPs) and the programme patient and public involvement representatives to inform the perspective, settings of care and main resource items to develop a new RUM. In the third stage, a first draft of the RUM was tested for content and face validity in a modified Delphi study comprising HCPs and carers. The measure was revised and, in the final stage, piloted in the first 3 months of the BELIDE trial to assess acceptability and feasibility of collecting the economic outcomes and the completeness of data collection. The key drivers of resource use and costs were assessed, and appropriate face validity checks were applied to ensure accurate description of the treatment pathways.

Carers and family of people living with NMLD recruited from Rare Dementia Support members in the UK, and a broad range of HCPs with experience of working with people who have NMLD to capture the different dimensions of experience, grade and skill mix.

In total, 20 people participated in the modified Delphi study, 11 HCPs and 9 carers. Rare Dementia Support groups and 1:1 calls were highly rated, as were general practitioner appointments. The greatest consensus was in the productivity and carer tasks; all caring tasks were highly rated. Healthcare practitioners rated healthcare items as higher importance than carers themselves.

Unpaid carers and HCPs are the experts in the resource impact of caring for someone with NMLD and have been underserved in research to date. This research, as part of preparatory stages of the BELIDE trial, has enabled the timely development of a comprehensive and valid RUM for unpaid carers of people with NMLD.

CRD42022356943.

NCT06241287.

Although multiple studies have offered self-collection for human papillomavirus (HPV)-based cervical screening in community settings, there are no randomised controlled trials (RCTs) that have compared implementation outcomes of programme approaches for self-collection. This trial will compare two such approaches in low-resource settings in the states of Tamil Nadu and Mizoram, India.

A cluster RCT will be conducted over a year, offering self-collection to 3000 women aged 30–49 from 28 clusters (average size 101) in selected districts. Clusters in tribal, rural and urban low-income settings will be randomised to two arms. The intervention arm, co-designed with multiple stakeholders, will involve campaigns to offer self-collection in the community. The comparison arm will be offered self-collection at the nearest health facilities.

HPV-based cervical screening will be performed at central laboratories using clinically validated screening assays that can identify the highest risk carcinogenic HPV types (Group 1a–c - HPV16/18/31/33/45/52/58, ±35). Ablative treatment will be based on positivity with this extended genotyping triage, while those with any of the lower carcinogenic HPV types (Group 1d - 39, 51, 56, 59, ±35, Groups 2a/b - 66, 68) will undergo further assessment with visual inspection with acetic acid. Outcomes will be evaluated quantitatively and qualitatively using RE-AIM and the Theoretical Framework of Acceptability.

The primary outcome will be percentage of women well-managed (screened and appropriately treated) in both arms, with secondary outcomes including proportion screened, proportion treated, acceptability (willingness to screen, rescreen, and/or recommend to others) to women, community and healthcare providers, adoption (by providers), implementation fidelity, costs, sustainability assessment and systematically identified implementation barriers and facilitators. The reach, effectiveness and acceptability of community-based self-collection and the use of extended genotyping for triage in resource-constrained, hard-to-reach populations will be assessed, with lessons that can inform future statewide and national programmes.

Ethics approval has been obtained from the Institutional Review Board (IRB) and Ethics Committee of the Christian Medical College Vellore, Tamil Nadu, India (IRB Min. No 14314; INTERVEN), the Alfred Hospital Ethics Committee (HREC Ref 80134, Local Reference: project 601/21), Melbourne, Australia, the IARC Ethics Committee (IEC 21-32), Lyon, France, the Salem Polyclinic Institutional Ethics Committee (SPCIEC/2022/June/01/02), Tamil Nadu, India and the Institutional Ethics Committee, Civil Hospital, Aizawl, Mizoram, India (No.B.12018/1/13-CHA(A)/IEC/115). The study is also approved by the State Scientific Advisory Committee, Directorate of Public Health and Preventive Medicine, Chennai, Tamil Nadu (R. No. 011575/HEB/A2/2023). The Alfred Hospital Approval, as an authorised Australian ethics committee for national mutual recognition, is recognised and registered with the University of Melbourne Human Research Ethics Committee (2024-25255-57650-1). Written informed consent will be obtained from participants. The results of the trial will be disseminated through a peer-reviewed medical journal, and also through workshops, reports and conferences.

