Dataset Information

Comparison of methods for predicting COVID-19-related death in the general population using the OpenSAFELY platform.

ABSTRACT:

Background

Obtaining accurate estimates of the risk of COVID-19-related death in the general population is challenging in the context of changing levels of circulating infection.

Methods

We propose a modelling approach to predict 28-day COVID-19-related death which explicitly accounts for COVID-19 infection prevalence using a series of sub-studies from new landmark times incorporating time-updating proxy measures of COVID-19 infection prevalence. This was compared with an approach ignoring infection prevalence. The target population was adults registered at a general practice in England in March 2020. The outcome was 28-day COVID-19-related death. Predictors included demographic characteristics and comorbidities. Three proxies of local infection prevalence were used: model-based estimates, rate of COVID-19-related attendances in emergency care, and rate of suspected COVID-19 cases in primary care. We used data within the TPP SystmOne electronic health record system linked to Office for National Statistics mortality data, using the OpenSAFELY platform, working on behalf of NHS England. Prediction models were developed in case-cohort samples with a 100-day follow-up. Validation was undertaken in 28-day cohorts from the target population. We considered predictive performance (discrimination and calibration) in geographical and temporal subsets of data not used in developing the risk prediction models. Simple models were contrasted to models including a full range of predictors.

Results

Prediction models were developed on 11,972,947 individuals, of whom 7999 experienced COVID-19-related death. All models discriminated well between individuals who did and did not experience the outcome, including simple models adjusting only for basic demographics and number of comorbidities: C-statistics 0.92-0.94. However, absolute risk estimates were substantially miscalibrated when infection prevalence was not explicitly modelled.

Conclusions

Our proposed models allow absolute risk estimation in the context of changing infection prevalence but predictive performance is sensitive to the proxy for infection prevalence. Simple models can provide excellent discrimination and may simplify implementation of risk prediction tools.

SUBMITTER: OpenSAFELY Collaborative

PROVIDER: S-EPMC8865947 | biostudies-literature | 2022 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Comparison of methods for predicting COVID-19-related death in the general population using the OpenSAFELY platform.

Williamson Elizabeth J EJ Tazare John J Bhaskaran Krishnan K McDonald Helen I HI Walker Alex J AJ Tomlinson Laurie L Wing Kevin K Bacon Sebastian S Bates Chris C Curtis Helen J HJ Forbes Harriet J HJ Minassian Caroline C Morton Caroline E CE Nightingale Emily E Mehrkar Amir A Evans David D Nicholson Brian D BD Leon David A DA Inglesby Peter P MacKenna Brian B Davies Nicholas G NG DeVito Nicholas J NJ Drysdale Henry H Cockburn Jonathan J Hulme William J WJ Morley Jessica J Douglas Ian I Rentsch Christopher T CT Mathur Rohini R Wong Angel A Schultze Anna A Croker Richard R Parry John J Hester Frank F Harper Sam S Grieve Richard R Harrison David A DA Steyerberg Ewout W EW Eggo Rosalind M RM Diaz-Ordaz Karla K Keogh Ruth R Evans Stephen J W SJW Smeeth Liam L Goldacre Ben B

Diagnostic and prognostic research 20220224 1

<h4>Background</h4>Obtaining accurate estimates of the risk of COVID-19-related death in the general population is challenging in the context of changing levels of circulating infection.<h4>Methods</h4>We propose a modelling approach to predict 28-day COVID-19-related death which explicitly accounts for COVID-19 infection prevalence using a series of sub-studies from new landmark times incorporating time-updating proxy measures of COVID-19 infection prevalence. This was compared with an approach ...[more]

