Blood-based epigenome-wide analyses of chronic low-grade inflammation across diverse population cohorts

Chronic inflammation is a hallmark of age-related disease states. The effectiveness of inflammatory proteins including C-reactive protein (CRP) in assessing long-term inflammation is hindered by their phasic nature. DNA methylation (DNAm) signatures of CRP may act as more reliable markers of chronic inflammation. We show that inter-individual differences in DNAm capture 50% of the variance in circulating CRP (N = 17,936, Generation Scotland). We develop a series of DNAm predictors of CRP using state-of-the-art algorithms. An elastic-net-regression-based predictor outperformed competing methods and explained 18% of phenotypic variance in the Lothian Birth Cohort of 1936 (LBC1936) cohort, doubling that of existing DNAm predictors. DNAm predictors performed comparably in four additional test cohorts (Avon Longitudinal Study of Parents and Children, Health for Life in Singapore, Southall and Brent Revisited, and LBC1921), including for individuals of diverse genetic ancestry and different age groups. The best-performing predictor surpassed assay-measured CRP and a genetic score in its associations with 26 health outcomes. Our findings forge new avenues for assessing chronic low-grade inflammation in diverse populations.

Ethnicity data resource in population-wide health records: completeness, coverage and granularity of diversity

Intersectional social determinants including ethnicity are vital in health research. We curated a population-wide data resource of self-identified ethnicity data from over 60 million individuals in England primary care, linking it to hospital records. We assessed ethnicity data in terms of completeness, consistency, and granularity and found one in ten individuals do not have ethnicity information recorded in primary care. By linking to hospital records, ethnicity data were completed for 94% of individuals. By reconciling SNOMED-CT concepts and census-level categories into a consistent hierarchy, we organised more than 250 ethnicity sub-groups including and beyond "White", "Black", "Asian", "Mixed" and "Other, and found them to be distributed in proportions similar to the general population. This large observational dataset presents an algorithmic hierarchy to represent self-identified ethnicity data collected across heterogeneous healthcare settings. Accurate and easily accessible ethnicity data can lead to a better understanding of population diversity, which is important to address disparities and influence policy recommendations that can translate into better, fairer health for all.

Undervaccination and severe COVID-19 outcomes: meta-analysis of national cohort studies in England, Northern Ireland, Scotland, and Wales

BACKGROUND

Undervaccination (receiving fewer than the recommended number of SARS-CoV-2 vaccine doses) could be associated with increased risk of severe COVID-19 outcomes-ie, COVID-19 hospitalisation or death-compared with full vaccination (receiving the recommended number of SARS-CoV-2 vaccine doses). We sought to determine the factors associated with undervaccination, and to investigate the risk of severe COVID-19 outcomes in people who were undervaccinated in each UK nation and across the UK.

METHODS

We used anonymised, harmonised electronic health record data with whole population coverage to carry out cohort studies in England, Northern Ireland, Scotland, and Wales. Participants were required to be at least 5 years of age to be included in the cohorts. We estimated adjusted odds ratios for undervaccination as of June 1, 2022. We also estimated adjusted hazard ratios (aHRs) for severe COVID-19 outcomes during the period June 1 to Sept 30, 2022, with undervaccination as a time-dependent exposure. We combined results from nation-specific analyses in a UK-wide fixed-effect meta-analysis. We estimated the reduction in severe COVID-19 outcomes associated with a counterfactual scenario in which everyone in the UK was fully vaccinated on June 1, 2022.

FINDINGS

The numbers of people undervaccinated on June 1, 2022 were 26 985 570 (45·8%) of 58 967 360 in England, 938 420 (49·8%) of 1 885 670 in Northern Ireland, 1 709 786 (34·2%) of 4 992 498 in Scotland, and 773 850 (32·8%) of 2 358 740 in Wales. People who were younger, from more deprived backgrounds, of non-White ethnicity, or had a lower number of comorbidities were less likely to be fully vaccinated. There was a total of 40 393 severe COVID-19 outcomes in the cohorts, with 14 156 of these in undervaccinated participants. We estimated the reduction in severe COVID-19 outcomes in the UK over 4 months of follow-up associated with a counterfactual scenario in which everyone was fully vaccinated on June 1, 2022 as 210 (95% CI 94-326) in the 5-15 years age group, 1544 (1399-1689) in those aged 16-74 years, and 5426 (5340-5512) in those aged 75 years or older. aHRs for severe COVID-19 outcomes in the meta-analysis for the age group of 75 years or older were 2·70 (2·61-2·78) for one dose fewer than recommended, 3·13 (2·93-3·34) for two fewer, 3·61 (3·13-4·17) for three fewer, and 3·08 (2·89-3·29) for four fewer.

INTERPRETATION

Rates of undervaccination against COVID-19 ranged from 32·8% to 49·8% across the four UK nations in summer, 2022. Undervaccination was associated with an elevated risk of severe COVID-19 outcomes.

FUNDING

UK Research and Innovation National Core Studies: Data and Connectivity.

Understanding the genetic complexity of puberty timing across the allele frequency spectrum

Pubertal timing varies considerably and has been associated with a range of health outcomes in later life. To elucidate the underlying biological mechanisms, we performed multi-ancestry genetic analyses in ~800,000 women, identifying 1,080 independent signals associated with age at menarche. Collectively these loci explained 11% of the trait variance in an independent sample, with women at the top and bottom 1% of polygenic risk exhibiting a ~11 and ~14-fold higher risk of delayed and precocious pubertal development, respectively. These common variant analyses were supported by exome sequence analysis of ~220,000 women, identifying several genes, including rare loss of function variants in ZNF483 which abolished the impact of polygenic risk. Next, we implicated 660 genes in pubertal development using a combination of in silico variant-to-gene mapping approaches and integration with dynamic gene expression data from mouse embryonic GnRH neurons. This included an uncharacterized G-protein coupled receptor GPR83, which we demonstrate amplifies signaling of MC3R, a key sensor of nutritional status. Finally, we identified several genes, including ovary-expressed genes involved in DNA damage response that co-localize with signals associated with menopause timing, leading us to hypothesize that the ovarian reserve might signal centrally to trigger puberty. Collectively these findings extend our understanding of the biological complexity of puberty timing and highlight body size dependent and independent mechanisms that potentially link reproductive timing to later life disease.

Refining epigenetic prediction of chronological and biological age

BACKGROUND

Epigenetic clocks can track both chronological age (cAge) and biological age (bAge). The latter is typically defined by physiological biomarkers and risk of adverse health outcomes, including all-cause mortality. As cohort sample sizes increase, estimates of cAge and bAge become more precise. Here, we aim to develop accurate epigenetic predictors of cAge and bAge, whilst improving our understanding of their epigenomic architecture.

METHODS

First, we perform large-scale (N = 18,413) epigenome-wide association studies (EWAS) of chronological age and all-cause mortality. Next, to create a cAge predictor, we use methylation data from 24,674 participants from the Generation Scotland study, the Lothian Birth Cohorts (LBC) of 1921 and 1936, and 8 other cohorts with publicly available data. In addition, we train a predictor of time to all-cause mortality as a proxy for bAge using the Generation Scotland cohort (1214 observed deaths). For this purpose, we use epigenetic surrogates (EpiScores) for 109 plasma proteins and the 8 component parts of GrimAge, one of the current best epigenetic predictors of survival. We test this bAge predictor in four external cohorts (LBC1921, LBC1936, the Framingham Heart Study and the Women's Health Initiative study).

RESULTS

Through the inclusion of linear and non-linear age-CpG associations from the EWAS, feature pre-selection in advance of elastic net regression, and a leave-one-cohort-out (LOCO) cross-validation framework, we obtain cAge prediction with a median absolute error equal to 2.3 years. Our bAge predictor was found to slightly outperform GrimAge in terms of the strength of its association to survival (HRGrimAge = 1.47 [1.40, 1.54] with p = 1.08 × 10-52, and HRbAge = 1.52 [1.44, 1.59] with p = 2.20 × 10-60). Finally, we introduce MethylBrowsR, an online tool to visualise epigenome-wide CpG-age associations.

CONCLUSIONS

The integration of multiple large datasets, EpiScores, non-linear DNAm effects, and new approaches to feature selection has facilitated improvements to the blood-based epigenetic prediction of biological and chronological age.

Charting a Course for Smartphones and Wearables to Transform Population Health Research

The use of data from smartphones and wearable devices has huge potential for population health research, given the high level of device ownership; the range of novel health-relevant data types available from consumer devices; and the frequency and duration with which data are, or could be, collected. Yet, the uptake and success of large-scale mobile health research in the last decade have not met this intensely promoted opportunity. We make the argument that digital person-generated health data are required and necessary to answer many top priority research questions, using illustrative examples taken from the James Lind Alliance Priority Setting Partnerships. We then summarize the findings from 2 UK initiatives that considered the challenges and possible solutions for what needs to be done and how such solutions can be implemented to realize the future opportunities of digital person-generated health data for clinically important population health research. Examples of important areas that must be addressed to advance the field include digital inequality and possible selection bias; easy access for researchers to the appropriate data collection tools, including how best to harmonize data items; analysis methodologies for time series data; patient and public involvement and engagement methods for optimizing recruitment, retention, and public trust; and methods for providing research participants with greater control over their data. There is also a major opportunity, provided through the linkage of digital person-generated health data to routinely collected data, to support novel population health research, bringing together clinician-reported and patient-reported measures. We recognize that well-conducted studies need a wide range of diverse challenges to be skillfully addressed in unison (eg, challenges regarding epidemiology, data science and biostatistics, psychometrics, behavioral and social science, software engineering, user interface design, information governance, data management, and patient and public involvement and engagement). Consequently, progress would be accelerated by the establishment of a new interdisciplinary community where all relevant and necessary skills are brought together to allow for excellence throughout the life cycle of a research study. This will require a partnership of diverse people, methods, and technologies. If done right, the synergy of such a partnership has the potential to transform many millions of people's lives for the better.

Harmonising electronic health records for reproducible research: challenges, solutions and recommendations from a UK-wide COVID-19 research collaboration

BACKGROUND

The CVD-COVID-UK consortium was formed to understand the relationship between COVID-19 and cardiovascular diseases through analyses of harmonised electronic health records (EHRs) across the four UK nations. Beyond COVID-19, data harmonisation and common approaches enable analysis within and across independent Trusted Research Environments. Here we describe the reproducible harmonisation method developed using large-scale EHRs in Wales to accommodate the fast and efficient implementation of cross-nation analysis in England and Wales as part of the CVD-COVID-UK programme. We characterise current challenges and share lessons learnt.