The trial has been registered with the Clinical Trials Registry - India: CTRI/2022/04/042327.

To examine the longitudinal impact of time-varying factors on US youth’s trajectories of initiation and use of e-cigarettes and cigarettes during the transition from adolescence to young adulthood.

Longitudinal.

Nationally representative US survey, the Population Assessment of Tobacco and Health (PATH) Study.

2682 US youth (aged 16–17) at wave (W)1 of the PATH Study across six waves (2013–2020) into young adulthood (aged 22–23).

Unweighted longitudinal latent class analyses identified trajectory classes of e-cigarette and cigarette use, separately. Nationally representative weighted multinomial logistic regression analyses examined time-varying harm perceptions, substance use problems and tobacco product first tried as predictors of these trajectory classes.

Five e-cigarette classes (2013–2020; 41.5% Persistent Never Use, 12.6% W5 Initiation, 19.9% W3 Initiation, 15.2% Prior Initiation, 10.8% High Frequency Past 30-Day (P30D) Use) and five cigarette classes (2013–2019; 58.6% Persistent Never Use, 11.5% W4 Initiation, 10.9% W2 Initiation, 9.6% Prior Initiation, 9.5% High Frequency P30D Use) were identified. Time-varying harm perceptions and substance use problems were associated with trajectories of initiation and use for both products. Cigarettes, cigarillos, other combustibles and any smokeless tobacco as first product tried were associated with e-cigarette initiation and/or progression to high frequency use. E-cigarettes and hookah as first product tried were associated with later cigarette initiation. High Frequency P30D Cigarette Use was less likely if the first product tried was e-cigarettes, cigarillos, hookah or any smokeless tobacco product.

Results reinforce the need for identification and intervention of early substance use among younger adolescents and targeted public health messaging to address changing harm perceptions and prevent initiation among older adolescents.

Suicide poses a significant global health challenge in low- and middle-income countries and is a leading cause of death worldwide. Although global suicide rates have decreased by 36% between 2000 and 2019, the decline in Africa has not been as substantial. Suicide surveillance data, particularly from pesticide ingestion, is lacking.

This scoping review aimed to identify pesticide suicide surveillance methods currently used in African countries and to assess their viability for improving reporting of pesticide suicide deaths.

A scoping review was conducted using a five-stage methodological framework across several databases.

Peer-reviewed articles published in English that investigated (a) pesticide poisoning and (b) suicide in humans only and specific to the African continent.

MEDLINE (via PubMed), Scopus (with substantial Embase content), Web of Science Core Collection, Biological Abstracts, SciELO (on Web of Science platform), Academic Search Premier, Africa-Wide Information, Biological and Agricultural Index, CINAHL, Health Source Nursing/Academic, APA PsycInfo and General Science (EBSCOhost platform). The search was performed in January 2022, and the review of the results took place from February to May 2022.

All authors developed and tested the data extraction tool. The charting framework remained dynamic, continuously updated as reviewers gained a deeper understanding of the studies included. Four independent reviewers charted the data to ensure inter-rater reliability. To address discrepancies, a sample of eligible articles was cross-rated by the reviewers.

We identified 110 relevant studies describing eight different surveillance systems, conducted in 30 different African countries. Of these studies, 50.1% (56/110) reported on the number of pesticide suicides (totalling 1554); however, 49% (54/110) did not differentiate the cause of death further than either a poisoning death or a suicide, making it difficult to determine if pesticides were involved in the death or if the death was an intentional poisoning case. The surveillance systems identified included health facility records (hospital admissions data), poison control centre (PCC) data, forensic/mortuary data (including police reports), media reports, surveys/interviews with patient/family, case reports, systematic/literature reviews and civil registration and vital statistics (CRVS) data. Hospital admissions data was the most frequently used surveillance system (61/110, 55.5%).

