Rationale: Modulator therapies like ivacaftor have revolutionized clinical management of cystic fibrosis, showing marked short-term benefits in trials but heterogeneous findings in long-term observational studies. Since newer modulators have become the standard of care for the majority living with cystic fibrosis in the United States, characterizing long-term effectiveness with real-world data is increasingly difficult because of the lack of contemporary comparator groups for performing between-subjects analyses.Objectives: To determine the extent to which ivacaftor preserves long-term lung function and compare the results of within- and between-subjects analyses for evaluating its real-world effectiveness.Methods: This retrospective cohort study used data from the U.S. Cystic Fibrosis Foundation Patient Registry (2003-2016). We used two approaches to evaluate ivacaftor effectiveness on percent predicted forced expiratory volume in 1 second (ppFEV1): 1) within-subject comparisons of ppFEV1 before and after ivacaftor initiation; and 2) comparisons between ivacaftor-treated and untreated individuals with similar disease pathology. We modeled data from 560 ivacaftor-treated individuals with the G551D variant. For between-subjects comparisons, we used propensity scores to match the treated group with 2,800 untreated F508del homozygous individuals. Modulator initiation bias was assessed and accounted for in each model.Results: Our results showed an initial average improvement in ppFEV1 in ivacaftor-treated children and adults (ranging from 4.54% to 6.53% predicted based on within-subject comparison of before vs. after ivacaftor initiation). There was a slower decline in adults, compared with children. These ivacaftor-treated cohorts experienced less decline relative to their F508del homozygous counterparts (between-group differences in treated vs. control ranged from 0.36% to 0.64% predicted). Both the within- and between-subjects comparisons demonstrated similar degrees of ivacaftor effectiveness. However, small differences between the two approaches were observed in younger individuals.Conclusions: Ivacaftor was associated with improved ppFEV1 across all age groups, with the magnitude of improvement roughly 50% of that observed in clinical trials. The results support the need to account for modulator initiation bias and the use of within-subject analysis in future CFTR (cystic fibrosis transmembrane conductance regulator) modulator effectiveness studies, but caution is advised in younger individuals because of developmental changes that may affect pre- and post-treatment comparability.

Background There is growing interest in widening the use of sodium-glucose co-transporter 2 inhibitors (SGLT2i) to all people with type 2 diabetes mellitus (T2DM). However, pivotal randomized controlled trials (RCTs) evaluated these drugs only in highly selected populations, often lacking generalizability to real-world populations. Understanding the effects of SGLT2i in populations where RCT evidence may be lacking is essential to help inform guideline development. To address this, we estimated the effect of empagliflozin in real-world users, many of whom would not have been eligible for the pivotal EMPA-REG RCT.Methods We designed a trial emulation in UK primary care data, based on the EMPA-REG RCT, to assess the effect of empagliflozin in a more clinically relevant population. Adults with T2DM initiating empagliflozin (intervention) or dipeptidyl peptidase-4 inhibitors (active control) between January 1, 2014 and December 31, 2022 were included. Eligibility was extended to both RCT-eligible and RCT-ineligible individuals. The effect of empagliflozin on all-cause mortality was estimated using an adjusted Cox proportional hazards model, with stratified analyses by RCT eligibility.Findings The majority of people prescribed empagliflozin would not have met the EMPA-REG RCT eligibility criteria (11,011/13,239, 83.2% RCT-ineligible). During follow-up, all-cause mortality occurred in 551 out of 13,239 (4.2%) in the empagliflozin group and 6,589 out of 49,264 (13.4%) in the active control group (adjusted HR 0.76, 95% CI 0.69 to 0.83). There was no evidence of differential treatment effect by RCT eligibility status (p-interaction=0.27).Interpretation Patients prescribed empagliflozin in real-world settings differ substantially from those enrolled in the EMPA-REG RCT. Using electronic health records, we demonstrate that the mortality benefit observed in EMPA-REG extends to a broader, more diverse real-world population, including those excluded from the original RCT. These findings provide a novel source of real-world evidence supporting the wider use of empagliflozin in routine clinical practice.