METHODS

Serving the scope and scalability of multiple study protocols, we used linked, anonymised individual-level EHR, demographic and administrative data held within the SAIL Databank for the population of Wales. The harmonisation method was implemented as a four-layer reproducible process, starting from raw data in the first layer. Then each of the layers two to four is framed by, but not limited to, the characterised challenges and lessons learnt. We achieved curated data as part of our second layer, followed by extracting phenotyped data in the third layer. We captured any project-specific requirements in the fourth layer.

RESULTS

Using the implemented four-layer harmonisation method, we retrieved approximately 100 health-related variables for the 3.2 million individuals in Wales, which are harmonised with corresponding variables for > 56 million individuals in England. We processed 13 data sources into the first layer of our harmonisation method: five of these are updated daily or weekly, and the rest at various frequencies providing sufficient data flow updates for frequent capturing of up-to-date demographic, administrative and clinical information.

CONCLUSIONS

We implemented an efficient, transparent, scalable, and reproducible harmonisation method that enables multi-nation collaborative research. With a current focus on COVID-19 and its relationship with cardiovascular outcomes, the harmonised data has supported a wide range of research activities across the UK.

The impact of the COVID-19 pandemic on cardiovascular disease prevention and management

How the Coronavirus Disease 2019 (COVID-19) pandemic has affected prevention and management of cardiovascular disease (CVD) is not fully understood. In this study, we used medication data as a proxy for CVD management using routinely collected, de-identified, individual-level data comprising 1.32 billion records of community-dispensed CVD medications from England, Scotland and Wales between April 2018 and July 2021. Here we describe monthly counts of prevalent and incident medications dispensed, as well as percentage changes compared to the previous year, for several CVD-related indications, focusing on hypertension, hypercholesterolemia and diabetes. We observed a decline in the dispensing of antihypertensive medications between March 2020 and July 2021, with 491,306 fewer individuals initiating treatment than expected. This decline was predicted to result in 13,662 additional CVD events, including 2,281 cases of myocardial infarction and 3,474 cases of stroke, should individuals remain untreated over their lifecourse. Incident use of lipid-lowering medications decreased by 16,744 patients per month during the first half of 2021 as compared to 2019. By contrast, incident use of medications to treat type 2 diabetes mellitus, other than insulin, increased by approximately 623 patients per month for the same time period. In light of these results, methods to identify and treat individuals who have missed treatment for CVD risk factors and remain undiagnosed are urgently required to avoid large numbers of excess future CVD events, an indirect impact of the COVID-19 pandemic.

Using national electronic health records for pandemic preparedness: validation of a parsimonious model for predicting excess deaths among those with COVID-19-a data-driven retrospective cohort study

OBJECTIVES

To use national, pre- and post-pandemic electronic health records (EHR) to develop and validate a scenario-based model incorporating baseline mortality risk, infection rate (IR) and relative risk (RR) of death for prediction of excess deaths.

DESIGN

An EHR-based, retrospective cohort study.

SETTING

Linked EHR in Clinical Practice Research Datalink (CPRD); and linked EHR and COVID-19 data in England provided in NHS Digital Trusted Research Environment (TRE).

PARTICIPANTS

In the development (CPRD) and validation (TRE) cohorts, we included 3.8 million and 35.1 million individuals aged ≥30 years, respectively.

MAIN OUTCOME MEASURES

One-year all-cause excess deaths related to COVID-19 from March 2020 to March 2021.

RESULTS

From 1 March 2020 to 1 March 2021, there were 127,020 observed excess deaths. Observed RR was 4.34% (95% CI, 4.31-4.38) and IR was 6.27% (95% CI, 6.26-6.28). In the validation cohort, predicted one-year excess deaths were 100,338 compared with the observed 127,020 deaths with a ratio of predicted to observed excess deaths of 0.79.

CONCLUSIONS

We show that a simple, parsimonious model incorporating baseline mortality risk, one-year IR and RR of the pandemic can be used for scenario-based prediction of excess deaths in the early stages of a pandemic. Our analyses show that EHR could inform pandemic planning and surveillance, despite limited use in emergency preparedness to date. Although infection dynamics are important in the prediction of mortality, future models should take greater account of underlying conditions.

A survey on clinical natural language processing in the United Kingdom from 2007 to 2022

Much of the knowledge and information needed for enabling high-quality clinical research is stored in free-text format. Natural language processing (NLP) has been used to extract information from these sources at scale for several decades. This paper aims to present a comprehensive review of clinical NLP for the past 15 years in the UK to identify the community, depict its evolution, analyse methodologies and applications, and identify the main barriers. We collect a dataset of clinical NLP projects (n = 94; £ = 41.97 m) funded by UK funders or the European Union's funding programmes. Additionally, we extract details on 9 funders, 137 organisations, 139 persons and 431 research papers. Networks are created from timestamped data interlinking all entities, and network analysis is subsequently applied to generate insights. 431 publications are identified as part of a literature review, of which 107 are eligible for final analysis. Results show, not surprisingly, clinical NLP in the UK has increased substantially in the last 15 years: the total budget in the period of 2019-2022 was 80 times that of 2007-2010. However, the effort is required to deepen areas such as disease (sub-)phenotyping and broaden application domains. There is also a need to improve links between academia and industry and enable deployments in real-world settings for the realisation of clinical NLP's great potential in care delivery. The major barriers include research and development access to hospital data, lack of capable computational resources in the right places, the scarcity of labelled data and barriers to sharing of pretrained models.

Effects of the COVID-19 pandemic on secondary care for cardiovascular disease in the UK: an electronic health record analysis across three countries

BACKGROUND

Although morbidity and mortality from COVID-19 have been widely reported, the indirect effects of the pandemic beyond 2020 on other major diseases and health service activity have not been well described.

METHODS

Analyses used national administrative electronic hospital records in England, Scotland and Wales for 2016-2021. Admissions and procedures during the pandemic (2020-2021) related to six major cardiovascular conditions (acute coronary syndrome, heart failure, stroke/transient ischaemic attack, peripheral arterial disease, aortic aneurysm, and venous thromboembolism) were compared to the annual average in the pre-pandemic period (2016-2019). Differences were assessed by time period and urgency of care.

RESULTS

In 2020, there were 31 064 (-6%) fewer hospital admissions (14 506 [-4%] fewer emergencies, 16 560 [-23%] fewer elective admissions) compared to 2016-2019 for the six major cardiovascular diseases combined. The proportional reduction in admissions was similar in all three countries. Overall, hospital admissions returned to pre-pandemic levels in 2021. Elective admissions remained substantially below expected levels for almost all conditions in all three countries (-10 996 [-15%] fewer admissions). However, these reductions were offset by higher than expected total emergency admissions (+25 878 [+6%] higher admissions), notably for heart failure and stroke in England, and for venous thromboembolism in all three countries. Analyses for procedures showed similar temporal variations to admissions.

CONCLUSION

This study highlights increasing emergency cardiovascular admissions during the pandemic, in the context of a substantial and sustained reduction in elective admissions and procedures. This is likely to increase further the demands on cardiovascular services over the coming years.

NCMP-09. IMPACT OF TUMOUR CHARACTERISTICS AND CANCER TREATMENT ON CEREBROVASCULAR MORTALITY AFTER GLIOMA DIAGNOSIS: EVIDENCE FROM A POPULATION-BASED CANCER REGISTRY

BACKGROUND

Brain tumour patients have the highest stroke mortality rates among cancer types, but the factors associated with fatal stroke remain unknown. We aimed to examine to what extent brain tumour grade, a marker of biological aggressiveness, tumour size and cancer treatment are associated with cerebrovascular mortality among patients with malignant glioma, the most common and aggressive brain tumour.

METHODS

We conducted a retrospective, observational cohort study using the US National Cancer Institute’s state and regional population-based cancer registries (NCI SEER). We identified adult patients with a diagnosis of malignant glioma (2000 to 2018, Nf72,916). The primary outcome of interest was death from cerebrovascular disease. Cox regression modelling estimated the associations with cerebrovascular mortality of tumour grade, tumour size and treatment (surgery, radiotherapy, chemotherapy), calculating hazard ratios (HR) adjusted for these factors as well as for age, sex, race, marital status and calendar year.

RESULTS

Higher grade (Grade IV vs Grade II: HR=2.47, 95% CI=1.69-3.61, p< 0.001) and larger brain tumours (size 3 to < 6 cm: HR=1.40, 95% CI=1.03 -1.89, p< 0.05; size ≥ 6 cm: HR=1.47, 95% CI=1.02-2.13, p< 0.05 compared to size < 3cm) were associated with increased cerebrovascular mortality. Having cancer treatment was associated with decreased risk (surgery: HR= 0.60, p< 0.001; chemotherapy: HR=0.42, p< 0.001; radiation: HR= 0.69, p< 0.05). However, among patents surviving five years or more from their cancer diagnosis radiotherapy was associated with higher risk of cerebrovascular mortality (HR 2.73, 95% CI 1.49-4.99, p< 0.01).

CONCLUSIONS

More aggressive tumour characteristics are associated with increased cerebrovascular mortality, and treatment was associated with lower risk within 5 years of diagnosis. Radiotherapy increased risk of fatal cerebrovascular outcome five-years after cancer diagnosis. Further research is needed to better understand the long-term cardiovascular consequences of radiation therapy, and whether the consequent risk can be mitigated.

Feasibility and ethics of using data from the Scottish newborn blood spot archive for research

Background

Newborn heel prick blood spots are routinely used to screen for inborn errors of metabolism and life-limiting inherited disorders. The potential value of secondary data from newborn blood spot archives merits ethical consideration and assessment of feasibility for public benefit. Early life exposures and behaviours set health trajectories in childhood and later life. The newborn blood spot is potentially well placed to create an unbiased and cost-effective population-level retrospective birth cohort study. Scotland has retained newborn blood spots for all children born since 1965, around 3 million in total. However, a moratorium on research access is currently in place, pending public consultation.

Methods

We conducted a Citizens' Jury as a first step to explore whether research use of newborn blood spots was in the public interest. We also assessed the feasibility and value of extracting research data from dried blood spots for predictive medicine.

