RESUMO
SARS-CoV-2 lineage B.1.1.7, a variant that was first detected in the UK in September 20201, has spread to multiple countries worldwide. Several studies have established that B.1.1.7 is more transmissible than pre-existing variants, but have not identified whether it leads to any change in disease severity2. Here we analyse a dataset that links 2,245,263 positive SARS-CoV-2 community tests and 17,452 deaths associated with COVID-19 in England from 1 November 2020 to 14 February 2021. For 1,146,534 (51%) of these tests, the presence or absence of B.1.1.7 can be identified because mutations in this lineage prevent PCR amplification of the spike (S) gene target (known as S gene target failure (SGTF)1). On the basis of 4,945 deaths with known SGTF status, we estimate that the hazard of death associated with SGTF is 55% (95% confidence interval, 39-72%) higher than in cases without SGTF after adjustment for age, sex, ethnicity, deprivation, residence in a care home, the local authority of residence and test date. This corresponds to the absolute risk of death for a 55-69-year-old man increasing from 0.6% to 0.9% (95% confidence interval, 0.8-1.0%) within 28 days of a positive test in the community. Correcting for misclassification of SGTF and missingness in SGTF status, we estimate that the hazard of death associated with B.1.1.7 is 61% (42-82%) higher than with pre-existing variants. Our analysis suggests that B.1.1.7 is not only more transmissible than pre-existing SARS-CoV-2 variants, but may also cause more severe illness.
Assuntos
COVID-19/mortalidade , COVID-19/virologia , Filogenia , SARS-CoV-2/classificação , SARS-CoV-2/patogenicidade , Adolescente , Adulto , Idoso , Idoso de 80 Anos ou mais , Criança , Pré-Escolar , Inglaterra/epidemiologia , Etnicidade , Evolução Molecular , Feminino , Instituição de Longa Permanência para Idosos , Humanos , Lactente , Masculino , Pessoa de Meia-Idade , Modelos de Riscos Proporcionais , Medição de Risco , SARS-CoV-2/genética , SARS-CoV-2/isolamento & purificação , Análise de Sobrevida , Fatores de Tempo , Adulto JovemRESUMO
Multi-state models are used to describe how individuals transition through different states over time. The distribution of the time spent in different states, referred to as 'length of stay', is often of interest. Methods for estimating expected length of stay in a given state are well established. The focus of this paper is on the distribution of the time spent in different states conditional on the complete pathway taken through the states, which we call 'conditional length of stay'. This work is motivated by questions about length of stay in hospital wards and intensive care units among patients hospitalised due to Covid-19. Conditional length of stay estimates are useful as a way of summarising individuals' transitions through the multi-state model, and also as inputs to mathematical models used in planning hospital capacity requirements. We describe non-parametric methods for estimating conditional length of stay distributions in a multi-state model in the presence of censoring, including conditional expected length of stay (CELOS). Methods are described for an illness-death model and then for the more complex motivating example. The methods are assessed using a simulation study and shown to give unbiased estimates of CELOS, whereas naive estimates of CELOS based on empirical averages are biased in the presence of censoring. The methods are applied to estimate conditional length of stay distributions for individuals hospitalised due to Covid-19 in the UK, using data on 42,980 individuals hospitalised from March to July 2020 from the COVID19 Clinical Information Network.