The number of studies and surveillance methods found was higher than anticipated. While this is a positive sign, several areas of improvement were identified for pesticide suicide surveillance in Africa. These included improving reporting of the specific pesticides (including the active ingredient) linked to suicide cases for improving policy (since this is required for pesticide regulation) and making use of more than one surveillance system to enhance surveillance of pesticide suicides, allowing for under-used sources, such as PCCs, to be used more effectively in pesticide suicide surveillance. Finally, although a comparison of these surveillance methods outside of Africa was not directly possible for each surveillance system due to a lack of similar high-quality reviews, we did refer to publications where similar pesticide suicide surveillance systems were discussed.

There is an absence of real-world evidence, especially from low- and middle-income countries (LMICs), on the implementation successes and challenges of COVID-19 Test and Treat (T&T) programmes. In 2022, nirmatrelvir/ritonavir was provided as standard of care for mild to moderate COVID-19 treatment in eight LMICs (Ghana, Kenya, Laos, Malawi, Nigeria, Rwanda, Uganda and Zambia). This manuscript describes a research protocol to study novel drug introduction during the COVID-19 health emergency, with implications and learnings for future pandemic preparedness. The goal of the study is to provide simultaneous programme learnings and improvements with programme rollout, to fill a gap in real-world implementation data on T&T programmes of oral antiviral treatment for COVID-19 and inform programme implementation and scale-up in other LMICs.

This multiple methods implementation research study is divided into three components to address key operational research objectives: (1) programme learnings, monitoring and evaluation; (2) patient-level programme impact; and (3) key stakeholder perspectives. Data collection will occur for a minimum of 6 months in each country up to the end of grant. Quantitative data will be analysed using descriptive statistics for each country and then aggregated across the programme countries. Stakeholder perspectives will be examined using the Consolidated Framework for Implementation Research implementation science framework and semistructured interviews.

This study was approved by the Duke University Institutional Review Board (Pro00111388). The study was also approved by the local institutional review boards in each country participating in individual-level data collection (objectives 2 and 3): Ghana, Malawi, Rwanda, Nigeria and Zambia. The study’s findings will be published in peer-reviewed journals and disseminated through dialogue events, national and international conferences and through social media.

NCT06360783.

Carers of people with non-memory-led dementias such as posterior cortical atrophy (PCA), primary progressive aphasia (PPA) and behavioural variant frontotemporal dementia (bvFTD) face unique challenges. Yet, little evidence-based support and guidance are available for this population. To address this gap in services, we have developed a novel, web-based educational programme: the Better Living with Non-memory-led Dementia programme (BELIDE). BELIDE was co-designed with people with lived experience of non-memory-led dementia and a previous pilot study confirmed its feasibility as an online intervention. This protocol outlines the randomised controlled trial (RCT) to evaluate the clinical and cost-effectiveness of BELIDE.

This is a parallel-group, single-blind, RCT of 238 unpaid caregivers of people diagnosed with PCA, PPA or bvFTD recruited internationally among members of the UK-based organisation Rare Dementia Support. The intervention (BELIDE programme) consists of six structured online educational modules tailored to each phenotype, a virtual onboarding session, real-life practice tasks and up to two follow-up facilitation sessions. The group receiving the intervention will be given access to the programme, while the control group will receive treatment as usual and be placed on a wait-list to receive access to the programme once they complete their participation in the trial. The allocation ratio will be 1:1 stratified by dementia diagnosis and gender. The primary outcome is reduction in caregiver depressive symptoms. Secondary outcomes include stress, anxiety, self-efficacy, quality of life and caregiver-patient relationship quality. Data will be collected online via Qualtrics surveys at baseline, 8 weeks and 6 months post-randomisation. A mixed-method process evaluation with a subgroup of intervention participants will explore barriers and facilitators for engagement. A health economics evaluation will also be conducted to assess cost-effectiveness. If effective, this programme could improve access to caregiver support for non-memory-led dementias by providing scalable, tailored education.

Ethical approval has been granted by University College London Research Ethics Committee (8545/007). The results will be disseminated via peer-reviewed publications, conferences, stakeholder events and open-access resources.

This trial has been registered prospectively on the Clinical Trials registry, first posted on 5 February 2024 under registration number NCT06241287.