Methotrexate (MTX) effectively controls rheumatoid arthritis (RA) but often leads to side effects (SE) such as gastrointestinal (GI) issues, liver toxicity and bone marrow suppression. To develop clinically interpretable machine learning (ML) models that accurately predict MTX-related SE in patients with RA taking MTX. The aim was to enhance predictive accuracy and to identify patient-specific risk factors using explainable artificial intelligence (XAI), thereby enabling transparent clinical interpretation. We specifically sought to address the unmet need for individualised risk stratification using real-world, multicentre observational data.

Retrospective case-control study.

Across 23 rheumatology clinics in South Korea, based on data from a nationwide multicentre cohort.

A total of 5077 patients with RA were initially enrolled from the Korean Observational Study Network for Arthritis. After excluding those with missing clinical, demographic or prescription data and those not receiving MTX, 2375 patients remained eligible. Among these, 1654 and 1218 patients were included in the overall SE and GI SE analysis groups, respectively, after 1:1 propensity score matching. All patients were aged ≥18 years and met the 1987 American College of Rheumatology classification criteria.

The primary outcome was the presence of SE in patients with RA taking MTX, categorised into overall SE and GI SE, based on standardised patient questionnaires and clinical assessments. The secondary outcome was the identification of key predictors using SHapley Additive exPlanations (SHAP) to enhance the interpretability of ML predictions.

Among six ML classifiers, extreme gradient boosting demonstrated the highest performance in predicting overall SE (area under the curve (AUC) 0.781, F1 score 0.672, area under the precision-recall curve (AUPRC) 0.757) and GI SE (AUC 0.701, F1 score 0.690, AUPRC 0.670). SHAP analysis identified key predictive features including age, physician visual analogue scale score, alanine aminotransferase, Health Assessment Questionnaire score, celecoxib use and drug adherence. Logistic regression confirmed statistical significance for multiple variables (eg, OR 4.63; 95% CI 1.41 to 20.90 for non-adherence >30 days; OR 1.45; 95% CI 1.14 to 1.85 for celecoxib use). DeLong’s test indicated that boosting models significantly outperformed support vector machine (p

Interpretable ML models using real-world clinical data can accurately predict SE in patients with RA taking MTX. These models may facilitate early identification of high-risk individuals and inform personalised treatment strategies. Integration into clinical decision support systems could improve MTX safety monitoring. Further prospective validation in external cohorts is warranted.

Our study investigated the age-adjusted incidence rates of non-fatal overdoses by HIV status and sex, and examined trends over time.

We used data from the Comparative Outcomes and Service Utilization Trends study, a population-based cohort study that includes clinical and administrative health data on virtually all people with HIV (PWH) and a 10% random sample of people without HIV in the province.

British Columbia, Canada.

Between April 2012 and March 2020, 11 050 PWH (81.8% male) and 473 952 people without HIV (50.3% male) who were 19 years and older contributed 68 035 and 3 285 824 person years (PY) of follow-up, respectively.

The primary outcome was age-adjusted incidence rates of non-fatal overdose events stratified by sex and HIV status. Trends over time were also assessed.

Age-adjusted non-fatal overdose incidence rates among males with and without HIV were 36.4 and 3.12 per 1000 PY, respectively (incidence rate ratio (IRR) = 11.7, 95% CI 10.9 to 12.5). For females with and without HIV, the age-adjusted incidence rates were 61.4 and 2.33 per 1000 PY, respectively (IRR=26.3, 95% CI 24.0 to 28.7). Between 2013 and 2019 (calendar years with full-year data), the age-adjusted non-fatal overdose rate increased significantly among males and females without HIV but not among PWH.

We observed a significantly higher non-fatal overdose rate among PWH compared to people without HIV. The rate was highest among females with HIV. These findings underline the need for policies and programmes oriented towards PWH to mitigate overdoses, especially for females.

Infant feeding practices in the first 2 years of life are linked to long-term weight trajectories. Despite the importance of obesity prevention interventions, there are no randomised controlled trials (RCTs) evaluating early childhood education and care (ECEC) and primary caregiver-targeted interventions on child weight and feeding outcomes.