Assuntos
COVID-19 , Tempo de Internação , Humanos , Unidades de Terapia Intensiva , Masculino , Feminino , Simulação por ComputadorRESUMO
BACKGROUND: Cystic fibrosis (CF) is a life-threatening genetic disease, affecting around 10 500 people in the UK. Precision medicines have been developed to treat specific CF-gene mutations. The newest, elexacaftor/tezacaftor/ivacaftor (ELEX/TEZ/IVA), has been found to be highly effective in randomised controlled trials (RCTs) and became available to a large proportion of UK CF patients in 2020. Understanding the potential health economic impacts of ELEX/TEZ/IVA is vital to planning service provision. METHODS: We combined observational UK CF Registry data with RCT results to project the impact of ELEX/TEZ/IVA on total days of intravenous (IV) antibiotic treatment at a population level. Registry data from 2015 to 2017 were used to develop prediction models for IV days over a 1-year period using several predictors, and to estimate 1-year population total IV days based on standards of care pre-ELEX/TEZ/IVA. We considered two approaches to imposing the impact of ELEX/TEZ/IVA on projected outcomes using effect estimates from RCTs: approach 1 based on effect estimates on FEV1% and approach 2 based on effect estimates on exacerbation rate. RESULTS: ELEX/TEZ/IVA is expected to result in significant reductions in population-level requirements for IV antibiotics of 16.1% (~17 800 days) using approach 1 and 43.6% (~39 500 days) using approach 2. The two approaches require different assumptions. Increased understanding of the mechanisms through which ELEX/TEZ/IVA acts on these outcomes would enable further refinements to our projections. CONCLUSIONS: This work contributes to increased understanding of the changing healthcare needs of people with CF and illustrates how Registry data can be used in combination with RCT evidence to estimate population-level treatment impacts.
Assuntos
Fibrose Cística , Aminofenóis/uso terapêutico , Antibacterianos/uso terapêutico , Benzodioxóis/uso terapêutico , Fibrose Cística/tratamento farmacológico , Fibrose Cística/genética , Regulador de Condutância Transmembrana em Fibrose Cística/genética , Humanos , Mutação , Estudos Observacionais como Assunto , Ensaios Clínicos Controlados Aleatórios como Assunto , Sistema de RegistrosRESUMO
Clinical prediction models (CPMs) are often used to guide treatment initiation, with individuals at high risk offered treatment. This implicitly assumes that the probability quoted from a CPM represents the risk to an individual of an adverse outcome in absence of treatment. However, for a CPM to correctly target this estimand requires careful causal thinking. One problem that needs to be overcome is treatment drop-in: where individuals in the development data commence treatment after the time of prediction but before the outcome occurs. In this issue of the Journal, Xu et al. (Am J Epidemiol. 2021;190(10):2000-2014) use causal estimates from external data sources, such as clinical trials, to adjust CPMs for treatment drop-in. This represents a pragmatic and promising approach to address this issue, and it illustrates the value of utilizing causal inference in prediction. Building causality into the prediction pipeline can also bring other benefits. These include the ability to make and compare hypothetical predictions under different interventions, to make CPMs more explainable and transparent, and to improve model generalizability. Enriching CPMs with causal inference therefore has the potential to add considerable value to the role of prediction in healthcare.
Assuntos
Causalidade , Humanos , ProbabilidadeRESUMO
BACKGROUND: Cross-sectional measures of body mass index (BMI) are associated with cardiovascular disease (CVD) incidence, but less is known about whether weight change affects the risk of CVD. METHODS: We estimated the effect of 2-y weight change interventions on 7-y risk of CVD (CVD death, myocardial infarction, stroke, hospitalization from coronary heart disease, and heart failure) by emulating hypothetical interventions using electronic health records. We identified 138,567 individuals with 45-69 years of age without chronic disease in England from 1998 to 2016. We performed pooled logistic regression, using inverse-probability weighting to adjust for baseline and time-varying confounders. We categorized each individual into a weight loss, maintenance, or gain group. RESULTS: Among those of normal weight, both weight loss [risk difference (RD) vs. weight maintenance = 1.5% (0.3% to 3.0%)] and gain [RD = 1.3% (0.5% to 2.2%)] were associated with increased risk for CVD compared with weight maintenance. Among overweight individuals, we observed moderately higher risk of CVD in both the weight loss [RD = 0.7% (-0.2% to 1.7%)] and the weight gain group [RD = 0.7% (-0.1% to 1.7%)], compared with maintenance. In the obese, those losing weight showed lower risk of coronary heart disease [RD = -1.4% (-2.4% to -0.6%)] but not of stroke. When we assumed that chronic disease occurred 1-3 years before the recorded date, estimates for weight loss and gain were attenuated among overweight individuals; estimates for loss were lower among obese individuals. CONCLUSION: Among individuals with obesity, the weight-loss group had a lower risk of coronary heart disease but not of stroke. Weight gain was associated with increased risk of CVD across BMI groups. See video abstract at, http://links.lww.com/EDE/B838.