Within-individual variability of health indicators measured over time is becoming commonly used to inform about disease progression. Simple summary statistics (e.g., the standard deviation for each individual) are often used but they are not suited to account for time changes. In addition, when these summary statistics are used as covariates in a regression model for time-to-event outcomes, the estimates of the hazard ratios are subject to regression dilution. To overcome these issues, a joint model is built where the association between the time-to-event outcome and multivariate longitudinal markers is specified in terms of the within-individual variability of the latter. A mixed-effect location-scale model is used to analyze the longitudinal biomarkers, their within-individual variability and their correlation. The time to event is modeled using a proportional hazard regression model, with a flexible specification of the baseline hazard, and the information from the longitudinal biomarkers is shared as a function of the random effects. The model can be used to quantify within-individual variability for the longitudinal markers and their association with the time-to-event outcome. We show through a simulation study the performance of the model in comparison with standard joint models with constant variance. The model is applied on a dataset of adult women from the UK cystic fibrosis registry, to evaluate the association between lung function, malnutrition and mortality.

Abstract Background Less than 20% of patients diagnosed with advanced lung cancer will survive beyond five years and half of these will suffer a serious adverse event (SAE) caused by systemic anticancer therapy (SACT) that will result in a hospital attendance. As multiple different SACT treatments are available for patients, a risk score that predicts the likelihood of a SAE following each type of SACT treatment would improve both communication with the patient and shared decision making with all those involved in delivering care for patients. There are currently no risk scores available for use in those with advanced stage lung cancer. Aim The overarching aim of this research is to develop and internally validate a risk score that will calculate the individualised risk of SAEs for different SACT treatments for patients with late stage lung cancer. Methods Utilising linked cancer registry data (National Cancer Registration and Analysis Service (NCRAS), England) for over 20,000 late stage lung cancer patients, a risk score will be developed using a multivariable logistic regression model to predict the risk of an acute admission within 30 days of SACT administration. Model performance will be summarised using calibration and discrimination. Internal validation will be used to quantify the degree of optimism due to overfitting, using re-sampling bootstrapping. Heterogeneity will be assessed, and the model will be fine-tuned. Fine-tuning and interrogation will be used to evaluate differences in performance between hospitals. The clinical utility will be assessed through calculating the net benefit in preventing SAEs. Conclusion A developed risk score (under each treatment strategy) has real potential to support individualised treatment decisions and optimise management of SACT-induced SAEs for patients and reduce hospital attendances.

Background We investigated frailty progression after severe infections in adults (65+ years) in the US and England. Methods We conducted parallel matched cohort studies using: US Veterans Aging Cohort Study (VACS-National, 2008-2019; median age 74 years; 98% male); and English Clinical Practice Research Datalink (2006-2019; median age 76 years; 45% male). Adults hospitalised primarily for infection (i.e., severe infection) were matched in calendar date order to individuals without severe infection on age, sex, care site, and US only, plus race and ethnicity. We measured frailty using VACS Index 2.0 (US) and Electronic Frailty Index (eFI; England). We estimated annual conditional mean frailty differences between adults with versus without severe infection using linear regression adjusting for baseline frailty, demographics, lifestyle factors, infection history, and US only, comorbidities. Results Mean baseline frailty was higher in those with severe infection than those without (US: 57 v 48; England: 0.17 v 0.12). At Year 1, adjusted mean frailty was higher among adults with severe infections than those without (US: VACS Index +2.0, 95% CI 1.9-2.0; England: eFI +0.005, 95% CI 0.005-0.006). At Years 2-5, adjusted mean frailty remained higher after severe infection; however, compared to Year 1, differences were smaller in US, and larger in England. Effects varied by infection type (strongest for lower respiratory tract infections, meningoencephalitis (UK only), urinary tract infections, and sepsis). Interpretation Individuals with severe infections had higher frailty at baseline and follow up than those without. Preventing both frailty and infections is important for improving health in older age. ### Competing Interest Statement The authors have declared no competing interest. ### Funding Statement Funding for this study is through a Wellcome Career Development Award (225868/Z/22/Z) received by CWG. The Veterans Aging Cohort Study is supported by the National Institute on Alcohol Abuse and Alcoholism [P01-AA029545, U01-AA026224, U24-AA020794, U01-AA020790, U10-AA013566]. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: Both analyses were approved by the London School of Hygiene & Tropical Medicine Research Ethics Committee (References: US, 31029; England, 31298). Additionally, the US study was approved by the institutional review boards of Yale University (Reference: #1506016006) and VA Connecticut Healthcare System (Reference: #AJ0013), and the English study by CPRD Independent Scientific Advisory Committee (Reference: 24_004305). I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes We do not have the rights to publicly share the data. However, Due to US Department of Veterans Affairs (VA) regulations and our ethics agreements, the analytic data sets used for this study are not permitted to leave the VA firewall without a data use agreement. This limitation is consistent with other studies based on VA data. However, VA data are made freely available to researchers with an approved VA study protocol. For more information, please visit https://www.virec.research.va.gov or contact the VA Information Resource Center at virec{at}va.gov. Data used for England cohort are available on request from CPRD. The code used to clean, and analyse the data are publicly available on the GitHub repositories (US, England)

