Bethan Davies

Background: Empirical evidence on the health impacts of brownfield land – previously developed land that may be contaminated – remains limited, mainly due to poorly defined exposures and pathways. Here we assessed associations between an area-level contamination hazard score and emergency hospital admissions across England, using the score as a proxy for the potential contamination associated with brownfield land. Methods: We used Lower-layer Super Output Area (LSOA)-level population-weighted contamination hazard scores in Bayesian hierarchical models to examine the risk of emergency hospital admission (age- and sex-standardised, 2010–2019) with a primary diagnosis of i) all causes (excluding external causes), ii) respiratory diseases, and iii) cardiovascular diseases (CVD). Models were adjusted for deprivation, population density, tobacco expenditure, green space, urbanicity, and ethnicity, and included spatially structured and unstructured random effects to account for residual spatial autocorrelation between neighbouring areas. Stratified analyses assessed potential effect modification by deprivation, urban/rural classification, and region. Results: We included 42,875,231 admissions for all natural causes, 6,657,050 for respiratory diseases, and 4,393,461 for CVD across 32,844 LSOAs in England. LSOAs with higher contamination hazard scores showed a non-linear decreasing trend in relative risk (RR) for all health outcomes. Spatial variation was evident, with stronger negative associations in urban and the least deprived LSOAs, while rural and more deprived LSOAs showed weaker associations or greater uncertainty. Region-specific analyses demonstrated inverse associations for all outcomes in London and for all natural causes and respiratory admissions in the North West and East of England. Other regions showed no consistent pattern across outcomes. Conclusion: Higher contamination hazard scores were associated with lower emergency hospital admission risks. This suggests that areas with a greater legacy of potentially contaminating land uses do not necessarily experience poorer contemporary population health. The inverse associations may partly reflect subsequent remediation and urban regeneration, although the underlying mechanisms require further investigation. Highlights: A national brownfield study used contamination hazard scores as an exposure proxy. Higher contamination hazard scores were linked to lower emergency admission risks. Inverse associations were strongest in urban areas, least deprived areas and London. Lower risks may reflect redevelopment, not a protective effect of brownfield land.

This study provides a unique deterministic modelling analysis of changes in air pollution in London between 2018 and 2023, during which the phased implementation of the Ultra Low Emission Zone (ULEZ) and associated policies for road vehicles coincided with the abrupt and substantial disruption to activity caused by COVID-19. Using counterfactual emissions scenarios within a combined chemical transport and road traffic dispersion model, we estimate the separate effects of local government policy and COVID-19 on ambient concentrations of nitrogen dioxide (NO2), PM10 and PM2.5 across the full period of ULEZ implementation and COVID-related restrictions. The methodology has broader international relevance for cities seeking to determine air pollution impacts of rapidly changing emissions conditions. Results show a sustained benefit of local transport policies on NO2 concentrations, the central area showing differences of between 2.18 and 3.21 µg.m-3 at urban background across the six-year period, and larger areas of the city showing improvements in NO2 concentrations as the scheme expanded. Impacts on PM were small at well under 0.5 µg.m-3 across all areas and years. The COVID-19 pandemic had a larger, shorter-term impact on both NO2 and PM; most notably, PM2.5 was reduced in central London by 1.52 µg.m-3 in 2020 at the height of COVID restrictions, but 5.25 µg.m-3 in 2021 when local restrictions were slightly more relaxed, as concentrations were influenced heavily by meteorology-dependent long-range transport. Taken together this demonstrates how local measures can effectively reduce NO2 to a large extent, but controlling PM levels requires coordinated international measures.

The cognitive mechanisms underlying post-acute COVID-19 sequelae (‘Long COVID’) remain poorly defined. While deficits in memory and executive function are well-documented, it is unclear whether these arise from primary neural dysfunction or are secondary consequences of fatigue-related factors. Here, we addressed this in a community cohort (N = 24 095) using a custom cognitive battery to dissect the interplay between sustained attention, fatigue, and memory. Among this subset, 11 925 participants had previously tested positive for COVID-19, and 548 reported ongoing symptoms consistent with Long COVID. Path analysis (structural equation modelling) was used to evaluate competing mechanistic models linking COVID-19-related features (e.g. duration of symptoms, severity of illness) with objective performance and subjective cognitive symptoms. Participants with persistent symptoms post-COVID-19 (unresolved beyond 12 weeks) showed significant deficits in sustained attention compared to non-infected individuals. Path models revealed complex relationships underlying these deficits. Unlike models of mental health disorders and general ageing, where cognitive performance was largely mediated by relationships to fatigue, the final Long COVID model represented a hybrid mechanism: COVID-19 features were associated with direct, independent effects on attention and memory, while simultaneously relating lower cognitive performance to on-task fatigue build-up. Findings also suggest that ‘brain fog’ reflects a distinct neurocognitive vulnerability associated with fatigue and reduced performance on sustained attention and memory tasks. These findings accord with the hypothesis that Long COVID has a heterogeneous, multi-faceted impact on cognitive function, with higher-order processes such as memory being particularly vulnerable due to their dependence on foundational attention and motivation processes.