Assuntos
Doenças Cardiovasculares , Índice de Massa Corporal , Doenças Cardiovasculares/epidemiologia , Estudos Transversais , Registros Eletrônicos de Saúde , Humanos , Sobrepeso/epidemiologia , Fatores de RiscoRESUMO
BACKGROUND: Predicting bed occupancy for hospitalised patients with COVID-19 requires understanding of length of stay (LoS) in particular bed types. LoS can vary depending on the patient's "bed pathway" - the sequence of transfers of individual patients between bed types during a hospital stay. In this study, we characterise these pathways, and their impact on predicted hospital bed occupancy. METHODS: We obtained data from University College Hospital (UCH) and the ISARIC4C COVID-19 Clinical Information Network (CO-CIN) on hospitalised patients with COVID-19 who required care in general ward or critical care (CC) beds to determine possible bed pathways and LoS. We developed a discrete-time model to examine the implications of using either bed pathways or only average LoS by bed type to forecast bed occupancy. We compared model-predicted bed occupancy to publicly available bed occupancy data on COVID-19 in England between March and August 2020. RESULTS: In both the UCH and CO-CIN datasets, 82% of hospitalised patients with COVID-19 only received care in general ward beds. We identified four other bed pathways, present in both datasets: "Ward, CC, Ward", "Ward, CC", "CC" and "CC, Ward". Mean LoS varied by bed type, pathway, and dataset, between 1.78 and 13.53 days. For UCH, we found that using bed pathways improved the accuracy of bed occupancy predictions, while only using an average LoS for each bed type underestimated true bed occupancy. However, using the CO-CIN LoS dataset we were not able to replicate past data on bed occupancy in England, suggesting regional LoS heterogeneities. CONCLUSIONS: We identified five bed pathways, with substantial variation in LoS by bed type, pathway, and geography. This might be caused by local differences in patient characteristics, clinical care strategies, or resource availability, and suggests that national LoS averages may not be appropriate for local forecasts of bed occupancy for COVID-19. TRIAL REGISTRATION: The ISARIC WHO CCP-UK study ISRCTN66726260 was retrospectively registered on 21/04/2020 and designated an Urgent Public Health Research Study by NIHR.
Assuntos
Ocupação de Leitos , COVID-19 , Inglaterra , Humanos , Tempo de Internação , SARS-CoV-2RESUMO
The increasing availability of genetic cohort data has led to many genome-wide association studies (GWAS) successfully identifying genetic associations with an ever-expanding list of phenotypic traits. Association, however, does not imply causation, and therefore methods have been developed to study the issue of causality. Under additional assumptions, Mendelian randomization (MR) studies have proved popular in identifying causal effects between two phenotypes, often using GWAS summary statistics. Given the widespread use of these methods, it is more important than ever to understand, and communicate, the causal assumptions upon which they are based, so that methods are transparent, and findings are clinically relevant. Causal graphs can be used to represent causal assumptions graphically and provide insights into the limitations associated with different analysis methods. Here we review GWAS and MR from a causal perspective, to build up intuition for causal diagrams in genetic problems. We also examine issues of confounding by ancestry and comment on approaches for dealing with such confounding, as well as discussing approaches for dealing with selection biases arising from study design.