Results

Jurors delivered an agreed verdict that conditional research access to the newborn blood spots was in the public interest. The Chief Medical Officer for Scotland authorised restricted lifting of the current research moratorium to allow a feasibility study. Newborn blood spots from consented Generation Scotland volunteers were retrieved and their potential for both epidemiological and biological research demonstrated.

Conclusions

Through the Citizens' Jury, we have begun to identify under what conditions, if any, should researchers in Scotland be granted access to the archive. Through the feasibility study, we have demonstrated the potential value of research access for health data science and predictive medicine.

Frequency and Phenotype Associations of Rare Variants in 5 Monogenic Cerebral Small Vessel Disease Genes in 200,000 UK Biobank Participants

BACKGROUND AND OBJECTIVES

Based on previous case reports and disease-based cohorts, a minority of patients with cerebral small vessel disease (cSVD) have a monogenic cause, with many also manifesting extracerebral phenotypes. We investigated the frequency, penetrance, and phenotype associations of putative pathogenic variants in cSVD genes in the UK Biobank (UKB), a large population-based study.

METHODS

We used a systematic review of previous literature and ClinVar to identify putative pathogenic rare variants in CTSA, TREX1, HTRA1, and COL4A1/2. We mapped phenotypes previously attributed to these variants (phenotypes-of-interest) to disease coding systems used in the UKB's linked health data from UK hospital admissions, death records, and primary care. Among 199,313 exome-sequenced UKB participants, we assessed the following: the proportion of participants carrying ≥1 variant(s); phenotype-of-interest penetrance; and the association between variant carrier status and phenotypes-of-interest using a binary (any phenotype present/absent) and phenotype burden (linear score of the number of phenotypes a participant possessed) approach.

RESULTS

Among UKB participants, 0.5% had ≥1 variant(s) in studied genes. Using hospital admission and death records, 4%-20% of variant carriers per gene had an associated phenotype. This increased to 7%-55% when including primary care records. Only COL4A1 variant carrier status was significantly associated with having ≥1 phenotype-of-interest and a higher phenotype score (OR = 1.29, p = 0.006).

DISCUSSION

While putative pathogenic rare variants in monogenic cSVD genes occur in 1:200 people in the UKB population, only approximately half of variant carriers have a relevant disease phenotype recorded in their linked health data. We could not replicate most previously reported gene-phenotype associations, suggesting lower penetrance rates, overestimated pathogenicity, and/or limited statistical power.

Cardiovascular Events After Primary Malignant and Non-Malignant Brain Tumour Diagnosis: A Population Matched Cohort Study in Wales (United Kingdom)

AIMS

It is uncertain whether the elevated cardiovascular diseases (CVD) risks in patients with brain tumours is due to differences in the distribution of risk factors. We compared CVD risks among patients with a primary brain tumour to a matched general population cohort.

METHOD

Using data from the Secured Anonymised Information Linkage (SAIL) Databank in Wales, we identified adults aged ≥18 years with a diagnosis of brain tumour identified in the cancer registry between 2000-2014, and a matched cohort (case-to-control ratio 1:5) by age, sex and primary care provider from the general population. Outcomes included fatal and non-fatal major vascular events and venous thromboembolism (VTE). We used multivariable Cox models adjusted for clinical risk factors to compare risks.

RESULTS

There were 2,869 and 3,931 people diagnosed with malignant and non-malignant brain tumours, respectively, in Wales. They were matched to 33,785 controls. Within the first year of tumour diagnosis, malignant tumour was associated with VTE (hazard ratio [HR] 21.58, 95% confidence interval 16.12-28.88) and stroke (HR 3.32, 2.44-4.53). People with malignant tumour surviving one year had higher risks of VTE (HR 2.20, 1.52-3.18) and stroke (HR 1.45, 1.00-2.10) compared to their matched controls. Individuals with non-malignant tumours had higher risks of VTE (HR 3.72, 2.73-5.06), stroke (HR 4.06, 3.35-4.93) and aortic and peripheral arterial disease (HR 2.09, 1.26-3.48) within the first year of diagnosis compared with their controls.

CONCLUSION

The elevated risks of CVD suggest risk reduction a potential strategy to improve life quality and survival in people with malignant and non-malignant brain tumour.

Association of COVID-19 With Major Arterial and Venous Thrombotic Diseases: A Population-Wide Cohort Study of 48 Million Adults in England and Wales

BACKGROUND

Infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) induces a prothrombotic state, but long-term effects of COVID-19 on incidence of vascular diseases are unclear.

METHODS

We studied vascular diseases after COVID-19 diagnosis in population-wide anonymized linked English and Welsh electronic health records from January 1 to December 7, 2020. We estimated adjusted hazard ratios comparing the incidence of arterial thromboses and venous thromboembolic events (VTEs) after diagnosis of COVID-19 with the incidence in people without a COVID-19 diagnosis. We conducted subgroup analyses by COVID-19 severity, demographic characteristics, and previous history.

RESULTS

Among 48 million adults, 125 985 were hospitalized and 1 319 789 were not hospitalized within 28 days of COVID-19 diagnosis. In England, there were 260 279 first arterial thromboses and 59 421 first VTEs during 41.6 million person-years of follow-up. Adjusted hazard ratios for first arterial thrombosis after COVID-19 diagnosis compared with no COVID-19 diagnosis declined from 21.7 (95% CI, 21.0-22.4) in week 1 after COVID-19 diagnosis to 1.34 (95% CI, 1.21-1.48) during weeks 27 to 49. Adjusted hazard ratios for first VTE after COVID-19 diagnosis declined from 33.2 (95% CI, 31.3-35.2) in week 1 to 1.80 (95% CI, 1.50-2.17) during weeks 27 to 49. Adjusted hazard ratios were higher, for longer after diagnosis, after hospitalized versus nonhospitalized COVID-19, among Black or Asian versus White people, and among people without versus with a previous event. The estimated whole-population increases in risk of arterial thromboses and VTEs 49 weeks after COVID-19 diagnosis were 0.5% and 0.25%, respectively, corresponding to 7200 and 3500 additional events, respectively, after 1.4 million COVID-19 diagnoses.

CONCLUSIONS

High relative incidence of vascular events soon after COVID-19 diagnosis declines more rapidly for arterial thromboses than VTEs. However, incidence remains elevated up to 49 weeks after COVID-19 diagnosis. These results support policies to prevent severe COVID-19 by means of COVID-19 vaccines, early review after discharge, risk factor control, and use of secondary preventive agents in high-risk patients.

A retrospective cohort study predicting and validating impact of the COVID-19 pandemic in individuals with chronic kidney disease

Chronic kidney disease (CKD) is associated with increased risk of baseline mortality and severe COVID-19, but analyses across CKD stages, and comorbidities are lacking. In prevalent and incident CKD, we investigated comorbidities, baseline risk, COVID-19 incidence, and predicted versus observed one-year excess death. In a national dataset (NHS Digital Trusted Research Environment [NHSD TRE]) for England encompassing 56 million individuals), we conducted a retrospective cohort study (March 2020 to March 2021) for prevalence of comorbidities by incident and prevalent CKD, SARS-CoV-2 infection and mortality. Baseline mortality risk, incidence and outcome of infection by comorbidities, controlling for age, sex and vaccination were assessed. Observed versus predicted one-year mortality at varying population infection rates and pandemic-related relative risks using our published model in pre-pandemic CKD cohorts (NHSD TRE and Clinical Practice Research Datalink [CPRD]) were compared. Among individuals with CKD (prevalent:1,934,585, incident:144,969), comorbidities were common (73.5% and 71.2% with one or more condition[s] in respective data sets, and 13.2% and 11.2% with three or more conditions, in prevalent and incident CKD), and associated with SARS-CoV-2 infection, particularly dialysis/transplantation (odds ratio 2.08, 95% confidence interval 2.04-2.13) and heart failure (1.73, 1.71-1.76), but not cancer (1.01, 1.01-1.04). One-year all-cause mortality varied by age, sex, multi-morbidity and CKD stage. Compared with 34,265 observed excess deaths, in the NHSD-TRE and CPRD databases respectively, we predicted 28,746 and 24,546 deaths (infection rates 10% and relative risks 3.0), and 23,754 and 20,283 deaths (observed infection rates 6.7% and relative risks 3.7). Thus, in this largest, national-level study, individuals with CKD have a high burden of comorbidities and multi-morbidity, and high risk of pre-pandemic and pandemic mortality. Hence, treatment of comorbidities, non-pharmaceutical measures, and vaccination are priorities for people with CKD and management of long-term conditions is important during and beyond the pandemic.

COVID-19 trajectories among 57 million adults in England: a cohort study using electronic health records

BACKGROUND

Updatable estimates of COVID-19 onset, progression, and trajectories underpin pandemic mitigation efforts. To identify and characterise disease trajectories, we aimed to define and validate ten COVID-19 phenotypes from nationwide linked electronic health records (EHR) using an extensible framework.

METHODS

In this cohort study, we used eight linked National Health Service (NHS) datasets for people in England alive on Jan 23, 2020. Data on COVID-19 testing, vaccination, primary and secondary care records, and death registrations were collected until Nov 30, 2021. We defined ten COVID-19 phenotypes reflecting clinically relevant stages of disease severity and encompassing five categories: positive SARS-CoV-2 test, primary care diagnosis, hospital admission, ventilation modality (four phenotypes), and death (three phenotypes). We constructed patient trajectories illustrating transition frequency and duration between phenotypes. Analyses were stratified by pandemic waves and vaccination status.

FINDINGS

Among 57 032 174 individuals included in the cohort, 13 990 423 COVID-19 events were identified in 7 244 925 individuals, equating to an infection rate of 12·7% during the study period. Of 7 244 925 individuals, 460 737 (6·4%) were admitted to hospital and 158 020 (2·2%) died. Of 460 737 individuals who were admitted to hospital, 48 847 (10·6%) were admitted to the intensive care unit (ICU), 69 090 (15·0%) received non-invasive ventilation, and 25 928 (5·6%) received invasive ventilation. Among 384 135 patients who were admitted to hospital but did not require ventilation, mortality was higher in wave 1 (23 485 [30·4%] of 77 202 patients) than wave 2 (44 220 [23·1%] of 191 528 patients), but remained unchanged for patients admitted to the ICU. Mortality was highest among patients who received ventilatory support outside of the ICU in wave 1 (2569 [50·7%] of 5063 patients). 15 486 (9·8%) of 158 020 COVID-19-related deaths occurred within 28 days of the first COVID-19 event without a COVID-19 diagnoses on the death certificate. 10 884 (6·9%) of 158 020 deaths were identified exclusively from mortality data with no previous COVID-19 phenotype recorded. We observed longer patient trajectories in wave 2 than wave 1.