G-formula is a popular approach for estimating the effects of time-varying treatments or exposures from longitudinal data. G-formula is typically implemented using Monte-Carlo simulation, with non-parametric bootstrapping used for inference. In longitudinal data settings missing data are a common issue, which are often handled using multiple imputation, but it is unclear how G-formula and multiple imputation should be combined. We show how G-formula can be implemented using Bayesian multiple imputation methods for synthetic data, and that by doing so, we can impute missing data and simulate the counterfactuals of interest within a single coherent approach. We describe how this can be achieved using standard multiple imputation software and explore its performance using a simulation study and an application from cystic fibrosis.

Background: Elexacaftor, Tezacaftor, Ivacaftor (ETI) became available in the UK in August 2020 to treat people with Cystic Fibrosis (CF) aged >12 years. We report a real-world study of clinical outcomes in young people treated with ETI at our CF centre within the first two years of its availability. Methods: Participants aged 12-17 were identified within our clinic, with demographic data supplemented by the UK CF registry. Comprehensive outcome data spanning two years pre- and two years post-initiation of CFTR modulators were compiled from various local sources, including patient records, medication delivery logs, and clinical notes. Results: Of the 62 patients started on ETI (32 male, mean age 13.3 years), most (76 %) were homozygous for the F508del mutation. Three discontinuations occurred: one pregnancy, two related to side effects. Adherence was high (Proportion of Days covered >90 % both years). Following ETI initiation there was a significant increase in mean FEV1% (+11.7 units; 95 % CI 7.4-15.6), sustained throughout the two-year treatment period. There was no association between baseline lung function and the degree of improvement or rate of decline post-treatment. Improvements were similar for all treatable genotypes. BMI z-score increased by 0.25 units after four months of treatment, returning to baseline by 24 months. Intravenous antibiotic use decreased by 88 % (median IV days/ year reduced from 32 to 4 days, p < 0.01). Conclusions: ETI use in adolescents in a real-world setting led to sustained improvements in health outcomes, consistent with those seen in open trial extension studies.