Objectives: To quantify the clinical burden and healthcare resource use (HCRU) associated with difficult-to-treat (D2T) rheumatoid arthritis (RA) compared with non-D2T RA and evaluate clinical characteristics and factors associated with D2T RA as well as treatment patterns in Japan. Design: Longitudinal, retrospective observational study using electronic medical record data from 1992 to 2024 and a physician survey administered between November 2022 and November 2025. Setting: A single academic medical centre real-world clinical practice setting in Japan. Participants: A total of 1581 patients with RA were included. Primary and secondary outcome measures: The primary outcome measures were the comparisons of clinical burden and HCRU in the D2T RA group compared with the non-D2T RA group. D2T RA status was defined according to the European Alliance of Associations for Rheumatology definition. Clinical burden included disease activity, physical function, patient-reported outcomes, serological markers and renal function. HCRU included outpatient visits and hospitalisations evaluated over 12-month, 24-month and 36 month windows pre-index using overlap weighting and regression models. Factors associated with D2T RA status were evaluated using logistic regression. Treatment patterns were reconstructed longitudinally as lines of therapy across biologic and targeted synthetic disease-modifying antirheumatic drugs. Results: Among 1581 patients with RA, 102 (6.5%) met the D2T criteria. Treatment trajectories showed earlier escalation and frequent switching across biologic and targeted synthetic disease-modifying antirheumatic drugs in the D2T RA group. The D2T RA group had modestly elevated outpatient visit rates (rate ratio 1.08–1.09) and were more likely to be hospitalised (OR 4.35–6.12) than the non-D2T RA group. Factors associated with D2T RA included RA onset at age <50 (adjusted OR versus ≥50 years: 2.01, 95% CI 1.19 to 3.39), female sex (2.38, 95% CI 0.95 to 5.94) and pulmonary comorbidities (2.82, 95% CI 1.38 to 5.74). Conclusions: In this real-world Japanese cohort, D2T RA was associated with higher disease burden and HCRU, particularly hospital-based care. Earlier identification and targeted management of D2T RA may help reduce the overall clinical and healthcare burden in RA.

We used the September 2013 age-based rollout of the live-attenuated shingles vaccine in England as a natural experiment to estimate the effect of vaccine eligibility on shingles and dementia diagnoses in linked hospital records. Individuals born just before and after the eligibility cutoff were compared using regression discontinuity methods, with follow-up for up to eight years after programme introduction. Eligibility was associated with a clear reduction in hospital-coded shingles diagnoses (RD estimate - 0.12 percentage points, 95% CI -0.153 to -0.079; p = 5.9 × 10 ^-10 ), but there was no evidence of a corresponding reduction in hospital-coded dementia diagnoses (RD estimate -0.06 percentage points, 95% CI -0.40 to 0.27; p = 0.72). Results were robust across denominator definitions, diagnostic-code specifications, estimator choice, placebo cutoffs, and negative-control analyses. The dementia estimate was also close to null in an independently conducted analysis using a separately held HES extract. Comparator analyses in Welsh data with linked primary care and death data did not suggest these results were driven by our reliance on hospital data. These findings do not support a detectable intention-to-treat effect of live-attenuated shingles vaccine eligibility on hospital-coded dementia in England.

Major incidents, such as large fires and floods, often trigger a complex series of environmental and health risk assessments involving multiple institutions and authorities. They can also lead to long-lasting concerns from members of the local communities about the possible impact on their health and environment. Here, we use the example of a large fire that occurred in 2000 at a chemical storage site in Sandhurst, Gloucestershire, England, to contrast evidence from official reports and scientific studies with the expectations of members of the local community. We reflect on how the handling of such an incident has evolved over the last two decades and how public involvement can be further improved in the future. Firstly, we present the results of a 20-year follow-up (2001–2020) retrospective small-area ecological epidemiological study with rates of overall cancer incidence, all-cause mortality and hospital admissions for respiratory disease in the exposed area compared with rates in the South West region of England. We also studied the ten years preceding the fire (1991–2000). Secondly, we discuss the limitations of these findings to alleviate the concerns of the local community. Finally, we use this case study to make recommendations about how to better manage this balance between scientific evidence and public concerns for future incidents. In line with earlier reports, the 20-year follow-up study did not identify any major increase in relative risks for cancer or mortality. Although an increase in admissions for respiratory disease was identified, this was also observed before the fire, suggesting that this may be due to other local factors. Despite these findings, members of the local community still believe that some local cases of cancer were caused by the fire. As rigorous as a long-term epidemiological study can be, it cannot determine causality between an incident and individual cases for a given health outcome, nor can it make up for the absence of data not—or only partly—collected after the incident (e.g., individual-level health register). This case study further highlights that engaging with local communities and managing their expectations and planning potential long-term follow-up studies immediately after a major incident should be carefully considered.