INTERPRETATION

Our analyses illustrate the wide spectrum of disease trajectories as shown by differences in incidence, survival, and clinical pathways. We have provided a modular analytical framework that can be used to monitor the impact of the pandemic and generate evidence of clinical and policy relevance using multiple EHR sources.

FUNDING

British Heart Foundation Data Science Centre, led by Health Data Research UK.

Systematic Review of Cerebral Phenotypes Associated With Monogenic Cerebral Small-Vessel Disease

Background Cerebral small-vessel disease (cSVD) is an important cause of stroke and vascular dementia. Most cases are multifactorial, but an emerging minority have a monogenic cause. While NOTCH3 is the best-known gene, several others have been reported. We aimed to summarize the cerebral phenotypes associated with these more recent cSVD genes. Methods and Results We performed a systematic review (PROSPERO [International Prospective Register of Systematic Reviews]: CRD42020196720), searching Medline/Embase (conception to July 2020) for any language publications describing COL4A1/2, TREX1, HTRA1, ADA2, or CTSA pathogenic variant carriers. We extracted data about individuals' characteristics and clinical and vascular radiological cerebral phenotypes. We summarized phenotype frequencies per gene, comparing patterns across genes. We screened 6485 publications including 402, and extracted data on 390 individuals with COL4A1, 123 with TREX1, 44 with HTRA1 homozygous, 41 with COL4A2, 346 with ADA2, 82 with HTRA1 heterozygous, and 14 with CTSA. Mean age ranged from 15 (ADA2) to 59 years (HTRA1 heterozygotes). Clinical phenotype frequencies varied widely: stroke, 9% (TREX1) to 52% (HTRA1 heterozygotes); cognitive features, 0% (ADA2) to 64% (HTRA1 homozygotes); and psychiatric features, 0% (COL4A2; ADA2) to 57% (CTSA). Among individuals with neuroimaging, vascular radiological phenotypes appeared common, ranging from 62% (ADA2) to 100% (HTRA1 homozygotes; CTSA). White matter lesions were the most common pathology, except in ADA2 and COL4A2 cases, where ischemic and hemorrhagic lesions dominated, respectively. Conclusions There appear to be differences in cerebral manifestations across cSVD genes. Vascular radiological changes were more common than clinical neurological phenotypes, and present in the majority of individuals with reported neuroimaging. However, these results may be affected by age and biases inherent to case reports. In the future, better characterization of associated phenotypes, as well as insights from population-based studies, should improve our understanding of monogenic cSVD to inform genetic testing, guide clinical management, and help unravel underlying disease mechanisms.

Prospective validation of the 4C prognostic models for adults hospitalised with COVID-19 using the ISARIC WHO Clinical Characterisation Protocol

PURPOSE

To prospectively validate two risk scores to predict mortality (4C Mortality) and in-hospital deterioration (4C Deterioration) among adults hospitalised with COVID-19.

METHODS

Prospective observational cohort study of adults (age ≥18 years) with confirmed or highly suspected COVID-19 recruited into the International Severe Acute Respiratory and emerging Infections Consortium (ISARIC) WHO Clinical Characterisation Protocol UK (CCP-UK) study in 306 hospitals across England, Scotland and Wales. Patients were recruited between 27 August 2020 and 17 February 2021, with at least 4 weeks follow-up before final data extraction. The main outcome measures were discrimination and calibration of models for in-hospital deterioration (defined as any requirement of ventilatory support or critical care, or death) and mortality, incorporating predefined subgroups.

RESULTS

76 588 participants were included, of whom 27 352 (37.4%) deteriorated and 12 581 (17.4%) died. Both the 4C Mortality (0.78 (0.77 to 0.78)) and 4C Deterioration scores (pooled C-statistic 0.76 (95% CI 0.75 to 0.77)) demonstrated consistent discrimination across all nine National Health Service regions, with similar performance metrics to the original validation cohorts. Calibration remained stable (4C Mortality: pooled slope 1.09, pooled calibration-in-the-large 0.12; 4C Deterioration: 1.00, -0.04), with no need for temporal recalibration during the second UK pandemic wave of hospital admissions.

CONCLUSION

Both 4C risk stratification models demonstrate consistent performance to predict clinical deterioration and mortality in a large prospective second wave validation cohort of UK patients. Despite recent advances in the treatment and management of adults hospitalised with COVID-19, both scores can continue to inform clinical decision making.

TRIAL REGISTRATION NUMBER

ISRCTN66726260.

TeenCovidLife: a resource to understand the impact of the COVID-19 pandemic on adolescents in Scotland

TeenCovidLife is part of Generation Scotland's CovidLife projects, a set of longitudinal observational studies designed to assess the psychosocial and health impacts of the COVID-19 pandemic. TeenCovidLife focused on how adolescents in Scotland were coping during the pandemic. As of September 2021, Generation Scotland had conducted three TeenCovidLife surveys. Participants from previous surveys were invited to participate in the next, meaning the age ranges shifted over time. TeenCovidLife Survey 1 consists of data from 5,543 young people age 12 to 17, collected from 22 May to 5 July 2020, during the first school closures period in Scotland. TeenCovidLife Survey 2 consists of data from 2,245 young people aged 12 to 18, collected from 18 August to 14 October 2020, when the initial lockdown measures were beginning to ease, and schools reopened in Scotland. TeenCovidLife Survey 3 consists of data from 597 young people age 12 to 19, collected from 12 May to 27 June 2021, a year after the first survey, after the schools returned following the second lockdown in 2021. A total of 316 participants took part in all three surveys. TeenCovidLife collected data on general health and well-being, as well as topics specific to COVID-19, such as adherence to COVID-19 health guidance, feelings about school closures, and the impact of exam cancellations. Limited work has examined the impact of the COVID-19 pandemic on young people. TeenCovidLife provides relevant and timely data to assess the impact of the pandemic on young people in Scotland. The dataset is available under authorised access from Generation Scotland; see the Generation Scotland website for more information.

TeenCovidLife: a resource to understand the impact of the COVID-19 pandemic on adolescents in Scotland

TeenCovidLife is part of Generation Scotland's CovidLife projects, a set of longitudinal observational studies designed to assess the psychosocial and health impacts of the COVID-19 pandemic. TeenCovidLife focused on how adolescents in Scotland were coping during the pandemic. As of September 2021, Generation Scotland had conducted three TeenCovidLife surveys. Participants from previous surveys were invited to participate in the next, meaning the age ranges shifted over time. TeenCovidLife Survey 1 consists of data from 5,543 young people age 12 to 17, collected from 22 May to 5 July 2020, during the first school closures period in Scotland. TeenCovidLife Survey 2 consists of data from 2,245 young people aged 12 to 18, collected from 18 August to 14 October 2020, when the initial lockdown measures were beginning to ease, and schools reopened in Scotland. TeenCovidLife Survey 3 consists of data from 597 young people age 12 to 19, collected from 12 May to 27 June 2021, a year after the first survey, after the schools returned following the second lockdown in 2021. A total of 316 participants took part in all three surveys. TeenCovidLife collected data on general health and well-being, as well as topics specific to COVID-19, such as adherence to COVID-19 health guidance, feelings about school closures, and the impact of exam cancellations. Limited work has examined the impact of the COVID-19 pandemic on young people. TeenCovidLife provides relevant and timely data to assess the impact of the pandemic on young people in Scotland. The dataset is available under authorised access from Generation Scotland; see the Generation Scotland website for more information.

RuralCovidLife: A new resource for the impact of the pandemic on rural Scotland

RuralCovidLife is part of Generation Scotland’s CovidLife project, investigating the impact of the COVID-19 pandemic and mitigation measures on people in Scotland. The RuralCovidLife project focuses on Scotland’s rural communities, and how they have been impacted by the pandemic. During survey development, Generation Scotland consulted with people living or working in rural communities, and collaborated with a patient and public involvement and engagement (PPIE) group composed of rural community leaders. Through this consultation work, the RuralCovidLife survey was developed to assess the issues most pertinent to people in rural communities, such as mental health, employment, transport, connectivity, and local communities. Between 14th October and 30th November 2020, 3,365 participants from rural areas in Scotland took part in the survey. Participant ages ranged from 16 to 96 (mean = 58.4, standard deviation [SD] = 13.3), and the majority of the participants were female (70.5%). Over half (51.3%) had taken part in the original CovidLife survey. RuralCovidLife includes a subsample (n = 523) of participants from the Generation Scotland cohort. Pre-pandemic data on health and lifestyle, as well as biological samples, are available for these participants. These participants’ data can also be linked to past and future healthcare records, allowing analysis of retrospective and prospective health outcomes. Like Generation Scotland, RuralCovidLife is designed as a resource for researchers. RuralCovidLife data, as well as the linked Generation Scotland data, is available for use by external researchers following approval from the Generation Scotland Access Committee. RuralCovidLife can be used to investigate mental health, well-being, and behaviour in participants living in rural areas during the COVID-19 pandemic, as well as comparisons with non-rural samples. Moreover, the sub-sample with full Generation Scotland data and linkage can be used to investigate the long-term health consequences of the COVID-19 pandemic in rural communities.

Reproducible disease phenotyping at scale: Example of coronary artery disease in UK Biobank

IMPORTANCE

A lack of internationally agreed standards for combining available data sources at scale risks inconsistent disease phenotyping limiting research reproducibility.

OBJECTIVE

To develop and then evaluate if a rules-based algorithm can identify coronary artery disease (CAD) sub-phenotypes using electronic health records (EHR) and questionnaire data from UK Biobank (UKB).

DESIGN

Case-control and cohort study.

SETTING

Prospective cohort study of 502K individuals aged 40-69 years recruited between 2006-2010 into the UK Biobank with linked hospitalization and mortality data and genotyping.

PARTICIPANTS

We included all individuals for phenotyping into 6 predefined CAD phenotypes using hospital admission and procedure codes, mortality records and baseline survey data. Of these, 408,470 unrelated individuals of European descent had a polygenic risk score (PRS) for CAD estimated.