OBJECTIVES:This study extends methods to estimate average causal effect of aneurysm repair surgery on (i) overall survival and (ii) aneurysm-related mortality, accounting for competing risks using data from the Effective Treatment for Thoracic Aortic Aneurysm (ETTAA) cohort. STUDY DESIGN AND SETTING:ETTAA, a prospective cohort study, recruited 886 patients between 2014 and 2018. Patients were linked to UK national hospital and mortality databases by National Health Service digital and followed-up for later surgeries and deaths. We compared a strategy of open or endovascular surgery (whichever appropriate) within 12 months of enrollment to ETTAA with no surgery within 12 months using the trial emulation framework and cloning-censoring-weighting (CCW) analysis. Key confounders at baseline were controlled for using inverse probability weighting methods. RESULTS:In complete case analysis, if everyone received surgery within a 12-month grace period, an estimated 7-year survival probability was 57.4% (95% CI: 47.3%, 67.4%) vs 49.9% (44.0%, 55.0%) if no one received surgery. This benefit was primarily attributable to reduction in aneurysm-related deaths (difference -8.7%, 95% CI: -14.0%, -3.9%), with no significant effect on deaths from other causes. The findings were consistent under sensitivity analyses, including multiple imputation of missing confounders. Our CCW approach addressed selection-for-treatment, allowed for surgery to be received within a grace period, and used appropriate methods to separate aneurysm-related mortality from competing risks. CONCLUSION:The study demonstrates the utility of trial emulation and counterfactual methods in estimation of causal effects on competing risks using observational data. The findings suggest a benefit for aneurysm-related survival up to 7 years after enrollment. PLAIN LANGUAGE SUMMARY:This study shows how to estimate effects of surgery on different causes of death, when we cannot do a clinical trial, and illustrates this using an example from heart surgery. The aorta is the main artery that carries oxygen-rich blood from the heart to the body. In some people, a part of the vessel wall becomes weak and loses its elastic properties, so it doesn't return to its normal shape after the blood has passed through. This can lead to swelling or bulging in the aorta, called an aneurysm. A thoracic aortic aneurysm, or TAA for short, is an aneurysm in the section of the aorta in the chest (https://www.bhf.org.uk/informationsupport/conditions/thoracic-aortic-aneurysm). We have used data from the Effective Treatment for Thoracic Aortic Aneurysm (ETTAA) study, which investigated aneurysm growth rates, patient outcomes, quality of life, and costs, in 886 patients diagnosed with TAA. ETTAA compared two surgical treatments, Open Heart Surgery, where the section of the aorta that contains the aneurysm is removed and replaced by a new aorta made from a synthetic material, and Stent Grafting, where tubes are inserted into arteries to allow blood to flow freely using less invasive "keyhole" surgery. ETTAA reviewed existing research evidence but data comparing the effectiveness of these two approaches to each other and to outcomes without surgery were of sparse or limited quality and outdated. The results of ETTAA up to 2020 have been published in a monograph. (https://pubmed.ncbi.nlm.nih.gov/35094747/). Two findings from ETTAA motivated this study. First, there were no clinical trials comparing surgery with no surgery and no studies that mimic clinical trials. Second, we had not considered whether surgery overall prevents deaths due to aneurysm or deaths from other causes. We call these two types of death, competing risks. It is unlikely that a clinical trial comparing surgery with no surgery will ever be completed because the number of people who are diagnosed with TAA is small. Also, TAA can become a serious problem if left untreated. On the other hand, surgery for TAA is difficult and can result in serious complications, including death. Therefore, it is important to know how much surgery improves survival related to the aneurysm and whether it improves survival overall. Recent developments in statistics provided methods for investigating survival in a way which increases confidence in the cause-effect relationship between surgery and outcomes. In this study, we show how these statistical methods can be used to estimate the proportion of patients who die from the competing risks, if all patients had surgery within 12 months compared with if no patients had surgery within 12 months. We take into account the different times between diagnosis of TAA and surgery and adjust for the main differences between surgery and no surgery patients. Using these methods, we estimate that surgery reduces deaths due to aneurysms at 7 years by 8.7%, with no effect on deaths from other causes. The benefit of surgery was significant by 3 years after diagnosis. We also provide discussion about using routine medical records to repeat this type of study.

Clinicians increasingly rely on prediction models to guide treatment choices. Most prediction models, however, are developed using observational data that include some patients who have already received the treatment the prediction model is meant to inform. Special attention to the causal role of those earlier treatments is required when interpreting the resulting predictions. "Causal blind spots" were identified in 3 common approaches to handling treatment when developing a prediction model: including treatment as a predictor, restricting to persons taking a certain treatment, and ignoring treatment. Through several real examples, this article illustrates how the risks obtained from models developed using such approaches may be misinterpreted and can lead to misinformed decision making. The discussion covers issues attributable to confounding, selection, mediation, and changes in treatment protocols over time. An extension of guidelines for the development, reporting, and evaluation of prediction models is advocated to avoid such misinterpretations. Developers must ensure that the intended target population for the model, and the treatment conditions under which predictions hold, are clearly communicated. When prediction models are intended to inform treatment decisions, they need to provide estimates of risk under the specific treatment (or intervention) options being considered, known as "prediction under interventions." Next to suitable data, this requires causal reasoning and causal inference techniques during model development and evaluation. Being clear about what a given prediction model can and cannot be used for prevents misinformed treatment decisions and thereby prevents potential harm to patients.

Introduction Target trial emulation is a framework for evaluating the effects of treatments using observational data. The trial emulation approach involves specifying key elements of a protocol for a target trial (a randomised controlled trial designed to address the question of interest) and then describing how best to emulate the trial using observational data. Recent years have seen an uptake of target trial emulation in several disease areas, although there are limited examples in cystic fibrosis (CF). This protocol describes a study which aims to assess the applicability of target trial emulation in CF. We aim to emulate an existing trial in CF and assess to what extent the results from the trial can be replicated using registry data.Methods and analysis The target trial is a published randomised controlled trial which found evidence for beneficial effects of azithromycin use on lung function in young adults with CF. Two emulated trials are planned: one using data from the UK CF Registry and one using data from the US CF Registry. The inclusion and exclusion criteria, treatment and outcome definitions, follow-up period, and estimand of interest are all designed to match the published trial as closely as possible. The analysis step of the trial emulations will use causal inference methods to control for confounding. Results obtained in the emulated trials using registry data will be compared with those from the target trial.Ethics and dissemination Ethical approval has been granted by the London School of Hygiene and Tropical Medicine Ethics Committee (Ref: 29609). This study has also been approved by the UK CF Registry Research Committee and the North Star Review Board. The results of this study will be published in a peer-reviewed journal and presented at relevant scientific conferences.