Assuntos
Estudo de Associação Genômica Ampla/métodos , Análise da Randomização Mendeliana/métodos , Neoplasias/genética , Causalidade , Estudos de Coortes , Variação Genética , Humanos , Modelos Estatísticos , FenótipoRESUMO
Two paradigms for the evaluation of surrogate markers in randomized clinical trials have been proposed: the causal effects paradigm and the causal association paradigm. Each of these paradigms rely on assumptions that must be made to proceed with estimation and to validate a candidate surrogate marker (S) for the true outcome of interest (T). We consider the setting in which S and T are Gaussian and are generated from structural models that include an unobserved confounder. Under the assumed structural models, we relate the quantities used to evaluate surrogacy within both the causal effects and causal association frameworks. We review some of the common assumptions made to aid in estimating these quantities and show that assumptions made within one framework can imply strong assumptions within the alternative framework. We demonstrate that there is a similarity, but not exact correspondence between the quantities used to evaluate surrogacy within each framework, and show that the conditions for identifiability of the surrogacy parameters are different from the conditions, which lead to a correspondence of these quantities.
Assuntos
Biomarcadores , Causalidade , Distribuição Normal , Interpretação Estatística de Dados , Humanos , Modelos Estatísticos , Ensaios Clínicos Controlados Aleatórios como Assunto/métodosRESUMO
BACKGROUND: Depression is common and is associated with poor outcomes among elderly care-home residents. Exercise is a promising low-risk intervention for depression in this population. We tested the hypothesis that a moderate intensity exercise programme would reduce the burden of depressive symptoms in residents of care homes. METHODS: We did a cluster-randomised controlled trial in care homes in two regions in England; northeast London, and Coventry and Warwickshire. Residents aged 65 years or older were eligible for inclusion. A statistician independent of the study randomised each home (1 to 1·5 ratio, stratified by location, minimised by type of home provider [local authority, voluntary, private and care home, private and nursing home] and size of home [<32 or ≥32 residents]) into intervention and control groups. The intervention package included depression awareness training for care-home staff, 45 min physiotherapist-led group exercise sessions for residents (delivered twice weekly), and a whole home component designed to encourage more physical activity in daily life. The control consisted of only the depression awareness training. Researchers collecting follow-up data from individual participants and the participants themselves were inevitably aware of home randomisation because of the physiotherapists' activities within the home. A researcher masked to study allocation coded NHS routine data. The primary outcome was number of depressive symptoms on the geriatric depression scale-15 (GDS-15). Follow-up was for 12 months. This trial is registered with ISRCTN Register, number ISRCTN43769277. FINDINGS: Care homes were randomised between Dec 15, 2008, and April 9, 2010. At randomisation, 891 individuals in 78 care homes (35 intervention, 43 control) had provided baseline data. We delivered 3191 group exercise sessions attended on average by five study participants and five non-study residents. Of residents with a GDS-15 score, 374 of 765 (49%) were depressed at baseline; 484 of 765 (63%) provided 12 month follow-up scores. Overall the GDS-15 score was 0·13 (95% CI -0·33 to 0·60) points higher (worse) at 12 months for the intervention group compared with the control group. Among residents depressed at baseline, GDS-15 score was 0·22 (95% CI -0·52 to 0·95) points higher at 6 months in the intervention group than in the control group. In an end of study cross-sectional analysis, including 132 additional residents joining after randomisation, the odds of being depressed were 0·76 (95% CI 0·53 to 1·09) for the intervention group compared with the control group. INTERPRETATION: This moderately intense exercise programme did not reduce depressive symptoms in residents of care homes. In this frail population, alternative strategies to manage psychological symptoms are required. FUNDING: National Institute for Health Research Health Technology Assessment.