EXPOSURE

CAD Phenotypes.

MAIN OUTCOMES AND MEASURES

Association with baseline risk factors, mortality (n = 14,419 over 7.8 years median f/u), and a PRS for CAD.

RESULTS

The algorithm classified individuals with CAD into prevalent MI (n = 4,900); incident MI (n = 4,621), prevalent CAD without MI (n = 10,910), incident CAD without MI (n = 8,668), prevalent self-reported MI (n = 2,754); prevalent self-reported CAD without MI (n = 5,623), yielding 37,476 individuals with any type of CAD. Risk factors were similar across the six CAD phenotypes, except for fewer men in the self-reported CAD without MI group (46.7% v 70.1% for the overall group). In age- and sex- adjusted survival analyses, mortality was highest following incident MI (HR 6.66, 95% CI 6.07-7.31) and lowest for prevalent self-reported CAD without MI at baseline (HR 1.31, 95% CI 1.15-1.50) compared to disease-free controls. There were similar graded associations across the six phenotypes per SD increase in PRS, with the strongest association for prevalent MI (OR 1.50, 95% CI 1.46-1.55) and the weakest for prevalent self-reported CAD without MI (OR 1.08, 95% CI 1.05-1.12). The algorithm is available in the open phenotype HDR UK phenotype library (https://portal.caliberresearch.org/).

CONCLUSIONS

An algorithmic, EHR-based approach distinguished six phenotypes of CAD with distinct survival and PRS associations, supporting adoption of open approaches to help standardize CAD phenotyping and its wider potential value for reproducible research in other conditions.

Physician-Confirmed and Administrative Definitions of Stroke in UK Biobank Reflect the Same Underlying Genetic Trait

Background

Stroke in UK Biobank (UKB) is ascertained via linkages to coded administrative datasets and self-report. We studied the accuracy of these codes using genetic validation.

Methods

We compiled stroke-specific and broad cerebrovascular disease (CVD) code lists (Read V2/V3, ICD-9/-10) for medical settings (hospital, death record, primary care) and self-report. Among 408,210 UKB participants, we identified all with a relevant code, creating 12 stroke definitions based on the code type and source. We performed genome-wide association studies (GWASs) for each definition, comparing summary results against the largest published stroke GWAS (MEGASTROKE), assessing genetic correlations, and replicating 32 stroke-associated loci.

Results

The stroke case numbers identified varied widely from 3,976 (primary care stroke-specific codes) to 19,449 (all codes, all sources). All 12 UKB stroke definitions were significantly correlated with the MEGASTROKE summary GWAS results (rg.81-1) and each other (rg.4-1). However, Bonferroni-corrected confidence intervals were wide, suggesting limited precision of some results. Six previously reported stroke-associated loci were replicated using ≥1 UKB stroke definition.

Conclusions

Stroke case numbers in UKB depend on the code source and type used, with a 5-fold difference in the maximum case-sample size. All stroke definitions are significantly genetically correlated with the largest stroke GWAS to date.

Association of COVID-19 vaccines ChAdOx1 and BNT162b2 with major venous, arterial, or thrombocytopenic events: A population-based cohort study of 46 million adults in England

BACKGROUND

Thromboses in unusual locations after the Coronavirus Disease 2019 (COVID-19) vaccine ChAdOx1-S have been reported, although their frequency with vaccines of different types is uncertain at a population level. The aim of this study was to estimate the population-level risks of hospitalised thrombocytopenia and major arterial and venous thromboses after COVID-19 vaccination.

METHODS AND FINDINGS

In this whole-population cohort study, we analysed linked electronic health records from adults living in England, from 8 December 2020 to 18 March 2021. We estimated incidence rates and hazard ratios (HRs) for major arterial, venous, and thrombocytopenic outcomes 1 to 28 and >28 days after first vaccination dose for ChAdOx1-S and BNT162b2 vaccines. Analyses were performed separately for ages <70 and ≥70 years and adjusted for age, age2, sex, ethnicity, and deprivation. We also prespecified adjustment for anticoagulant medication, combined oral contraceptive medication, hormone replacement therapy medication, history of pulmonary embolism or deep vein thrombosis, and history of coronavirus infection in analyses of venous thrombosis; and diabetes, hypertension, smoking, antiplatelet medication, blood pressure lowering medication, lipid lowering medication, anticoagulant medication, history of stroke, and history of myocardial infarction in analyses of arterial thromboses. We selected further covariates with backward selection. Of 46 million adults, 23 million (51%) were women; 39 million (84%) were <70; and 3.7 million (8.1%) Asian or Asian British, 1.6 million (3.5%) Black or Black British, 36 million (79%) White, 0.7 million (1.5%) mixed ethnicity, and 1.5 million (3.2%) were of another ethnicity. Approximately 21 million (46%) adults had their first vaccination between 8 December 2020 and 18 March 2021. The crude incidence rates (per 100,000 person-years) of all venous events were as follows: prevaccination, 140 [95% confidence interval (CI): 138 to 142]; ≤28 days post-ChAdOx1-S, 294 (281 to 307); >28 days post-ChAdOx1-S, 359 (338 to 382), ≤28 days post-BNT162b2-S, 241 (229 to 253); >28 days post-BNT162b2-S 277 (263 to 291). The crude incidence rates (per 100,000 person-years) of all arterial events were as follows: prevaccination, 546 (95% CI: 541 to 555); ≤28 days post-ChAdOx1-S, 1,211 (1,185 to 1,237); >28 days post-ChAdOx1-S, 1678 (1,630 to 1,726), ≤28 days post-BNT162b2-S, 1,242 (1,214 to 1,269); >28 days post-BNT162b2-S, 1,539 (1,507 to 1,572). Adjusted HRs (aHRs) 1 to 28 days after ChAdOx1-S, compared with unvaccinated rates, at ages <70 and ≥70 years, respectively, were 0.97 (95% CI: 0.90 to 1.05) and 0.58 (0.53 to 0.63) for venous thromboses, and 0.90 (0.86 to 0.95) and 0.76 (0.73 to 0.79) for arterial thromboses. Corresponding aHRs for BNT162b2 were 0.81 (0.74 to 0.88) and 0.57 (0.53 to 0.62) for venous thromboses, and 0.94 (0.90 to 0.99) and 0.72 (0.70 to 0.75) for arterial thromboses. aHRs for thrombotic events were higher at younger ages for venous thromboses after ChAdOx1-S, and for arterial thromboses after both vaccines. Rates of intracranial venous thrombosis (ICVT) and of thrombocytopenia in adults aged <70 years were higher 1 to 28 days after ChAdOx1-S (aHRs 2.27, 95% CI: 1.33 to 3.88 and 1.71, 1.35 to 2.16, respectively), but not after BNT162b2 (0.59, 0.24 to 1.45 and 1.00, 0.75 to 1.34) compared with unvaccinated. The corresponding absolute excess risks of ICVT 1 to 28 days after ChAdOx1-S were 0.9 to 3 per million, varying by age and sex. The main limitations of the study are as follows: (i) it relies on the accuracy of coded healthcare data to identify exposures, covariates, and outcomes; (ii) the use of primary reason for hospital admission to measure outcome, which improves the positive predictive value but may lead to an underestimation of incidence; and (iii) potential unmeasured confounding.

CONCLUSIONS

In this study, we observed increases in rates of ICVT and thrombocytopenia after ChAdOx1-S vaccination in adults aged <70 years that were small compared with its effect in reducing COVID-19 morbidity and mortality, although more precise estimates for adults aged <40 years are needed. For people aged ≥70 years, rates of arterial or venous thrombotic events were generally lower after either vaccine compared with unvaccinated, suggesting that either vaccine is suitable in this age group.

Towards nationally curated data archives for clinical radiology image analysis at scale: Learnings from national data collection in response to a pandemic

The prevalence of the coronavirus SARS-CoV-2 disease has resulted in the unprecedented collection of health data to support research. Historically, coordinating the collation of such datasets on a national scale has been challenging to execute for several reasons, including issues with data privacy, the lack of data reporting standards, interoperable technologies, and distribution methods. The coronavirus SARS-CoV-2 disease pandemic has highlighted the importance of collaboration between government bodies, healthcare institutions, academic researchers and commercial companies in overcoming these issues during times of urgency. The National COVID-19 Chest Imaging Database, led by NHSX, British Society of Thoracic Imaging, Royal Surrey NHS Foundation Trust and Faculty, is an example of such a national initiative. Here, we summarise the experiences and challenges of setting up the National COVID-19 Chest Imaging Database, and the implications for future ambitions of national data curation in medical imaging to advance the safe adoption of artificial intelligence in healthcare.

RuralCovidLife: Study protocol and description of the data

RuralCovidLife is part of Generation Scotland’s CovidLife project, investigating the impact of the COVID-19 pandemic and mitigation measures on people in Scotland. The RuralCovidLife project focuses on Scotland’s rural communities, and how they have been impacted by the pandemic. During survey development, Generation Scotland consulted with people living or working in rural communities, and collaborated with a patient and public involvement and engagement (PPIE) group composed of rural community leaders. Through this consultation work, the RuralCovidLife survey was developed to assess the issues most pertinent to people in rural communities, such as mental health, employment, transport, connectivity, and local communities. Between 14th October and 30th November 2020, 3,365 participants from rural areas in Scotland took part in the survey. Participant ages ranged from 16 to 96 (mean = 58.4, standard deviation [SD] = 13.3), and the majority of the participants were female (70.5%). Over half (51.3%) had taken part in the original CovidLife survey. RuralCovidLife includes a subsample (n = 523) of participants from the Generation Scotland cohort. Pre-pandemic data on health and lifestyle, as well as biological samples, are available for these participants. These participants’ data can also be linked to past and future healthcare records, allowing analysis of retrospective and prospective health outcomes. Like Generation Scotland, RuralCovidLife is designed as a resource for researchers. RuralCovidLife data, as well as the linked Generation Scotland data, is available for use by external researchers following approval from the Generation Scotland Access Committee. RuralCovidLife can be used to investigate mental health, well-being, and behaviour in participants living in rural areas during the COVID-19 pandemic, as well as comparisons with non-rural samples. Moreover, the sub-sample with full Generation Scotland data and linkage can be used to investigate the long-term health consequences of the COVID-19 pandemic in rural communities.