To assess the efficacy of an 18-month digital health intervention (Tiny Bites) delivered to ECEC services and primary caregivers of children aged 4 to ≤12 months on child age-adjusted and sex-adjusted body mass index-for-age z-score (zBMI) relative to usual care control in the Hunter New England (HNE) region of New South Wales, Australia.

This type 1 hybrid cluster RCT will include up to 60 ECEC services and 540 children/caregiver dyads. The intervention supports ECEC services and caregivers to deliver recommended responsive feeding practices to infants. ECEC services will receive access to an online assessment platform, training and resources, and implementation support. Primary caregivers will receive text messages, monthly e-newsletters, online links and direct communication from ECEC services. We will assess the impact on child zBMI at 18-month follow-up. Secondary outcomes include duration of consuming any breastmilk, child diet and caregiver responsive feeding practices. We will also assess ECEC policy and practice implementation related to targeted feeding practices, programme cost effectiveness, adverse effects and engagement with the programme (ECECs and caregivers). For the primary outcome, between-group differences will be assessed for paired data using two-level hierarchical linear regression models.

Ethics approval has been provided by HNE Human Research Ethics Committee (HREC) (2023/ETH01158), Deakin University (2024-202) and University of Newcastle HREC (R-2024-0039). Trial results will be submitted for publication in peer-reviewed journals, presented at scientific conferences locally and internationally and to relevant practice stakeholders.

ACTRN12624000576527.

A substantial number of patients with major depressive disorder experience non-remission despite treatment with psychotherapy and several antidepressant drugs. This has increased the interest in new treatment strategies, including non-invasive brain stimulation methods. This randomised controlled trial examines a new treatment method using transcranial-pulsed electromagnetic fields (T-PEMF) delivered via a MoodHeadBand (MHB). The main study objective is to examine the antidepressant effect of T-PEMF on moderate to severe depression.

A double-blinded, randomised (1:1), sham-controlled trial with 96 participants diagnosed with moderate to severe depression without psychotic symptoms, recruited from Psychiatric Center Copenhagen. Participants receive daily 30 min at-home T-PEMF or sham treatment for 8 weeks with the MHB. Depressive symptomatology is examined using the Inventory of Depressive Symptomatology (Self-Report) (primary outcome) and Hamilton-D-17 Items Rating Scale at baseline and treatment completion. Cognitive functions are examined using a battery of emotion-laden and non-emotion-laden tests at baseline, after 1 week and at treatment completion. The participants fill out the Quick Inventory of Depressive Symptomatology (Self-Report) and sleep logs weekly. Summary statistics, generalised linear models, proc mixed models, mixed model repeated measures and Kaplan-Meier analysis will be applied as appropriate to analyse data on depressive symptoms, cognition and sleep.

The Danish Medicines Agency (CIV-23-01-041987) and the Medical Research Ethics Committee (2215332) have approved the trial. The project concurs with the EU directive for medical devices 2017/745 of 5 April 2017 and the ISO 14155 standard. Participants give written informed consent before any trial-related activities. Results will be published in peer-reviewed journals and presented at relevant conferences.

NCT06005103.

Obesity is a leading risk factor for global morbidity and mortality, associated with significant healthcare costs that exceed US$260 billion annually in the USA. Weight cycling, the repeated pattern of intentional weight loss followed by unintentional regain, can exacerbate obesity-related health complications. This study aimed to assess the health economic consequences of weight cycling in patients with obesity defined by a body mass index (BMI) ≥30 kg/m², comparing ‘weight cyclers’ with ‘non-cyclers’, and evaluating the impact of a high-protein oral nutritional supplement (HP-ONS) as a maintenance strategy following weight loss via glucagon-like peptide-1 receptor agonists (GLP-1RAs).

Lifetime state transition modelling study with monthly cycles to simulate obesity-associated disease progression.

US healthcare system; societal perspective.

Simulated cohort of adult patients with obesity (BMI ≥30 kg/m²), stratified by weight cycling status.

Weight loss via GLP-1RAs with or without HP-ONS for weight maintenance.

Key outcomes included costs per obesity-related event avoided, life years (LYs) gained and quality-adjusted life years (QALYs) gained, calculated from a US societal perspective. Transition probabilities for disease states were derived from meta-analyses and adjusted for weight cycling and other relevant risks. Costs and health utilities were based on published US studies with future costs discounted at 3% per year. Uncertainty was investigated by deterministic and probabilistic sensitivity analyses.