Assuntos
Depressão/reabilitação , Terapia por Exercício/métodos , Idoso , Idoso de 80 Anos ou mais , Análise por Conglomerados , Estudos Transversais , Inglaterra , Feminino , Instituição de Longa Permanência para Idosos , Humanos , Masculino , Casas de Saúde , Resultado do TratamentoRESUMO
BACKGROUND: Missing data are a potential source of bias, and their handling in the statistical analysis can have an important impact on both the likelihood and degree of such bias. Inadequate handling of the missing data may also result in invalid variance estimation. The handling of missing values is more complex in cluster randomised trials, but there are no reviews of practice in this field. OBJECTIVES: A systematic review of published trials was conducted to examine how missing data are reported and handled in cluster randomised trials. METHODS: We systematically identified cluster randomised trials, published in English in 2011, using the National Library of Medicine (MEDLINE) via PubMed. Non-randomised and pilot/feasibility trials were excluded, as were reports of secondary analyses, interim analyses, and economic evaluations and those where no data were at the individual level. We extracted information on missing data and the statistical methods used to deal with them from a random sample of the identified studies. RESULTS: We included 132 trials. There was evidence of missing data in 95 (72%). Only 32 trials reported handling missing data, 22 of them using a variety of single imputation techniques, 8 using multiple imputation without accommodating the clustering and 2 stating that their likelihood-based complete case analysis accounted for missing values because the data were assumed Missing-at-Random. LIMITATIONS: The results presented in this study are based on a large random sample of cluster randomised trials published in 2011, identified in electronic searches and therefore possibly missing some trials, most likely of poorer quality. Also, our results are based on information in the main publication for each trial. These reports may omit some important information on the presence of, and reasons for, missing data and on the statistical methods used to handle them. Our extraction methods, based on published reports, could not distinguish between missing data in outcomes and missing data in covariates. This distinction may be important in determining the assumptions about the missing data mechanism necessary for complete case analyses to be valid. CONCLUSIONS: Missing data are present in the majority of cluster randomised trials. However, they are poorly reported, and most authors give little consideration to the assumptions under which their analysis will be valid. The majority of the methods currently used are valid under very strong assumptions about the missing data, whose plausibility is rarely discussed in the corresponding reports. This may have important consequences for the validity of inferences in some trials. Methods which result in valid inferences under general Missing-at-Random assumptions are available and should be made more accessible.
Assuntos
Guias como Assunto , Ensaios Clínicos Controlados Aleatórios como Assunto , Estatística como Assunto , Interpretação Estatística de Dados , Humanos , Projetos de PesquisaRESUMO
Target trial emulation applies design principles from randomised controlled trials to the analysis of observational data for causal inference and is increasingly used within human epidemiology. Using anonymised veterinary clinical data from the VetCompass Programme, this study applied the target trial emulation framework to determine whether surgical (compared to non-surgical) management for cranial cruciate ligament (CCL) rupture in dogs causes improved short- and long-term lameness and analgesia outcomes. The emulated target trial included dogs diagnosed with CCL rupture between January 1, 2019 and December 31, 2019 within the VetCompass database. Inclusion in the emulated trial required dogs aged ≥ 1.5 and < 12 years, first diagnosed with unilateral CCL rupture during 2019 and with no prior history of contralateral ligament rupture or stifle surgery. Dogs were retrospectively observed to have surgical or non-surgical management. Informed from a directed acyclic graph derived from expert opinion, data on the following variables were collected: age, breed, bodyweight, neuter status, insurance status, non-orthopaedic comorbidities, orthopaedic comorbidities and veterinary group. Inverse probability of treatment weighting (IPTW) was used to adjust for confounding, with weights calculated based on a binary logistic regression exposure model. Censored dogs were accounted for in the IPTW analysis using inverse probability of censoring weighting (IPCW). The IPCWs were combined with IPTWs and used to weight each dog's contribution to binary logistic regression outcome models. Standardized mean differences (SMD) examined the balance of covariate distribution between treatment groups. The emulated trial included 615 surgical CCL rupture cases and 200 non-surgical cases. The risk difference for short-term lameness in surgically managed cases (compared with non-surgically managed cases) was -25.7% (95% confidence interval (CI) -36.7% to -15.9%) and the risk difference for long-term lameness -31.7% (95% CI -37.9% to -18.1%). The study demonstrated the application of the target trial framework to veterinary observational data. The findings show that surgical management causes a reduction in short- and long-term lameness compared with non-surgical management in dogs.