Differential cerebrovascular risks in glioblastoma and meningioma patients: a population-based matched cohort study in Wales (United Kingdom)

Aims

There is limited evidence on cerebrovascular risks in glioblastoma and meningioma patients. We aimed to compare cerebrovascular risks of these patients with the general population.

Method

We used population-based routine healthcare and administrative data linkage in this matched cohort study. Cases were adult glioblastoma and meningioma patients diagnosed in Wales 2000-2014 identified in the cancer registry. Controls from cancer-free general population were matched to cases (5:1 ratio) on age (±5 years), sex and GP practice. Factors included in multivariable models were age, sex, index of multiple deprivation, hypertension, diabetes, high cholesterol, history of cardiovascular disease, and medications for cardiovascular diseases. Outcomes were fatal and non-fatal haemorrhagic and ischaemic stroke. We used flexible parametric models adjusting for confounders to calculate the hazard ratios (HR).

Results

Final analytic population was 16,921 participants, of which 1,340 had glioblastoma and 1,498 had meningioma. The median follow-up time was 0.5 year for glioblastoma patients, 4.9 years for meningioma patients, and 6.6 years for controls. The number of haemorrhage and ischaemic stroke was 154 and 374 in the glioblastoma matched cohort, respectively, and 180 and 569 in the meningioma matched cohort, respectively. The adjusted HRs for haemorrhagic and ischaemic stroke were 3.74 (95%CI 1.87-6.57) and 5.62 (95%CI 2.56-10.42) in glioblastoma patients, respectively, and were 2.42 (95%CI 1.58-3.52) and 1.86 (95%CI 1.54-2.23) in meningioma patients compared with their controls.

Conclusion

Glioblastoma and meningioma patients had higher cerebrovascular risks; these risks were even higher for glioblastoma patients. Further assessment of these potentially modifiable risks may improve survivorship.

Benchmarking network-based gene prioritization methods for cerebral small vessel disease

Network-based gene prioritization algorithms are designed to prioritize disease-associated genes based on known ones using biological networks of protein interactions, gene-disease associations (GDAs) and other relationships between biological entities. Various algorithms have been developed based on different mechanisms, but it is not obvious which algorithm is optimal for a specific disease. To address this issue, we benchmarked multiple algorithms for their application in cerebral small vessel disease (cSVD). We curated protein-gene interactions (PGIs) and GDAs from databases and assembled PGI networks and disease-gene heterogeneous networks. A screening of algorithms resulted in seven representative algorithms to be benchmarked. Performance of algorithms was assessed using both leave-one-out cross-validation (LOOCV) and external validation with MEGASTROKE genome-wide association study (GWAS). We found that random walk with restart on the heterogeneous network (RWRH) showed best LOOCV performance, with median LOOCV rediscovery rank of 185.5 (out of 19 463 genes). The GenePanda algorithm had most GWAS-confirmable genes in top 200 predictions, while RWRH had best ranks for small vessel stroke-associated genes confirmed in GWAS. In conclusion, RWRH has overall better performance for application in cSVD despite its susceptibility to bias caused by degree centrality. Choice of algorithms should be determined before applying to specific disease. Current pure network-based gene prioritization algorithms are unlikely to find novel disease-associated genes that are not associated with known ones. The tools for implementing and benchmarking algorithms have been made available and can be generalized for other diseases.

CovidLife: a resource to understand mental health, well-being and behaviour during the COVID-19 pandemic in the UK

CovidLife is a longitudinal observational study designed to investigate the impact of the COVID-19 pandemic on mental health, well-being and behaviour in adults living in the UK. In total, 18,518 participants (mean age = 56.43, SD = 14.35) completed the first CovidLife questionnaire (CovidLife1) between April and June 2020. To date, participants have completed two follow-up assessments. CovidLife2 took place between July and August 2020 (n = 11,319), and CovidLife3 took place in February 2021 (n = 10,386). A range of social and psychological measures were administered at each wave including assessments of anxiety, depression, well-being, loneliness and isolation. Information on sociodemographic, health, and economic circumstances was also collected. Questions also assessed information on COVID-19 infections and symptoms, compliance to COVID-19 restrictions, and opinions on the UK and Scottish Governments' handling of the pandemic. CovidLife includes a subsample of 4,847 participants from the Generation Scotland cohort (N~24,000, collected 2006-2011); a well-characterised cohort of families in Scotland with pre-pandemic data on mental health, physical health, lifestyle, and socioeconomic factors, along with biochemical and genomic data derived from biological samples. These participants also consented to their study data being linked to Scottish health records. CovidLife and Generation Scotland data can be accessed and used by external researchers following approval from the Generation Scotland Access Committee. CovidLife can be used to investigate mental health, well-being and behaviour during COVID-19; how these vary according to sociodemographic, health and economic circumstances; and how these change over time. The Generation Scotland subsample with pre-pandemic data and linkage to health records can be used to investigate the predictors of health and well-being during COVID-19 and the future health consequences of the COVID-19 pandemic.

Abstract LB084: High cardiovascular disease mortality after central nervous system tumor diagnosis: Evidence from UK and USA population-based study

Background Accumulating data show cardiovascular disease (CVD) is a major cause of death in many cancer patients supporting cardio-oncology epidemiology and clinical studies. Although rare, patients diagnosed with brain and central nervous system (CNS) tumours have significant morbidity and mortality. Whether CVD is a major cause of death and if this differs by malignancy has not been comprehensively assessed. Methods We aimed to examine the risk of CVD mortality in patients diagnosed with malignant and non-malignant CNS tumours using cancer registry data in the UK and US. Analyses were conducted using Wales Cancer Registry, UK (Secure Anonymised Information Linkage, SAIL) for 8,743 patients diagnosed from 2000-2015 (54.9% of which died); and the US National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) for 188,526 patients diagnosed from 2005-2015 (40.0% of which died). Standardized mortality ratios (SMRs) and 95% confidence intervals (CI) were calculated for CVD cause of death (heart disease, cerebrovascular disease, hypertension, atherosclerosis, and aortic aneurysm/dissection) and adjusted for age, sex and calendar year compared to all Welsh and US residents. SMRs were stratified by tumour types (malignant and non-malignant tumours) and main histologic types (glioma and meningioma). Results CVD is the second major cause of death for CNS tumour patients in SAIL and SEER (9.5% & 12.1%, respectively). Patients with malignant and non-malignant CNS tumours had excess CVD mortality compared with the general population (SAIL SMR=2.64, 95% CI=2.39-2.90, SEER SMR=1.38, 95% CI=1.35-1.42). Patients were more likely to die of CVD compared to the general population regardless of whether they were diagnosed with non-malignant meningiomas subtypes (SAIL SMR=3.13, 95%CI=2.73-3.57; SEER SMR=1.36, 95%CI=1.32-1.40) or malignant gliomas (SAIL SMR=2.08, 95%CI=1.46-2.88; SEER SMR=2.21, 95%CI=2.05-2.38). Patients diagnosed younger than 50 years of age had excess risk from CVD mortality compared to general population than those diagnosed at older ages (SAIL SMR=4.58, 95%CI=2.38-7.84, SEER SMR 2.03-95%CI=1.79-2.03). Patients had greater risk of CVD mortality within the first year after CNS tumour diagnosis in both SAIL and SEER (SMR=2.98, 95% CI=2.39-3.66 & SMR=2.14 95%CI=2.03-2.25, respectively). Conclusion CVD mortality is high among patients diagnosed with CNS tumours compared to general population. Cross national datasets for different histologic types of CNS tumours could help define high risk groups that may be targeted for future prevention and clinical studies to clarify the aetiology of CVD among CNS cancer patients.

Citation Format: KAI JIN, Paul Brennan, Michael Poon, Cathie Sudlow, Jonine FIGUEROA. High cardiovascular disease mortality after central nervous system tumor diagnosis: Evidence from UK and USA population-based study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr LB084.

Metastatic bone disease: new quality performance indicator development

OBJECTIVES

Patients with metastatic bone disease (MBD) should receive the same standard of care regardless of which centre they are treated in. The aim was to develop and test a set of quality performance indicators (QPIs) to evaluate care for patients with MBD referred to orthopaedics.

METHODS

QPIs were adapted from the literature and ranked on feasibility and necessity during a modified RAND/Delphi consensus process. They were then validated and field tested in a retrospective cohort of 108 patients using indicator-specific targets set during consensus.

RESULTS

2568 articles including six guidelines were reviewed. 43 quality objectives were extracted and 40 proceeded to expert consensus. After two rounds, 18 QPIs for MBD care were generated, with the following generating the highest consensus: 'Patients with high fracture risk should receive urgent assessment' (combined mean 6.7/7, 95% CI 6.5 to 6.8) and 'preoperative workup should include full blood tests including group and save' (combined mean 6.7/7, 95% CI 6.5 to 6.9). In the pilot test, targets were met for 5/18 QPIs (mean 52%, standard deviation 22%). The median deviation from projected target was -14% (interquartile range -11% to -31%, range -74% to 11%). The highest scoring QPI was 'adults with fractures should have surgery within 7 days' (target 80%:actual 92%).

CONCLUSIONS

The published evidence and guidelines were adapted into a set of validated QPIs for MBD care which can be used to evaluate variation in care between centres. These QPIs should be correlated with outcome scores to determine whether they can act as predictors of outcome after surgery.

Developing automated methods for disease subtyping in UK Biobank: an exemplar study on stroke

BACKGROUND

Better phenotyping of routinely collected coded data would be useful for research and health improvement. For example, the precision of coded data for hemorrhagic stroke (intracerebral hemorrhage [ICH] and subarachnoid hemorrhage [SAH]) may be as poor as < 50%. This work aimed to investigate the feasibility and added value of automated methods applied to clinical radiology reports to improve stroke subtyping.

METHODS

From a sub-population of 17,249 Scottish UK Biobank participants, we ascertained those with an incident stroke code in hospital, death record or primary care administrative data by September 2015, and ≥ 1 clinical brain scan report. We used a combination of natural language processing and clinical knowledge inference on brain scan reports to assign a stroke subtype (ischemic vs ICH vs SAH) for each participant and assessed performance by precision and recall at entity and patient levels.

RESULTS

Of 225 participants with an incident stroke code, 207 had a relevant brain scan report and were included in this study. Entity level precision and recall ranged from 78 to 100%. Automated methods showed precision and recall at patient level that were very good for ICH (both 89%), good for SAH (both 82%), but, as expected, lower for ischemic stroke (73%, and 64%, respectively), suggesting coded data remains the preferred method for identifying the latter stroke subtype.