Background: Concerns remain about the potential harmful health impact of nuclear installations. Historical clusters of leukaemia and non-Hodgkin’s lymphoma (LNHL) in children living near Sellafield and Dounreay installations in Great Britain remain aetiologically unexplained, and the sites remain under surveillance. We assess the risk of LNHL, central nervous system (CNS) and all solid tumours in children aged 0–14 years living within 25 km of nuclear installations in Great Britain, between 1995 and 2016. Methods: We used a Poisson regression model to estimate the expected number of cases of each cancer type at the community-level in the study population, we present standardized incidence ratios compared to the national population. We used a hierarchical Poisson regression model to estimate the adjusted incidence rate ratios for each cancer type by distance between the community of residence and nearest nuclear installation. Results: We found no evidence of elevated incidence of LNHL, CNS, or all solid tumours in children resident in communities in proximity to nuclear sites. Within the 25-km zone, there was no evidence of an increased risk of childhood cancer in communities closer to installations. Conclusion: In post-1994 data, there was no evidence of an increased risk of childhood cancers in communities within 25 km of nuclear installations in Great Britain. Previously raised risks are no longer evident.

Background: Syndromic management is widely used to treat symptomatic sexually transmitted infections in settings without aetiologic diagnostics. However, underlying aetiologies and consequent treatment suitability are uncertain without regular assessment. This systematic review estimated the distribution, trends, and determinants of aetiologies for vaginal discharge, urethral discharge, and genital ulcer in sub-Saharan Africa (SSA). Methods and findings: We searched Embase, MEDLINE, Global Health, Web of Science, and grey literature from inception until December 20, 2023, for observational studies reporting aetiologic diagnoses among symptomatic populations in SSA. We adjusted observations for diagnostic test performance, used generalised linear mixed-effects meta-regressions to generate estimates, and critically appraised studies using an adapted Joanna Briggs Institute checklist. Of 4,418 identified records, 206 reports were included from 190 studies in 32 countries conducted between 1969 and 2022. In 2015, estimated primary aetiologies for vaginal discharge were candidiasis (69.4% [95% confidence interval (CI): 44.3% to 86.6%], n = 50), bacterial vaginosis (50.0% [95% CI: 32.3% to 67.8%], n = 39), chlamydia (16.2% [95% CI: 8.6% to 28.5%], n = 50), and trichomoniasis (12.9% [95% CI: 7.7% to 20.7%], n = 80); for urethral discharge were gonorrhoea (77.1% [95% CI: 68.1% to 84.1%], n = 68) and chlamydia (21.9% [95% CI: 15.4% to 30.3%], n = 48); and for genital ulcer were herpes simplex virus type 2 (HSV-2) (48.3% [95% CI: 32.9% to 64.1%], n = 47) and syphilis (9.3% [95% CI: 6.4% to 13.4%], n = 117). Temporal variation was substantial, particularly for genital ulcer where HSV-2 replaced chancroid as the primary cause. Aetiologic distributions for each symptom were largely the same across regions and population strata, despite HIV status and age being significantly associated with several infection diagnoses. Limitations of the review include the absence of studies in 16 of 48 SSA countries, substantial heterogeneity in study observations, and impeded assessment of this variability due to incomplete or inconsistent reporting across studies. Conclusions: In our study, syndrome aetiologies in SSA aligned with World Health Organization guidelines without strong evidence of geographic or demographic variation, supporting broad guideline applicability. Temporal changes underscore the importance of regular aetiologic re-assessment for effective syndromic management. PROSPERO number: CRD42022348045.

Introduction: Syndromic management is widely used to treat symptomatic sexually transmitted infections in settings lacking aetiologic diagnostics. However, heterogeneity in underlying aetiologies and consequent treatment suitability are uncertain without regular assessment. This systematic review characterised aetiologies for vaginal discharge, urethral discharge, and genital ulcer in sub-Saharan Africa (SSA). Methods: We searched Embase, MEDLINE, Global Health, and Web of Science until July 25, 2022, and grey literature until August 31, 2022, for studies reporting aetiologic diagnoses among symptomatic populations in SSA. We adjusted observations for diagnostic test performance and used generalised linear mixed-effects meta-regressions to estimate aetiologic distributions, trends, and determinants. Results: Of 4136 identified records, 198 reports were included from 183 studies in 32 countries between 1969 and 2022. In 2015, primary aetiologies for vaginal discharge were candidiasis (69.4% [95% CI:44.1-86.6%], n=50), bacterial vaginosis (50.0% [32.3-67.8%], n=39), chlamydia (16.5% [8.7-29.0%], n=49), and trichomoniasis (12.9% [7.7-20.7%], n=78); for urethral discharge were gonorrhoea (78.8% [70.9-85.1%], n=67) and chlamydia (22.2% [16.0-30.1%], n=48); and for genital ulcer were HSV-2 (56.1% [39.2-71.6%], n=46) and syphilis (7.8% [5.3-11.4%], n=115). Regional variation was marginal. Temporal variation was substantial, particularly for genital ulcer. For each symptom, HIV-status and age were significantly associated with infection diagnoses, although aetiologic hierarchies were largely the same by strata. Conclusion: Syndrome aetiologies in SSA align with WHO guidelines without strong evidence of contextual or demographic variation, supporting broad guideline applicability. Temporal changes underscore the need for aetiologic re-assessment. STI surveillance using syndrome-based assessments is noncomprehensive and requires studies among symptomatic and asymptomatic populations.