Assuntos
Lesões do Ligamento Cruzado Anterior , Doenças do Cão , Coxeadura Animal , Animais , Cães , Lesões do Ligamento Cruzado Anterior/veterinária , Lesões do Ligamento Cruzado Anterior/cirurgia , Doenças do Cão/terapia , Doenças do Cão/cirurgia , Coxeadura Animal/etiologia , Ruptura/veterinária , Ruptura/cirurgia , Estudos Retrospectivos , Masculino , Feminino , Ligamento Cruzado Anterior/cirurgia , Resultado do TratamentoRESUMO
There is growing evidence supporting clinically important associations between age at neutering in bitches and subsequent urinary incontinence (UI), although much of this evidence to date is considered weak. Target trial emulation is an innovative approach in causal inference that has gained substantial attention in recent years, aiming to simulate a hypothetical randomised controlled trial by leveraging observational data. Using anonymised veterinary clinical data from the VetCompass Programme, this study applied the target trial emulation framework to determine whether later-age neutering (≥ 7 to ≤ 18 months) causes decreased odds of early-onset UI (diagnosed < 8.5 years) compared to early-age neutering (3 to < 7 months). The study included bitches in the VetCompass database born from January 1, 2010, to December 31, 2012, and neutered between 3 and 18 months old. Bitches were retrospectively confirmed from the electronic health records as neutered early or later. The primary outcome was a diagnosis of early-onset UI. Informed from a directed acyclic graph, data on the following covariates were extracted: breed, insurance status, co-morbidities and veterinary group. Inverse probability of treatment weighting was used to adjust for confounding, with inverse probability of censoring weighting accounting for censored bitches. The emulated trial included 612 early-age neutered bitches and 888 later-age neutered bitches. A pooled logistic regression outcome model identified bitches neutered later at 0.80 times the odds (95% CI 0.54 to 0.97) of early-onset UI compared with bitches neutered early. The findings show that later-age neutering causes reduced odds of early-onset UI diagnosis compared with early-age neutering. Decision-making on the age of neutering should be carefully considered, with preference given to delaying neutering until after 7 months of age unless other major reasons justify earlier surgery. The study is one of the first to demonstrate successful application of the target trial framework to veterinary observational data.
Assuntos
Doenças do Cão , Incontinência Urinária , Animais , Cães , Feminino , Incontinência Urinária/veterinária , Incontinência Urinária/epidemiologia , Incontinência Urinária/etiologia , Doenças do Cão/epidemiologia , Fatores Etários , Estudos Retrospectivos , Castração/veterinária , Fatores de RiscoRESUMO
BACKGROUND: Previous reviews of cluster randomised trials have been critical of the quality of the trials reviewed, but none has explored determinants of the quality of these trials in a specific field over an extended period of time. Recent work suggests that correct conduct and reporting of these trials may require more than published guidelines. In this review, our aim was to assess the quality of cluster randomised trials conducted in residential facilities for older people, and to determine whether (1) statistician involvement in the trial and (2) strength of journal endorsement of the Consolidated Standards of Reporting Trials (CONSORT) statement influence quality. METHODS: We systematically identified trials randomising residential facilities for older people, or parts thereof, without language restrictions, up to the end of 2010, using National Library of Medicine (Medline) via PubMed and hand-searching. We based quality assessment criteria largely on the extended CONSORT statement for cluster randomised trials. We assessed statistician involvement based on statistician co-authorship, and strength of journal endorsement of the CONSORT statement from journal websites. RESULTS: 73 trials met our inclusion criteria. Of these, 20 (27%) reported accounting for clustering in sample size calculations and 54 (74%) in the analyses. In 29 trials (40%), methods used to identify/recruit participants were judged by us to have potentially caused bias or reporting was unclear to reach a conclusion. Some elements of quality improved over time but this appeared not to be related to the publication of the extended CONSORT statement for these trials. Trials with statistician/epidemiologist co-authors were more likely to account for clustering in sample size calculations (unadjusted odds ratio 5.4, 95% confidence interval 1.1 to 26.0) and analyses (unadjusted OR 3.2, 1.2 to 8.5). Journal endorsement of the CONSORT statement was not associated with trial quality. CONCLUSIONS: Despite international attempts to improve methods in cluster randomised trials, important quality limitations remain amongst these trials in residential facilities. Statistician involvement on trial teams may be more effective in promoting quality than further journal endorsement of the CONSORT statement. Funding bodies and journals should promote statistician involvement and co-authorship in addition to adherence to CONSORT guidelines.