CONCLUSIONS

Our automated method applied to radiology reports provides a feasible, scalable and accurate solution to improve disease subtyping when used in conjunction with administrative coded health data. Future research should validate these findings in a different population setting.

Linked electronic health records for research on a nationwide cohort of more than 54 million people in England: data resource

OBJECTIVE

To describe a novel England-wide electronic health record (EHR) resource enabling whole population research on covid-19 and cardiovascular disease while ensuring data security and privacy and maintaining public trust.

DESIGN

Data resource comprising linked person level records from national healthcare settings for the English population, accessible within NHS Digital's new trusted research environment.

SETTING

EHRs from primary care, hospital episodes, death registry, covid-19 laboratory test results, and community dispensing data, with further enrichment planned from specialist intensive care, cardiovascular, and covid-19 vaccination data.

PARTICIPANTS

54.4 million people alive on 1 January 2020 and registered with an NHS general practitioner in England.

MAIN MEASURES OF INTEREST

Confirmed and suspected covid-19 diagnoses, exemplar cardiovascular conditions (incident stroke or transient ischaemic attack and incident myocardial infarction) and all cause mortality between 1 January and 31 October 2020.

RESULTS

The linked cohort includes more than 96% of the English population. By combining person level data across national healthcare settings, data on age, sex, and ethnicity are complete for around 95% of the population. Among 53.3 million people with no previous diagnosis of stroke or transient ischaemic attack, 98 721 had a first ever incident stroke or transient ischaemic attack between 1 January and 31 October 2020, of which 30% were recorded only in primary care and 4% only in death registry records. Among 53.2 million people with no previous diagnosis of myocardial infarction, 62 966 had an incident myocardial infarction during follow-up, of which 8% were recorded only in primary care and 12% only in death registry records. A total of 959 470 people had a confirmed or suspected covid-19 diagnosis (714 162 in primary care data, 126 349 in hospital admission records, 776 503 in covid-19 laboratory test data, and 50 504 in death registry records). Although 58% of these were recorded in both primary care and covid-19 laboratory test data, 15% and 18%, respectively, were recorded in only one.

CONCLUSIONS

This population-wide resource shows the importance of linking person level data across health settings to maximise completeness of key characteristics and to ascertain cardiovascular events and covid-19 diagnoses. Although this resource was initially established to support research on covid-19 and cardiovascular disease to benefit clinical care and public health and to inform healthcare policy, it can broaden further to enable a wide range of research.

Conducting public involvement in dementia research: The contribution of the European Working Group of People with Dementia to the ROADMAP project

Background

Dementia outcomes include memory loss, language impairment, reduced quality of life and personality changes. Research suggests that outcomes selected for dementia clinical trials might not be the most important to people affected.

Objective

One of the goals of the ‘Real world Outcomes across the Alzheimer's Disease spectrum for better care: Multi‐modal data Access Platform’ (ROADMAP) project was to identify important outcomes from the perspective of people with dementia and their caregivers. We review how ROADMAP's Public Involvement shaped the programme, impacted the research process and gave voice to people affected by dementia.

Design

The European Working Group of People with Dementia (EWGPWD) were invited to participate. In‐person consultations were held with people with dementia and caregivers, with advance information provided on ROADMAP activities. Constructive criticism of survey content, layout and accessibility was sought, as were views and perspectives on terminology and key concepts around disease progression.

Results

The working group provided significant improvements to survey accessibility and acceptability. They promoted better understanding of concepts around disease progression and how researchers might approach measuring and interpreting findings. They effectively expressed difficult concepts through real‐world examples.

Conclusions

The role of the EWGPWD in ROADMAP was crucial, and its impact was highly influential. Involvement from the design stage helped shape the ethos of the programme and ultimately its meaningfulness.

Public contribution

People with dementia and their carers were involved through structured consultations and invited to provide feedback on project materials, methods and insight into terminology and relevant concepts.

Development and validation of the ISARIC 4C Deterioration model for adults hospitalised with COVID-19: a prospective cohort study

BACKGROUND

Prognostic models to predict the risk of clinical deterioration in acute COVID-19 cases are urgently required to inform clinical management decisions.

METHODS

We developed and validated a multivariable logistic regression model for in-hospital clinical deterioration (defined as any requirement of ventilatory support or critical care, or death) among consecutively hospitalised adults with highly suspected or confirmed COVID-19 who were prospectively recruited to the International Severe Acute Respiratory and Emerging Infections Consortium Coronavirus Clinical Characterisation Consortium (ISARIC4C) study across 260 hospitals in England, Scotland, and Wales. Candidate predictors that were specified a priori were considered for inclusion in the model on the basis of previous prognostic scores and emerging literature describing routinely measured biomarkers associated with COVID-19 prognosis. We used internal-external cross-validation to evaluate discrimination, calibration, and clinical utility across eight National Health Service (NHS) regions in the development cohort. We further validated the final model in held-out data from an additional NHS region (London).

FINDINGS

74 944 participants (recruited between Feb 6 and Aug 26, 2020) were included, of whom 31 924 (43·2%) of 73 948 with available outcomes met the composite clinical deterioration outcome. In internal-external cross-validation in the development cohort of 66 705 participants, the selected model (comprising 11 predictors routinely measured at the point of hospital admission) showed consistent discrimination, calibration, and clinical utility across all eight NHS regions. In held-out data from London (n=8239), the model showed a similarly consistent performance (C-statistic 0·77 [95% CI 0·76 to 0·78]; calibration-in-the-large 0·00 [-0·05 to 0·05]); calibration slope 0·96 [0·91 to 1·01]), and greater net benefit than any other reproducible prognostic model.

INTERPRETATION

The 4C Deterioration model has strong potential for clinical utility and generalisability to predict clinical deterioration and inform decision making among adults hospitalised with COVID-19.

FUNDING

National Institute for Health Research (NIHR), UK Medical Research Council, Wellcome Trust, Department for International Development, Bill & Melinda Gates Foundation, EU Platform for European Preparedness Against (Re-)emerging Epidemics, NIHR Health Protection Research Unit (HPRU) in Emerging and Zoonotic Infections at University of Liverpool, NIHR HPRU in Respiratory Infections at Imperial College London.

Patient-reported assessment of outcome after surgery for bone metastases

Regardless of prognosis, surgery is often considered in metastatic bone disease (MBD) as a palliative procedure to improve function and quality of life. Traditional focus on objective outcomes such as mortality is inappropriate in this group, and there is a drive to assess outcomes via patient-reported outcome measures (PROMs). This is an overview of current understanding of MBD outcomes and how this should influence future decision-making and research. The objectives of this review were to identify difficulties in measuring PROMs in the MBD patient population and explore alternatives to patientreported outcomes. We also provide an overview of current understanding of outcomes in MBD and how this should influence decision-making and direct research.

Face covering adherence is positively associated with better mental health and wellbeing: a longitudinal analysis of the CovidLife surveys

Background: Face masks or coverings are effective at reducing airborne infection rates, yet pandemic mitigation measures, including wearing face coverings, have been suggested to contribute to reductions in quality of life and poorer mental health. Complaints of inconvenience, discomfort, and other issues have been repeatedly and loudly voiced by critics, and adherence in many nations is not strong enough to suppress viral spread. We wished to see whether wearing face coverings is associated with mental health and wellbeing. Methods: We analysed survey 1 and 2 of the CovidLife study, a sample of more than 18,000 individuals living in the UK. The study asked a variety of questions about participants’ psychological, economic, and social lives while living under the coronavirus disease 2019 (COVID-19) pandemic in 2020. We measured individuals’ adherence to following guidance on wearing face coverings, as well as several mental health outcomes: depression, anxiety, wellbeing, life satisfaction, and loneliness. Results: We found no association between lower adherence to face covering guidelines and poorer mental health. The opposite appears to be true. Even after controlling for behavioural, social, and psychological confounds, including measures of pre-pandemic mental health, individuals who wore face coverings “most of the time” or “always” had better mental health and wellbeing than those who did not. Individuals who wore masks only “some of the time” or “never” tended to be male, lower income, and already had COVID-19 or COVID-19-like symptoms. Conclusions: These results suggest that wearing face coverings more often does not negatively impact mental health. Wearing a face covering more often is actually linked to better mental health and wellbeing. Implications are discussed and we highlight the potential pathways for addressing a lack of face covering that this study reveals.

Beyond the Brain

Background and Purpose:

An important minority of cerebral small vessel disease (cSVD) is monogenic. Many monogenic cSVD genes are recognized to be associated with extracerebral phenotypes. We assessed the frequency of these phenotypes in existing literature.

Methods:

We performed a systematic review following the PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), searching Medline/Embase for publications describing individuals with pathogenic variants in COL4A1/2 , TREX1 , HTRA1 , ADA2 , and CTSA genes (PROSPERO 74804). We included any publication reporting on ≥1 individual with a pathogenic variant and their clinically relevant phenotype. We extracted individuals’ characteristics and information about associated extracerebral phenotypes and stroke/transient ischemic attack. We noted any novel extracerebral phenotypes and looked for shared phenotypes between monogenic cSVDs.

Results:

After screening 6048 publications, we included 96 COL4A1 (350 individuals), 32 TREX1 (115 individuals), 43 HTRA1 (38 homozygous/61 heterozygous individuals), 16 COL4A2 (37 individuals), 119 ADA2 (209 individuals), and 3 CTSA (14 individuals) publications. The majority of individuals originated from Europe/North America, except for HTRA1 , where most were from Asia. Age varied widely, ADA2 individuals being youngest and heterozygous HTRA1/CTSA individuals oldest. Sex distribution appeared equal. Extracerebral phenotypes were common: 14% to 100% of individuals with a pathogenic variant manifested at least one extracerebral phenotype (14% COL4A2 , 43% HTRA1 heterozygotes, 47% COL4A1 , 57% TREX1 , 91% ADA2 , 94% HTRA1 homozygotes, and 100% CTSA individuals). Indeed, for 4 of 7 genes, an extracerebral phenotype was observed more frequently than stroke/transient ischemic attack. Ocular, renal, hepatic, muscle, and hematologic systems were each involved in more than one monogenic cSVD.

Conclusions:

Extracerebral phenotypes are common in monogenic cSVD with extracerebral system involvement shared between genes. However, inherent biases in the existing literature mean that further data from large-scale population-based longitudinal studies collecting health outcomes in a systematic unbiased way is warranted. The emerging knowledge will help to select patients for testing, inform clinical management, and provide further insights into the underlying mechanisms of cSVD.

Risk stratification of patients admitted to hospital with covid-19 using the ISARIC WHO Clinical Characterisation Protocol: development and validation of the 4C Mortality Score

OBJECTIVE

To develop and validate a pragmatic risk score to predict mortality in patients admitted to hospital with coronavirus disease 2019 (covid-19).

DESIGN

Prospective observational cohort study.

SETTING

International Severe Acute Respiratory and emerging Infections Consortium (ISARIC) World Health Organization (WHO) Clinical Characterisation Protocol UK (CCP-UK) study (performed by the ISARIC Coronavirus Clinical Characterisation Consortium-ISARIC-4C) in 260 hospitals across England, Scotland, and Wales. Model training was performed on a cohort of patients recruited between 6 February and 20 May 2020, with validation conducted on a second cohort of patients recruited after model development between 21 May and 29 June 2020. PARTICIPANTS: Adults (age ≥18 years) admitted to hospital with covid-19 at least four weeks before final data extraction.

MAIN OUTCOME MEASURE

In-hospital mortality.

RESULTS

35 463 patients were included in the derivation dataset (mortality rate 32.2%) and 22 361 in the validation dataset (mortality rate 30.1%). The final 4C Mortality Score included eight variables readily available at initial hospital assessment: age, sex, number of comorbidities, respiratory rate, peripheral oxygen saturation, level of consciousness, urea level, and C reactive protein (score range 0-21 points). The 4C Score showed high discrimination for mortality (derivation cohort: area under the receiver operating characteristic curve 0.79, 95% confidence interval 0.78 to 0.79; validation cohort: 0.77, 0.76 to 0.77) with excellent calibration (validation: calibration-in-the-large=0, slope=1.0). Patients with a score of at least 15 (n=4158, 19%) had a 62% mortality (positive predictive value 62%) compared with 1% mortality for those with a score of 3 or less (n=1650, 7%; negative predictive value 99%). Discriminatory performance was higher than 15 pre-existing risk stratification scores (area under the receiver operating characteristic curve range 0.61-0.76), with scores developed in other covid-19 cohorts often performing poorly (range 0.63-0.73).

CONCLUSIONS

An easy-to-use risk stratification score has been developed and validated based on commonly available parameters at hospital presentation. The 4C Mortality Score outperformed existing scores, showed utility to directly inform clinical decision making, and can be used to stratify patients admitted to hospital with covid-19 into different management groups. The score should be further validated to determine its applicability in other populations.

STUDY REGISTRATION

ISRCTN66726260.

Knowledge Driven Phenotyping

Extracting patient phenotypes from routinely collected health data (such as Electronic Health Records) requires translating clinically-sound phenotype definitions into queries/computations executable on the underlying data sources by clinical researchers. This requires significant knowledge and skills to deal with heterogeneous and often imperfect data. Translations are time-consuming, error-prone and, most importantly, hard to share and reproduce across different settings. This paper proposes a knowledge driven framework that (1) decouples the specification of phenotype semantics from underlying data sources; (2) can automatically populate and conduct phenotype computations on heterogeneous data spaces. We report preliminary results of deploying this framework on five Scottish health datasets.

Challenges for Optimizing Real-World Evidence in Alzheimer's Disease: The ROADMAP Project

ROADMAP is a public-private advisory partnership to evaluate the usability of multiple data sources, including real-world evidence, in the decision-making process for new treatments in Alzheimer's disease, and to advance key concepts in disease and pharmacoeconomic modeling. ROADMAP identified key disease and patient outcomes for stakeholders to make informed funding and treatment decisions, provided advice on data integration methods and standards, and developed conceptual cost-effectiveness and disease models designed in part to assess whether early treatment provides long-term benefit.

Mental health in UK Biobank - development, implementation and results from an online questionnaire completed by 157 366 participants: a reanalysis

BACKGROUND

UK Biobank is a well-characterised cohort of over 500 000 participants including genetics, environmental data and imaging. An online mental health questionnaire was designed for UK Biobank participants to expand its potential.

AIMS

Describe the development, implementation and results of this questionnaire.

METHOD

An expert working group designed the questionnaire, using established measures where possible, and consulting a patient group. Operational criteria were agreed for defining likely disorder and risk states, including lifetime depression, mania/hypomania, generalised anxiety disorder, unusual experiences and self-harm, and current post-traumatic stress and hazardous/harmful alcohol use.

RESULTS

A total of 157 366 completed online questionnaires were available by August 2017. Participants were aged 45-82 (53% were ≥65 years) and 57% women. Comparison of self-reported diagnosed mental disorder with a contemporary study shows a similar prevalence, despite respondents being of higher average socioeconomic status. Lifetime depression was a common finding, with 24% (37 434) of participants meeting criteria and current hazardous/harmful alcohol use criteria were met by 21% (32 602), whereas other criteria were met by less than 8% of the participants. There was extensive comorbidity among the syndromes. Mental disorders were associated with a high neuroticism score, adverse life events and long-term illness; addiction and bipolar affective disorder in particular were associated with measures of deprivation.

CONCLUSIONS

The UK Biobank questionnaire represents a very large mental health survey in itself, and the results presented here show high face validity, although caution is needed because of selection bias. Built into UK Biobank, these data intersect with other health data to offer unparalleled potential for crosscutting biomedical research involving mental health.

Body mass index, diet, physical inactivity, and the incidence of dementia in 1 million UK women

OBJECTIVE

To help determine whether midlife obesity is a cause of dementia and whether low body mass index (BMI), low caloric intake, and physical inactivity are causes or merely consequences of the gradual onset of dementia by recording these factors early in a large 20-year prospective study and relating them to dementia detection rates separately during follow-up periods of <5, 5 to 9, 10 to 14, and 15+ years.

METHODS

A total of 1,136,846 UK women, mean age 56 (SD 5) years, were recruited in 1996 to 2001 and asked about height, weight, caloric intake, and inactivity. They were followed up until 2017 by electronic linkage to National Health Service records, detecting hospital admissions with mention of dementia. Cox regression yielded adjusted rate ratios (RRs) for first dementia detection during particular follow-up periods.

RESULTS

Fifteen years after the baseline survey, only 1% were lost to follow-up, and 89% remained alive with no detected dementia, of whom 18,695 had dementia detected later, at a mean age of 77 (SD 4) years. Dementia detection during years 15+ was associated with baseline obesity (BMI 30+ vs 20-24 kg/m2: RR 1.21, 95% confidence interval 1.16-1.26, p < 0.0001) but not clearly with low BMI, low caloric intake, or inactivity at baseline. The latter 3 factors were associated with increased dementia rates during the first decade, but these associations weakened substantially over time, approaching null after 15 years.

CONCLUSIONS

Midlife obesity may well be a cause of dementia. In contrast, behavioral changes due to preclinical disease could largely or wholly account for associations of low BMI, low caloric intake, and inactivity with dementia detection during the first decade of follow-up.

Efficient Reuse of Natural Language Processing Models for Phenotype-Mention Identification in Free-text Electronic Medical Records: A Phenotype Embedding Approach

Background

Much effort has been put into the use of automated approaches, such as natural language processing (NLP), to mine or extract data from free-text medical records in order to construct comprehensive patient profiles for delivering better health care. Reusing NLP models in new settings, however, remains cumbersome, as it requires validation and retraining on new data iteratively to achieve convergent results.

Objective

The aim of this work is to minimize the effort involved in reusing NLP models on free-text medical records.

Methods

We formally define and analyze the model adaptation problem in phenotype-mention identification tasks. We identify “duplicate waste” and “imbalance waste,” which collectively impede efficient model reuse. We propose a phenotype embedding–based approach to minimize these sources of waste without the need for labelled data from new settings.

Results

We conduct experiments on data from a large mental health registry to reuse NLP models in four phenotype-mention identification tasks. The proposed approach can choose the best model for a new task, identifying up to 76% waste (duplicate waste), that is, phenotype mentions without the need for validation and model retraining and with very good performance (93%-97% accuracy). It can also provide guidance for validating and retraining the selected model for novel language patterns in new tasks, saving around 80% waste (imbalance waste), that is, the effort required in “blind” model-adaptation approaches.

Conclusions

Adapting pretrained NLP models for new tasks can be more efficient and effective if the language pattern landscapes of old settings and new settings can be made explicit and comparable. Our experiments show that the phenotype-mention embedding approach is an effective way to model language patterns for phenotype-mention identification tasks and that its use can guide efficient NLP model reuse.

UK phenomics platform for developing and validating electronic health record phenotypes: CALIBER

OBJECTIVE

Electronic health records (EHRs) are a rich source of information on human diseases, but the information is variably structured, fragmented, curated using different coding systems, and collected for purposes other than medical research. We describe an approach for developing, validating, and sharing reproducible phenotypes from national structured EHR in the United Kingdom with applications for translational research.

MATERIALS AND METHODS

We implemented a rule-based phenotyping framework, with up to 6 approaches of validation. We applied our framework to a sample of 15 million individuals in a national EHR data source (population-based primary care, all ages) linked to hospitalization and death records in England. Data comprised continuous measurements (for example, blood pressure; medication information; coded diagnoses, symptoms, procedures, and referrals), recorded using 5 controlled clinical terminologies: (1) read (primary care, subset of SNOMED-CT [Systematized Nomenclature of Medicine Clinical Terms]), (2) International Classification of Diseases-Ninth Revision and Tenth Revision (secondary care diagnoses and cause of mortality), (3) Office of Population Censuses and Surveys Classification of Surgical Operations and Procedures, Fourth Revision (hospital surgical procedures), and (4) DM+D prescription codes.

RESULTS

Using the CALIBER phenotyping framework, we created algorithms for 51 diseases, syndromes, biomarkers, and lifestyle risk factors and provide up to 6 validation approaches. The EHR phenotypes are curated in the open-access CALIBER Portal (https://www.caliberresearch.org/portal) and have been used by 40 national and international research groups in 60 peer-reviewed publications.

CONCLUSIONS

We describe a UK EHR phenomics approach within the CALIBER EHR data platform with initial evidence of validity and use, as an important step toward international use of UK EHR data for health research.