PMID: 35197114

Similar Datasets

Project description:ObjectiveTo assess the association between learning disability and risk of hospital admission and death from covid-19 in England among adults and children.DesignPopulation based cohort study on behalf of NHS England using the OpenSAFELY platform.SettingPatient level data were obtained for more than 17 million people registered with a general practice in England that uses TPP software. Electronic health records were linked with death data from the Office for National Statistics and hospital admission data from NHS Secondary Uses Service.ParticipantsAdults (aged 16-105 years) and children (<16 years) from two cohorts: wave 1 (registered with a TPP practice as of 1 March 2020 and followed until 31 August 2020); and wave 2 (registered 1 September 2020 and followed until 8 February 2021). The main exposure group consisted of people on a general practice learning disability register; a subgroup was defined as those having profound or severe learning disability. People with Down's syndrome and cerebral palsy were identified (whether or not they were on the learning disability register).Main outcome measureCovid-19 related hospital admission and covid-19 related death. Non-covid-19 deaths were also explored.ResultsFor wave 1, 14 312 023 adults aged ≥16 years were included, and 90 307 (0.63%) were on the learning disability register. Among adults on the register, 538 (0.6%) had a covid-19 related hospital admission; there were 222 (0.25%) covid-19 related deaths and 602 (0.7%) non-covid deaths. Among adults not on the register, 29 781 (0.2%) had a covid-19 related hospital admission; there were 13 737 (0.1%) covid-19 related deaths and 69 837 (0.5%) non-covid deaths. Wave 1 hazard ratios for adults on the learning disability register (adjusted for age, sex, ethnicity, and geographical location) were 5.3 (95% confidence interval 4.9 to 5.8) for covid-19 related hospital admission and 8.2 (7.2 to 9.4) for covid-19 related death. Wave 2 produced similar estimates. Associations were stronger among those classified as having severe to profound learning disability, and among those in residential care. For both waves, Down's syndrome and cerebral palsy were associated with increased hazards for both events; Down's syndrome to a greater extent. Hazard ratios for non-covid deaths followed similar patterns with weaker associations. Similar patterns of increased relative risk were seen for children, but covid-19 related deaths and hospital admissions were rare, reflecting low event rates among children.ConclusionsPeople with learning disability have markedly increased risks of hospital admission and death from covid-19, over and above the risks observed for non-covid causes of death. Prompt access to covid-19 testing and healthcare is warranted for this vulnerable group, and prioritisation for covid-19 vaccination and other targeted preventive measures should be considered.

Project description:BackgroundWhether HIV infection is associated with risk of death due to COVID-19 is unclear. We aimed to investigate this association in a large-scale population-based study in England.MethodsWe did a retrospective cohort study. Working on behalf of NHS England, we used the OpenSAFELY platform to analyse routinely collected electronic primary care data linked to national death registrations. We included all adults (aged ≥18 years) alive and in follow-up on Feb 1, 2020, and with at least 1 year of continuous registration with a general practitioner before this date. People with a primary care record for HIV infection were compared with people without HIV. The outcome was COVID-19 death, defined as the presence of International Classification of Diseases 10 codes U07.1 or U07.2 anywhere on the death certificate. Cox regression models were used to estimate the association between HIV infection and COVID-19 death; they were initially adjusted for age and sex, then we added adjustment for index of multiple deprivation and ethnicity, and then for a broad range of comorbidities. Interaction terms were added to assess effect modification by age, sex, ethnicity, comorbidities, and calendar time.Results17 282 905 adults were included, of whom 27 480 (0·16%) had HIV recorded. People living with HIV were more likely to be male, of Black ethnicity, and from a more deprived geographical area than the general population. 14 882 COVID-19 deaths occurred during the study period, with 25 among people with HIV. People living with HIV had higher risk of COVID-19 death than those without HIV after adjusting for age and sex: hazard ratio (HR) 2·90 (95% CI 1·96-4·30; p<0·0001). The association was attenuated, but risk remained high, after adjustment for deprivation, ethnicity, smoking and obesity: adjusted HR 2·59 (95% CI 1·74-3·84; p<0·0001). There was some evidence that the association was larger among people of Black ethnicity: HR 4·31 (95% CI 2·42-7·65) versus 1·84 (1·03-3·26) in non-Black individuals (p-interaction=0·044).InterpretationPeople with HIV in the UK seem to be at increased risk of COVID-19 mortality. Targeted policies should be considered to address this raised risk as the pandemic response evolves.FundingWellcome, Royal Society, National Institute for Health Research, National Institute for Health Research Oxford Biomedical Research Centre, UK Medical Research Council, Health Data Research UK.