Assuntos
Instituição de Longa Permanência para Idosos/normas , Análise por Conglomerados , Interpretação Estatística de Dados , Humanos , Melhoria de Qualidade , Ensaios Clínicos Controlados Aleatórios como Assunto , Relatório de Pesquisa/normasRESUMO
Adjustment for baseline covariates in randomized trials has been shown to lead to gains in power and can protect against chance imbalances in covariates. For continuous covariates, there is a risk that the the form of the relationship between the covariate and outcome is misspecified when taking an adjusted approach. Using a simulation study focusing on individually randomized trials with small sample sizes, we explore whether a range of adjustment methods are robust to misspecification, either in the covariate-outcome relationship or through an omitted covariate-treatment interaction. Specifically, we aim to identify potential settings where G-computation, inverse probability of treatment weighting (IPTW), augmented inverse probability of treatment weighting (AIPTW) and targeted maximum likelihood estimation (TMLE) offer improvement over the commonly used analysis of covariance (ANCOVA). Our simulations show that all adjustment methods are generally robust to model misspecification if adjusting for a few covariates, sample size is 100 or larger, and there are no covariate-treatment interactions. When there is a non-linear interaction of treatment with a skewed covariate and sample size is small, all adjustment methods can suffer from bias; however, methods that allow for interactions (such as G-computation with interaction and IPTW) show improved results compared to ANCOVA. When there are a high number of covariates to adjust for, ANCOVA retains good properties while other methods suffer from under- or over-coverage. An outstanding issue for G-computation, IPTW and AIPTW in small samples is that standard errors are underestimated; they should be used with caution without the availability of small-sample corrections, development of which is needed. These findings are relevant for covariate adjustment in interim analyses of larger trials.
Assuntos
Ensaios Clínicos Controlados Aleatórios como Assunto , Humanos , Simulação por Computador , Probabilidade , Tamanho da AmostraRESUMO
BACKGROUND: Over time, the performance of clinical prediction models may deteriorate due to changes in clinical management, data quality, disease risk and/or patient mix. Such prediction models must be updated in order to remain useful. In this study, we investigate dynamic model updating of clinical survival prediction models. In contrast to discrete or one-time updating, dynamic updating refers to a repeated process for updating a prediction model with new data. We aim to extend previous research which focused largely on binary outcome prediction models by concentrating on time-to-event outcomes. We were motivated by the rapidly changing environment seen during the COVID-19 pandemic where mortality rates changed over time and new treatments and vaccines were introduced. METHODS: We illustrate three methods for dynamic model updating: Bayesian dynamic updating, recalibration, and full refitting. We use a simulation study to compare performance in a range of scenarios including changing mortality rates, predictors with low prevalence and the introduction of a new treatment. Next, the updating strategies were applied to a model for predicting 70-day COVID-19-related mortality using patient data from QResearch, an electronic health records database from general practices in the UK. RESULTS: In simulated scenarios with mortality rates changing over time, all updating methods resulted in better calibration than not updating. Moreover, dynamic updating outperformed ad hoc updating. In the simulation scenario with a new predictor and a small updating dataset, Bayesian updating improved the C-index over not updating and refitting. In the motivating example with a rare outcome, no single updating method offered the best performance. CONCLUSIONS: We found that a dynamic updating process outperformed one-time discrete updating in the simulations. Bayesian updating offered good performance overall, even in scenarios with new predictors and few events. Intercept recalibration was effective in scenarios with smaller sample size and changing baseline hazard. Refitting performance depended on sample size and produced abrupt changes in hazard ratio estimates between periods.