Non-cyclers experienced 0.090 fewer obesity-associated events, gained 0.602 LYs and achieved 0.518 QALYs compared with cyclers, resulting in total cost savings of approximately US$4592 per patient. In the second scenario, the combination of GLP-1RA treatment and HP-ONS for weight maintenance yielded effective health outcomes with a cost-effectiveness ratio of US$24 276 per QALY gained, well within accepted cost-effectiveness thresholds in the USA, ranging from US$100 000 to US$150 000 per QALY gained.

Weight cycling significantly impacts health and economic outcomes for patients with obesity, underscoring the need for effective weight management programmes, including the use of HP-ONS focused on sustained weight maintenance after weight loss to curtail associated risks and costs.

To identify the minimum effective dose of a multi-behaviour change technique (BCT) intervention to increase physical activity among individuals on primary statin therapy using the time-to-event continual reassessment method (TiTE-CRM).

A large New York metropolitan area healthcare system comprising approximately 85 000 employees and 5.5 million patient encounters annually.

42 participants enrolled in 13 cohorts of 3 participants, 1 cohort of 2 participants and 1 cohort of 1 participant. The sample was composed of 16.7% individuals aged 66 and older (n=7), 64.3% women (n=27), 69.0% white individuals (n=29) and 7.1% Hispanic individuals (n=3).

A variable-duration, four-BCT text message intervention and a 2-week follow-up. Dose assignment relied on TiTE-CRM to adjust the duration of the intervention based on adherence of participants in prior cohorts. Five mechanisms of action (MoAs) were assessed: self-efficacy, intrinsic regulation, discrepancy in behaviour, motivation and barriers to activity.

The primary outcome measure was the proportion of participants who achieved a 2000 step/day increase between baseline and follow-up. The secondary outcomes were within-participant changes in daily steps (examined as a continuous variable at the daily level) and potential MoAs for increased physical activity.

Of the 40 participants who completed follow-up, 7 (17.5%) achieved the goal of 2000 or more steps per day during their follow-up period. Though participants did increase the number of steps they walked during the intervention (B(SE)=373.1 (154.7) steps; p=0.016), there was no association between increased intervention duration and increased daily average steps. The intervention was also associated with increases in self-efficacy (p=0.002), intrinsic regulation (p=0.037), discrepancy in behaviour (p

The results of this trial did not show a traditional dose-response curve to increasing the length of a multicomponent BCT intervention. Results did show that the intervention successfully increased steps during the intervention period and that the benefit of the intervention dwindled during follow-up. Further, potential MoAs for the intervention were confirmed.

NCT05273723.

Our objectives were (1) to characterise the age-sex-standardised prevalence of comorbidities among people living with HIV (PLWH) and people not living with HIV (PnLWH) between 2001 and 2019 and (2) to examine the effect of comorbidities on direct healthcare costs among PLWH and PnLWH.

This was a retrospective, matched cohort study conducted with the Comparative Outcomes and Service Utilisation Trends (COAST) cohort, which contained all known PLWH in British Columbia (BC), Canada and a general population sample.

BC, Canada.

A total of 9554 PLWH and 47 770 PnLWH from the COAST cohort were followed between 2001 and 2019. Participants were at least 19 years old and 82% male in both groups.

The primary outcomes were the age-sex-standardised prevalence of 16 comorbidities, calculated annually, among PLWH and PnLWH. Secondary outcomes included direct healthcare costs associated with each comorbidity among PLWH and PnLWH. Outcomes were ascertained from administrative health databases.

PLWH exhibited a higher age-sex-standardised prevalence of most comorbidities compared with PnLWH over the study period. Relative disparities in liver and kidney diseases markedly decreased since 2008. Disparities in the prevalence of mental health disorders and substance use disorder (SUD) were consistently large throughout the study period. Comorbidities were associated with high healthcare costs, especially among PLWH.

This study underscores the persistent and evolving burden of non-AIDS-defining comorbidities among PLWH, even in the context of improved HIV management. The high prevalence of mental health disorders and SUD, coupled with the substantial healthcare costs associated with these conditions, emphasises the need for holistic and integrated care models that address the full spectrum of health challenges faced by PLWH.