Project description:BackgroundEarly descriptions of patients admitted to hospital during the COVID-19 pandemic showed a lower prevalence of asthma and chronic obstructive pulmonary disease (COPD) than would be expected for an acute respiratory disease like COVID-19, leading to speculation that inhaled corticosteroids (ICSs) might protect against infection with severe acute respiratory syndrome coronavirus 2 or the development of serious sequelae. We assessed the association between ICS and COVID-19-related death among people with COPD or asthma using linked electronic health records (EHRs) in England, UK.MethodsIn this observational study, we analysed patient-level data for people with COPD or asthma from primary care EHRs linked with death data from the Office of National Statistics using the OpenSAFELY platform. The index date (start of follow-up) for both cohorts was March 1, 2020; follow-up lasted until May 6, 2020. For the COPD cohort, individuals were eligible if they were aged 35 years or older, had COPD, were a current or former smoker, and were prescribed an ICS or long-acting β agonist plus long-acting muscarinic antagonist (LABA-LAMA) as combination therapy within the 4 months before the index date. For the asthma cohort, individuals were eligible if they were aged 18 years or older, had been diagnosed with asthma within 3 years of the index date, and were prescribed an ICS or short-acting β agonist (SABA) only within the 4 months before the index date. We compared the outcome of COVID-19-related death between people prescribed an ICS and those prescribed alternative respiratory medications: ICSs versus LABA-LAMA for the COPD cohort, and low-dose or medium-dose and high-dose ICSs versus SABAs only in the asthma cohort. We used Cox regression models to estimate hazard ratios (HRs) and 95% CIs for the association between exposure categories and the outcome in each population, adjusted for age, sex, and all other prespecified covariates. We calculated e-values to quantify the effect of unmeasured confounding on our results.FindingsWe identified 148 557 people with COPD and 818 490 people with asthma who were given relevant respiratory medications in the 4 months before the index date. People with COPD who were prescribed ICSs were at increased risk of COVID-19-related death compared with those prescribed LABA-LAMA combinations (adjusted HR 1·39 [95% CI 1·10-1·76]). Compared with those prescribed SABAs only, people with asthma who were prescribed high-dose ICS were at an increased risk of death (1·55 [1·10-2·18]), whereas those given a low or medium dose were not (1·14 [0·85-1·54]). Sensitivity analyses showed that the apparent harmful association we observed could be explained by relatively small health differences between people prescribed ICS and those not prescribed ICS that were not recorded in the database (e value lower 95% CI 1·43).InterpretationOur results do not support a major role for regular ICS use in protecting against COVID-19-related death among people with asthma or COPD. Observed increased risks of COVID-19-related death can be plausibly explained by unmeasured confounding due to disease severity.FundingUK Medical Research Council.

Project description:BackgroundMortality from COVID-19 shows a strong relationship with age and pre-existing medical conditions, as does mortality from other causes. We aimed to investigate how specific factors are differentially associated with COVID-19 mortality as compared to mortality from causes other than COVID-19.MethodsWorking on behalf of NHS England, we carried out a cohort study within the OpenSAFELY platform. Primary care data from England were linked to national death registrations. We included all adults (aged ≥18 years) in the database on 1st February 2020 and with >1 year of continuous prior registration; the cut-off date for deaths was 9th November 2020. Associations between individual-level characteristics and COVID-19 and non-COVID deaths, classified according to the presence of a COVID-19 code as the underlying cause of death on the death certificate, were estimated by fitting age- and sex-adjusted logistic models for these two outcomes.Findings17,456,515 individuals were included. 17,063 died from COVID-19 and 134,316 from other causes. Most factors associated with COVID-19 death were similarly associated with non-COVID death, but the magnitudes of association differed. Older age was more strongly associated with COVID-19 death than non-COVID death (e.g. ORs 40.7 [95% CI 37.7-43.8] and 29.6 [28.9-30.3] respectively for ≥80 vs 50-59 years), as was male sex, deprivation, obesity, and some comorbidities. Smoking, history of cancer and chronic liver disease had stronger associations with non-COVID than COVID-19 death. All non-white ethnic groups had higher odds than white of COVID-19 death (OR for Black: 2.20 [1.96-2.47], South Asian: 2.33 [2.16-2.52]), but lower odds than white of non-COVID death (Black: 0.88 [0.83-0.94], South Asian: 0.78 [0.75-0.81]).InterpretationSimilar associations of most individual-level factors with COVID-19 and non-COVID death suggest that COVID-19 largely multiplies existing risks faced by patients, with some notable exceptions. Identifying the unique factors contributing to the excess COVID-19 mortality risk among non-white groups is a priority to inform efforts to reduce deaths from COVID-19.FundingWellcome, Royal Society, National Institute for Health Research, National Institute for Health Research Oxford Biomedical Research Centre, UK Medical Research Council, Health Data Research UK.