RESUMO
Target trial emulation applies design principles from randomised controlled trials to the analysis of observational data for causal inference and is increasingly used within human epidemiology. Veterinary electronic clinical records represent a potentially valuable source of information to estimate real-world causal effects for companion animal species. This study employed the target trial framework to evaluate the usefulness on veterinary observational data. Acute diarrhoea in dogs was used as a clinical exemplar. Inclusion required dogs aged ≥ 3 months and < 10 years, presenting for veterinary primary care with acute diarrhoea during 2019. Treatment strategies were: 1. antimicrobial prescription compared to no antimicrobial prescription and 2. gastrointestinal nutraceutical prescription compared to no gastrointestinal nutraceutical prescription. The primary outcome was clinical resolution (defined as no revisit with ongoing diarrhoea within 30 days from the date of first presentation). Informed from a directed acyclic graph, data on the following covariates were collected: age, breed, bodyweight, insurance status, comorbidities, vomiting, reduced appetite, haematochezia, pyrexia, duration, additional treatment prescription and veterinary group. Inverse probability of treatment weighting was used to balance covariates between the treatment groups for each of the two target trials. The risk difference (RD) of 0.4% (95% CI -4.5% to 5.3%) was non-significant for clinical resolution in dogs treated with antimicrobials compared with dogs not treated with antimicrobials. The risk difference (RD) of 0.3% (95% CI -4.5% to 5.0%) was non-significant for clinical resolution in dogs treated with gastrointestinal nutraceuticals compared with dogs not treated with gastrointestinal nutraceuticals. This study successfully applied the target trial framework to veterinary observational data. The findings show that antimicrobial or gastrointestinal prescription at first presentation of acute diarrhoea in dogs causes no difference in clinical resolution. The findings support the recommendation for veterinary professionals to limit antimicrobial use for acute diarrhoea in dogs.
Assuntos
Anti-Infecciosos , Animais , Cães , Anti-Infecciosos/uso terapêutico , Diarreia/tratamento farmacológico , Diarreia/veterinária , Diarreia/epidemiologia , Prescrições , Reino Unido/epidemiologia , VômitoRESUMO
HIGHLIGHTS: This article examines a causal machine-learning approach, causal forests (CF), for exploring the heterogeneity of treatment effects, without prespecifying a specific functional form.The CF approach is considered in the reanalysis of the 65 Trial and was found to provide similar estimates of subgroup effects to using a fixed parametric model.The CF approach also provides estimates of individual-level treatment effects that suggest that for most patients in the 65 Trial, the intervention is expected to reduce 90-d mortality but with wide levels of statistical uncertainty.The study illustrates how individual-level treatment effect estimates can be analyzed to generate hypotheses for further research about those patients who are likely to benefit most from an intervention.
Assuntos
Aprendizado de Máquina , HumanosRESUMO
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.
RESUMO
We consider mediated effects of an exposure, X on an outcome, Y, via a mediator, M, under no unmeasured confounding assumptions in the setting where models for the conditional expectation of the mediator and outcome are partially linear. We propose G-estimators for the direct and indirect effects and demonstrate consistent asymptotic normality for indirect effects when models for the conditional means of M, or X and Y are correctly specified, and for direct effects, when models for the conditional means of Y, or X and M are correct. This marks an improvement, in this particular setting, over previous 'triple' robust methods, which do not assume partially linear mean models. Testing of the no-mediation hypothesis is inherently problematic due to the composite nature of the test (either X has no effect on M or M no effect on Y), leading to low power when both effect sizes are small. We use generalized methods of moments (GMM) results to construct a new score testing framework, which includes as special cases the no-mediation and the no-direct-effect hypotheses. The proposed tests rely on an orthogonal estimation strategy for estimating nuisance parameters. Simulations show that the GMM-based tests perform better in terms of power and small sample performance compared with traditional tests in the partially linear setting, with drastic improvement under model misspecification. New methods are illustrated in a mediation analysis of data from the COPERS trial, a randomized trial investigating the effect of a non-pharmacological intervention of patients suffering from chronic pain. An accompanying R package implementing these methods can be found at github.com/ohines/plmed.