The DECOVID database contains harmonized pseudonymized electronic health record (EHR) data on all adult (>= 18 years old) patients presenting to two large, digitally mature centers in the United Kingdom between 1 January 2020 and 28 February 2021, with follow-up until at least 28 March 2021. The database was originally developed to support the COVID-19 response but is now available via the PIONEER data hub for researchers to explore a wide range of research questions, including exploratory analyses, risk factor assessment, prediction modeling, and comparative effectiveness studies. Raw data were extracted from local EHRs and transformed into a standardized form (Observational Health Data Sciences and Informatics-Common Data Model version 5.3.1). The database includes 165,420 patients across 256,804 hospital presentations. For these patients, highly granular data are available, including patient demographics, longitudinal vital signs, physiology, treatments, laboratory findings, clinical diagnoses, and outcomes. There are 10,030 patients with COVID-19, of whom 1472 died in hospital.

Xeroderma pigmentosum (XP) is a rare autosomal recessive disorder of DNA repair, characterized by extreme sensitivity to ultraviolet radiation. Patients are at risk of skin cancers and ocular surface disease, and approximately one-third of patients experience progressive neurodegeneration. Life expectancy varies significantly based on complementation group, disease severity and access to preventive measures. Obtaining reliable survival estimates for this rare and genetically heterogeneous disease can be challenging, leading to variability in reported survival rates. A National Institutes of Health study of 106 patients (1971–2009) reported a median age at death of 29 years for patients with XP with neurodegeneration, and 37 years for those without. These statistics are frequently cited in academic literature and widely reported on publicly accessible websites. Our aim was to assess the life expectancy of patients with XP receiving multidisciplinary care at the UK National XP Service and assess patients’ and families’ understanding of their prognosis. We conducted a longitudinal study of 89 patients seen annually in the UK National XP Service from 2015 to 2024. Mortality events were documented and categorized between those with and without neurodegeneration. To evaluate patient perceptions on life expectancy, 24 clinic attendees completed a short survey between February and July 2023. Over a 9-year period, 15 deaths (17%) were recorded, comprising 11 cases with neurodegeneration (groups A, B, D, F and G) and four cases without (groups A, C and V). There were no deaths attributable to skin cancer. Four deaths were linked to internal malignancies. The median survival age was calculated as 50 years (95% confidence interval 34 to not calculable) for those with neurological involvement, and 81 years (95% confidence interval 81 to not calculable) for those without. The latter is comparable with the general UK population (78.6 years for men, 82.6 years for women; 2020–2022 data). Only 11 of 24 survey respondents (46%) believed that life expectancy in neurologically unaffected patients with XP was normal. One respondent (4%) estimated life expectancy at 20–29 years, eight (31%) at 30–39 years and one (4%) at 50–59 years, and three reported uncertainty. We present evidence of improved survival outcomes in patients with XP in the UK, both with and without neurological involvement. While life expectancy can be limited, particularly in the presence of neurodegeneration, advances in medical management and strict adherence to preventive measures have improved outcomes. It is important that outdated and potentially harmful information online is updated. Discussions with patients and families should convey that regular medical follow-ups, diligent photoprotection and timely intervention for skin cancers can enhance longevity and quality of life for those affected by XP.

Estimating risks or survival probabilities conditional on individual characteristics based on censored time-to-event data is a commonly faced task. This may be for the purpose of developing a prediction model or may be part of a wider estimation procedure, such as in causal inference. A challenge is that it is impossible to know at the outset which of a set of candidate models will provide the best predictions. The super learner is a powerful approach for finding the best model or combination of models ('ensemble') among a pre-specified set of candidate models or 'learners', which can include parametric and machine learning models. Super learners for time-to-event outcomes have been developed, but the literature is technical and a reader may find it challenging to gather together the full details of how these methods work and can be implemented. In this paper we provide a practical tutorial on super learner methods for time-to-event outcomes. An overview of the general steps involved in the super learner is given, followed by details of three specific implementations for time-to-event outcomes. We cover discrete-time and continuous-time versions of the super learner, as described by Polley and van der Laan (2011), Westling et al. (2023) and Munch and Gerds (2024). We compare the properties of the methods and provide information on how they can be implemented in R. The methods are illustrated using an open access data set and R code is provided.

The Fine-Gray model for the subdistribution hazard is commonly used for estimating associations between covariates and competing risks outcomes. When there are missing values in the covariates included in a given model, researchers may wish to multiply impute them. Assuming interest lies in estimating the risk of only one of the competing events, this paper develops a substantive-model-compatible multiple imputation approach that exploits the parallels between the Fine-Gray model and the standard (single-event) Cox model. In the presence of right-censoring, this involves first imputing the potential censoring times for those failing from competing events, and thereafter imputing the missing covariates by leveraging methodology previously developed for the Cox model in the setting without competing risks. In a simulation study, we compared the proposed approach to alternative methods, such as imputing compatibly with cause-specific Cox models. The proposed method performed well (in terms of estimation of both subdistribution log hazard ratios and cumulative incidences) when data were generated assuming proportional subdistribution hazards, and performed satisfactorily when this assumption was not satisfied. The gain in efficiency compared to a complete-case analysis was demonstrated in both the simulation study and in an applied data example on competing outcomes following an allogeneic stem cell transplantation. For individual-specific cumulative incidence estimation, assuming proportionality on the correct scale at the analysis phase appears to be more important than correctly specifying the imputation procedure used to impute the missing covariates.

When estimating heterogeneous treatment effects, missing outcome data can complicate treatment effect estimation, causing certain subgroups of the population to be poorly represented. In this work, we discuss this commonly overlooked problem and consider the impact that missing at random outcome data has on causal machine learning estimators for the conditional average treatment effect (CATE). We propose 2 de-biased machine learning estimators for the CATE, the mDR-learner, and mEP-learner, which address the issue of under-representation by integrating inverse probability of censoring weights into the DR-learner and EP-learner, respectively. We show that under reasonable conditions, these estimators are oracle efficient and illustrate their favorable performance through simulated data settings, comparing them to existing CATE estimators, including comparison to estimators that use common missing data techniques. We present an example of their application using the GBSG2 trial, exploring treatment effect heterogeneity when comparing hormonal therapies to non-hormonal therapies among breast cancer patients post surgery, and offer guidance on the decisions a practitioner must make when implementing these estimators.

OBJECTIVES:If randomized controlled trials can be successfully emulated using real-world data (RWD), confidence in the validity of RWD for estimating treatment effects for questions that have not been assessed in trials increases. We used routinely collected administrative and clinical national linked datasets from England to emulate the PR07 trial for high-risk prostate cancer patients, which compared the effects of radiotherapy added to hormone therapy (RT+HT) within 8 weeks of randomization (the "grace period") and hormone therapy (HT) only on all-cause mortality. We highlight methodological choices required and challenges encountered in emulating this trial. STUDY DESIGN AND SETTING:Patients diagnosed with prostate cancer from 2014 to 2020 were identified from the routine national linked datasets. Diagnosis was taken as the time zero. As few patients initiated radiotherapy within 8 weeks of diagnosis, we considered target trials with grace periods of 4-6 months. Estimands of interest were hazard ratios (HRs) and survival probabilities over 7 years. The cloning-censoring-and-weighting (CCW) approach was used to control for measured confounding and to allow for the grace period. We also used an extension (the "landmark-CCW" approach), in which we consider several time-origins post-diagnosis, enabling us to use a grace period of 8 weeks as in PR07. RESULTS:A total of 2,690 patients were eligible for inclusion in the emulated trial. The CCW analysis using a grace period of 6 months gave an estimated HR of 0.48 (95% confidence interval [CI]: 0.34-0.60) and 7-year survival estimates of 80.7% (95% CI: 74.3-87.0) for the RT+HT strategy and 65.6% (95% CI: 62.8-68.1) for HT only strategy, and corresponding risk difference of 15.1% (95% CI: 11.5-18.9). The corresponding HR from the landmark-CCW approach was 0.58 (95% CI: 0.51-0.65) and with survival estimates of 80.7% (95% CI: 77.7-83.8) for RT+HT strategy and 69.8% (95% CI: 68.2-71.4) for the HT only strategy, and a risk difference of 10.9% (95% CI: 6.3-15.9). CONCLUSION:Our findings from the emulated trial using RWD are broadly consistent with those from PR07, with RT+HT estimated to result in better survival compared to HT only. However, the findings were not replicated exactly, with PR07 reporting an HR of 0.77 (95% CI: 0.61-0.98) over 7 years of follow-up. Differences may be in part due to challenges in defining time zero and allowing for a treatment grace period of the same duration as in PR07. Our study considered ways in which these challenges can be addressed, and our findings affirm the utility of RWD for estimating treatment effects. PLAIN LANGUAGE SUMMARY:Clinical trials are the best way to test whether treatments work, but they are expensive, take years to complete, and focus on narrow research questions. If we can use real-world data (RWD) such as patient health records to mimic these trials, we may be able to answer additional medical questions relevant to patients who are prescribed these medications. This study aimed to see if we could recreate the results of a past clinical trial (PR07) using national health data from England. The PR07 trial looked at two treatments for high-risk prostate cancer: hormone therapy (HT) alone and radiotherapy added to hormone therapy (RT+HT) within 8 weeks of starting the study. This trial was chosen for several reasons. It included high-risk patients who were studied for up to 7 years, meaning that there was enough information to study survival outcomes. Being a major UK-based trial, it was useful for comparing with the data recorded in UK national health data. The trial also allowed some flexibility in when treatment started, which was an interesting factor to examine. Its main goal was to see if adding radiotherapy to hormone therapy provided extra benefits to patients. We wanted to see whether the trial results were replicated using the UK health data. If results were replicated, it would provide confidence in further exploration of questions not covered by the trial. In a trial, randomization time is the point where doctors allocate patients to one of the treatments being assessed, usually 1 new treatment and 1 control treatment. Patients in the study are then followed up from that point. However, defining a starting point in a non-trial setting, using UK national health data, is not as straightforward. In the PR07 trial, patients started RT+HT within 8 weeks of randomization. In the UK datasets, we use diagnosis as a starting point, because it is available for both treatment groups. However, very few patients started RT+HT within 8 weeks of diagnosis. To address this, we allowed for longer periods of 4-6 months from diagnosis for RT+HT initiation. Recent developments provided statistical methods for estimating the cause-effect relationship between treatment received and survival patterns. In this study, we show how these approaches can be used to estimate survival patterns if all patients had RT+HT within 4-6 months from diagnosis compared with if no patients had any RT within 4-6 months. We account for the different treatment initiation times and the main differences between the RT+HT and HT only patients. Using these methods, we estimate that RT+HT has an 11%-15% higher survival rate at 7 years compared to HT only. We discuss similarities and differences between these findings and those in the original PR07 trial.

BACKGROUND:Children with cystic fibrosis (CF) from socioeconomically deprived areas have poorer growth, worse lung function, and shorter life expectancy than their less-deprived peers. While early growth is associated with lung function around age 6, it is unclear whether improving early growth in the most deprived children reduces inequalities in lung function. METHODS:We used data from the UK CF Registry, tracking children born 2000-2010 up to 2016. We extended the interventional disparity effects approach to the setting of a longitudinally measured mediator. Applying this approach, we estimated the association between socioeconomic deprivation (children in the least vs. most deprived population quintile; exposure) and lung function at first measurement (ages 6-8, outcome), and the role of early weight trajectories (ages 0-6) as mediators of this relationship. We adjusted for baseline confounding by sex, birthyear, and genotype and time-varying intermediate confounding by lung infection. RESULTS:The study included 853 children, with 165 children from the least and 172 from the most deprived quintiles. The average lung function difference between the least and most deprived quintiles was 4.5% of predicted forced expiratory volume in 1 second (95% confidence interval: 1.1-7.9). If the distribution of early weight trajectories in the most deprived children matched that in the least deprived children, this difference would reduce to 4% (95% confidence interval: 0.57- 7.4). CONCLUSION:Socioeconomic deprivation has a strong negative association with lung function for children with CF. We estimate that improving early weight trajectories in the most deprived children would only marginally reduce these inequalities.

BACKGROUND:Prenatal or infant wasting malnutrition followed by later overweight is associated with increased risk of chronic diseases, including type 2 diabetes. OBJECTIVES:In a pooled analysis of 6 longitudinal cohorts, we investigated associations between prior malnutrition (PM) early in life or in adulthood and subsequent glycemic status. METHODS:We identified cohorts in Tanzania, Zambia, India, and the Philippines in whom low birth weight or wasting malnutrition in childhood or as adults following human immunodeficiency virus or tuberculosis infection had been measured. Anthropometry, body composition, and glycemic status, determined by hemoglobin A1c (HbA1c), and glucose at 120 min in an oral glucose tolerance test (glucose120), were assessed 3-38 y after PM and in non-PM (NPM) controls. HbA1c and glucose120 were compared between PM and NPM participants by linear regression, controlling for age, sex, and socioeconomic status and, in pooled analyses, for cohort also. RESULTS:In the full cohort of 2251 participants, there was no overall association between PM and diabetes risk. Child participants aged ∼12 y who were hospitalized with PM when <2 y had higher glucose120 compared to NPM (difference 0.50 mmol/L; 95% CI: 0.10, 0.91 mmol/L). In pooled analyses across adult cohorts controlling for cohort, age, sex, and socioeconomic status, PM participants, compared to NPM, may have higher glucose120 (difference 0.33 mmol/L; 95% CI: -0.27, 0.92 mmol/L) if still underweight, and higher HbA1c (difference 0.41%; 95% CI: -0.07%, 0.89%) and glucose120 (difference 0.70 mmol/L; 95% CI: -0.25, 1.66 mmol/L) if currently obese. CONCLUSIONS:Childhood PM is associated with greater adult dysglycemia, whereas adulthood PM may have heterogeneous outcomes dependent on subsequent presence/absence of weight gain. Clinicians and public health managers should be aware of the long-term risk and intervene to promote some weight gain but prevent excess weight gain in people who were previously malnourished.

Background Leptospirosis is an underdiagnosed infectious disease with non-specific clinical presentation that requires laboratory confirmation for diagnosis. The serologic reference standard remains the microscopic agglutination test (MAT) on paired serum samples. However, reported estimates of MAT’s sensitivity vary. We evaluated the accuracy of four index tests, MAT on paired samples as well as alternative standards for leptospirosis diagnosis: MAT on single acute-phase samples, polymerase chain reaction (PCR) with the target gene Lfb1 , and ELISA IgM with Leptospira fainei serovar Hurstbridge as an antigen. Methods We performed a systematic review of studies reporting results of leptospirosis diagnostic tests. We searched eight electronic databases and selected studies that tested human blood samples and compared index tests with blood culture and/or PCR and/or MAT (comparator tests). For MAT selection criteria we defined a threshold for single acute-phase samples according to a national classification of leptospirosis endemicity. We used a Bayesian random-effect meta-analysis to estimate the sensitivity and specificity of MAT in single acute-phase and paired samples separately, and assessed risk of bias using the Quality Assessment of Studies of Diagnostic Accuracy Approach- 2 (QUADAS-2) tool. Results For the MAT accuracy evaluation, 15 studies were included, 11 with single acute-phase serum, and 12 with paired sera. Two included studies used PCR targeting the Lfb1 gene, and one included study used IgM ELISA with Leptospira fainei serovar Hurstbridge as antigen. For MAT in single acute-phase samples, the pooled sensitivity and specificity were 14% (95% credible interval [CrI] 3–38%) and 86% (95% CrI 59–96%), respectively, and the predicted sensitivity and specificity were 14% (95% CrI 0–90%) and 86% (95% CrI 9–100%). Among paired MAT samples, the pooled sensitivity and specificity were 68% (95% CrI 32–92%) and 75% (95% CrI 45–93%) respectively, and the predicted sensitivity and specificity were 69% (95% CrI 2–100%) and 75% (2–100%). Conclusions Based on our analysis, the accuracy of MAT in paired samples was not high, but it remains the reference standard until a more accurate diagnostic test is developed. Future studies that include larger numbers of participants with paired samples will improve the certainty of accuracy estimates.

Prediction models are used amongst others to inform medical decisions on interventions. Typically, individuals with high risks of adverse outcomes are advised to undergo an intervention while those at low risk are advised to refrain from it. Standard prediction models do not always provide risks that are relevant to inform such decisions: e.g., an individual may be estimated to be at low risk because similar individuals in the past received an intervention which lowered their risk. Therefore, prediction models supporting decisions should target risks belonging to defined intervention strategies. Previous works on prediction under interventions assumed that the prediction model was used only at one time point to make an intervention decision. In clinical practice, intervention decisions are rarely made only once: they might be repeated, deferred and re-evaluated. This requires estimated risks under interventions that can be reconsidered at several potential decision moments. In the current work, we highlight key considerations for formulating estimands in sequential prediction under interventions that can inform such intervention decisions. We illustrate these considerations by giving examples of estimands for a case study about choosing between vaginal delivery and cesarean section for women giving birth. Our formalization of prediction tasks in a sequential, causal, and estimand context provides guidance for future studies to ensure that the right question is answered and appropriate causal estimation approaches are chosen to develop sequential prediction models that can inform intervention decisions.