Project description:BackgroundCOVID-19 has disproportionately affected minority ethnic populations in the UK. Our aim was to quantify ethnic differences in SARS-CoV-2 infection and COVID-19 outcomes during the first and second waves of the COVID-19 pandemic in England.MethodsWe conducted an observational cohort study of adults (aged ≥18 years) registered with primary care practices in England for whom electronic health records were available through the OpenSAFELY platform, and who had at least 1 year of continuous registration at the start of each study period (Feb 1 to Aug 3, 2020 [wave 1], and Sept 1 to Dec 31, 2020 [wave 2]). Individual-level primary care data were linked to data from other sources on the outcomes of interest: SARS-CoV-2 testing and positive test results and COVID-19-related hospital admissions, intensive care unit (ICU) admissions, and death. The exposure was self-reported ethnicity as captured on the primary care record, grouped into five high-level census categories (White, South Asian, Black, other, and mixed) and 16 subcategories across these five categories, as well as an unknown ethnicity category. We used multivariable Cox regression to examine ethnic differences in the outcomes of interest. Models were adjusted for age, sex, deprivation, clinical factors and comorbidities, and household size, with stratification by geographical region.FindingsOf 17 288 532 adults included in the study (excluding care home residents), 10 877 978 (62·9%) were White, 1 025 319 (5·9%) were South Asian, 340 912 (2·0%) were Black, 170 484 (1·0%) were of mixed ethnicity, 320 788 (1·9%) were of other ethnicity, and 4 553 051 (26·3%) were of unknown ethnicity. In wave 1, the likelihood of being tested for SARS-CoV-2 infection was slightly higher in the South Asian group (adjusted hazard ratio 1·08 [95% CI 1·07-1·09]), Black group (1·08 [1·06-1·09]), and mixed ethnicity group (1·04 [1·02-1·05]) and was decreased in the other ethnicity group (0·77 [0·76-0·78]) relative to the White group. The risk of testing positive for SARS-CoV-2 infection was higher in the South Asian group (1·99 [1·94-2·04]), Black group (1·69 [1·62-1·77]), mixed ethnicity group (1·49 [1·39-1·59]), and other ethnicity group (1·20 [1·14-1·28]). Compared with the White group, the four remaining high-level ethnic groups had an increased risk of COVID-19-related hospitalisation (South Asian group 1·48 [1·41-1·55], Black group 1·78 [1·67-1·90], mixed ethnicity group 1·63 [1·45-1·83], other ethnicity group 1·54 [1·41-1·69]), COVID-19-related ICU admission (2·18 [1·92-2·48], 3·12 [2·65-3·67], 2·96 [2·26-3·87], 3·18 [2·58-3·93]), and death (1·26 [1·15-1·37], 1·51 [1·31-1·71], 1·41 [1·11-1·81], 1·22 [1·00-1·48]). In wave 2, the risks of hospitalisation, ICU admission, and death relative to the White group were increased in the South Asian group but attenuated for the Black group compared with these risks in wave 1. Disaggregation into 16 ethnicity groups showed important heterogeneity within the five broader categories.InterpretationSome minority ethnic populations in England have excess risks of testing positive for SARS-CoV-2 and of adverse COVID-19 outcomes compared with the White population, even after accounting for differences in sociodemographic, clinical, and household characteristics. Causes are likely to be multifactorial, and delineating the exact mechanisms is crucial. Tackling ethnic inequalities will require action across many fronts, including reducing structural inequalities, addressing barriers to equitable care, and improving uptake of testing and vaccination.FundingMedical Research Council.

Dataset Information

Comparison of methods for predicting COVID-19-related death in the general population using the OpenSAFELY platform.

Background

Methods

Results

Conclusions

Publications

Comparison of methods for predicting COVID-19-related death in the general population using the OpenSAFELY platform.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets