Ruth H. Keogh

Background: Evidence on frailty progression following severe infections is limited. We compared rates of transition to greater frailty or death between adults with and without severe infection in England. Methods: We conducted a matched-cohort study among adults aged ≥65 years (1,452,117: median age 76 years, 45% male) in Clinical Practice Research Datalink Aurum (2006-2019). Adults with severe infection (hospitalised primarily due to infection) were matched on calendar time to individuals without severe infection on age, sex, and primary care practice. The admission date was used as index date and same was assigned to matched unexposed adults. We measured frailty using Electronic Frailty Index, a proportion of 36 health deficits in validated categories (Fit 0-0.12, Mild >0.12-0.24, Moderate >0.24-0.36, Severe >0.36). In a time-varying Markov multistate model, we focused on forward transitions from baseline or intermediate frailty states to higher states or death. For each transition, we used Cox regression to estimate cause-specific transition hazard ratios (HR) with 95% confidence intervals (CIs), comparing adults with and without severe infection. We adjusted for baseline frailty score, age, sex, deprivation, harmful alcohol use, smoking, and primary care infection history 5 years before index date. We estimated state occupancy probabilities, and expected length of stay (ELOS) in each state at year five among adults with and without severe infection. We explored effect modification by infection type. Results: Across all transitions, severe infection was associated with higher adjusted hazards of transitioning to worsening frailty or death, HR, 95% CI: ( fit to : mild[1.56, 1.54-1.58], moderate[2.51, 1.79-3.51], death[4.57, 4.50-4.65]; mild to : moderate[1.52, 1.50-1.53], severe[1.90, 1.43-2.52], death[2.67, 2.64-2.70]; moderate to : severe[1.40, 1.38-1.42], death[1.87, 1.85-1.90]; severe to death[1.48, 1.46-1.50]). Transition hazard ratios were strongest for lower respiratory tract infections, followed by sepsis, urinary tract infections, meningitis/encephalitis, gastroenteritis, and skin and soft tissue infections. At five years, adults with severe infection had higher probabilities of transitioning to greater frailty or death across all transitions and lower ELOS in each frailty state than those without severe infection. Interpretation: Severe infections may accelerate frailty deterioration in older age. Prevention through vaccination, early detection, and prompt management may help mitigate this decline. Research in context: Evidence before this study: We searched PubMed (inception to May 08, 2026), for published articles evaluating the association between infections and frailty progression with multistate modelling including studies that focused on one transition only with no language restrictions. We used the search terms [(infection OR infectious) AND (frailty OR frail) AND optional (progression OR multistate OR multi-state OR transition)]. We identified five longitudinal studies that compared transition hazard rates from a baseline fit state to pre-frailty or frailty during follow-up between individuals with specific infections (SARS-CoV-2, HIV, history of 20 infections based on International Classification of Diseases [ICD] codes and prescribed medications, cytomegalovirus antibody concentration quartiles, and seropositivity of herpes viruses [HSV 1 and 2, varicella-zoster, and Epstein-Barr]) versus those without infection. The studies found that infection or infection history increased the hazards of transition from a fit/pre-frail state to frailty after adjusting for confounders. The hazards of frailty recovery from frail to fit state was higher among individuals with HIV infection than those without. None of the studies accounted for the competing risk of death, and all modelled transitions along a single pathway between fit/pre-frail and frail states. Furthermore, frailty was conceptualised as a binary outcome, whereas validated frailty risk prediction models typically define frailty in progressive categories of deterioration, commonly classified as fit, mild, moderate, and severe. There is limited evidence on the risk of transition from baseline and intermediate frailty states to more advanced states following infection, while accounting for competing risk of death. To identify subgroups for targeted care, it is important to determine whether frailty deterioration after severe infection varies by infection type (sepsis, urinary tract infection, skin and soft tissue infection, meningitis/encephalitis, lower respiratory tract infection, and gastroenteritis), age, sex, social deprivation, diabetes, and dementia. Added value of this study: Our study compared rates of transition to greater frailty or death between older adults with and without severe infection (hospitalisation primarily for infection). We found that severe infection increased the hazards of transition from baseline and intermediate frailty states to more advanced frailty states or death. From each same starting frailty state, outgoing transition hazards increased progressively towards more advanced frailty states or death. The hazards of progression to greater frailty or death also varied by infection type, with the strongest associations observed for LRTI, followed by sepsis, UTI, meningitis/encephalitis, gastroenteritis, and skin and soft tissue infections. Implications of all the available evidence: Our findings underscore the importance of infection prevention in reducing the risk of frailty progression in older age. Additional studies are required to explore other wider life-course influences on frailty, to guide the development of comprehensive preventive strategies.

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.

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. Funding: Wellcome Research in context: Evidence before this study: We searched PubMed (inception to October 27, 2025), for published articles evaluating the association between infections and frailty, with no language restrictions. We used the search terms [(infection OR infectious) AND (frailty OR frail)]. We found fifteen observational studies investigating associations between individual infections (including: HIV, cytomegalovirus, SARS-CoV-2, acute respiratory infection, urinary tract infection, and influenza) and frailty in adults. Frailty measures varied: eight studies used Fried’s phenotype index, six used versions of the cumulative deficit index (i.e., Edmonton Frail Scale, FRAIL-NH Scale, Hospital Frailty Risk Score, Clinical Frailty Score, Veterans Affairs Frailty Index, Vulnerable Elders Survey-13), and one study used the Timed Up and Go Test. Results from identified studies were mixed, with nearly half (7/15) reporting a positive association between the infection studied and frailty, and the remaining eight finding no evidence of association. In cross-sectional analyses, HIV, SARS-CoV-2, cytomegalovirus, and urinary tract infection, were each associated with higher mean frailty scores or frailty prevalence. In longitudinal analysis, hospitalisation for acute respiratory infection was followed by higher mean hospital frailty risk scores two years post-discharge. SARS-CoV-2 infection was associated with early onset (i.e., higher hazard) of frailty over three years follow-up. However, other studies found no association between HIV, SARS-CoV-2, acute respiratory infection and influenza, and frailty prevalence, incidence, or transition between frailty states. These mixed findings may reflect methodological differences between the studies, including variation in frailty measures, and study limitations. Frailty exists along a continuum of vulnerability, and progression after infection may be an important outcome, yet current evidence is scarce. It remains unclear whether severe infections or different types of infection, are associated with faster frailty deterioration. Similarly, it is uncertain whether post-infection frailty risk varies by pathogen (bacterial, viral, parasitic, fungal), infection type (sepsis, urinary tract infection, skin and soft tissue infection, meningitis/encephalitis, lower respiratory tract, gastroenteritis), or by age, sex, social deprivation, and pre-existing comorbidities. Added value of this study: Our study compared frailty progression over a five-year period between adults aged ≥65 years with severe infection (hospitalisation primarily due to infection) versus comparators without severe infection. We found higher baseline frailty at severe infection onset than in matched comparators. We saw evidence of increased frailty progression over time in people following severe infections compared to those without, however, these differences were small. We also saw higher risk of worsening frailty progression in older adults and those with dementia. Further, worsening frailty progression varied by infection type (strongest for lower respiratory tract infections, meningoencephalitis (UK only), urinary tract infections, and sepsis). Implications of all the available evidence: Our findings underscore the importance of both frailty and infection prevention in improving health in older age. Additional studies are required to explore other wider life-course influences on frailty, to guide the development of comprehensive preventive strategies.

Since the COVID-19 pandemic, most general practices in England have adopted online consultation (OC) systems for easier access and demand management. However, there are concerns that easier access could increase demand relating to low-risk minor illness, with consequential changes in prescribing patterns. Using GP practice-level data we investigate effects of adoption of an OC system in English general practice on antibiotic prescribing rates. The target estimand is the average effect of adoption on the outcome in each month post-adoption within groups of practices defined by month of adoption. Estimation is using a doubly robust staggered difference-in-differences approach. We also estimate an effect aggregated across all adopting practices and across time since adoption. Overall, we find evidence that adoption of an OC system results in higher antibiotic prescribing rates, with our primary analysis indicating an overall 5.1% increase. However, using recently proposed methods, we find that our estimates are sensitive to possible violations of the key identification assumption, conditional parallel trends. When allowing for relatively small violations of this assumption, we would no longer conclude that there is statistically significant evidence of an effect and smaller in magnitude effects become more plausible, which may be considered reassuring for health policy.

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.

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.

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.

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.

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.