Glen P. Martin

Pre-implantation Balloon Aortic Valvuloplasty and Clinical Outcomes Following Transcatheter Aortic Valve Implantation: A Propensity Score Analysis of the UK Registry

Background Aortic valve predilation with balloon aortic valvuloplasty (BAV) is recommended before transcatheter aortic valve implantation (TAVI), despite limited data around the requirement of this preprocedural step and the potential risks of embolization. This study aimed to investigate the trends in practice and associations of BAV on short‐term outcomes in the UK TAVI registry. Methods and Results Eleven clinical endpoints were investigated, including 30‐day mortality, myocardial infarction, aortic regurgitation, valve dysfunction, and composite early safety. All endpoints were defined as per the VARC‐2 definitions. Odd ratios of each endpoint were estimated using logistic regression, with data analyzed in balloon‐ and self‐expandable valve subgroups. Propensity scores were calculated using patient demographics and procedural variables, which were included in the models of each endpoint to adjust for measured confounding. Between 2007 and 2014, 5887 patients met the study inclusion criteria, 1421 (24.1%) of whom had no BAV before TAVI valve deployment. We observed heterogeneity in the use of BAV nationally, both temporally and by center experience; rates of BAV in pre‐TAVI workup varied between 30% and 97% across TAVI centers. All endpoints were similar between treatment groups in SAPIEN (Edwards Lifesciences Inc., Irvine, CA) valve patients. After correction for multiple testing, none of the endpoints in CoreValve (Medtronic, Minneapolis, MN) patients were significantly different between patients with or without predilation. Conclusions Performing TAVI without predilation was not associated with adverse short‐term outcomes post procedure, especially when using a balloon‐expandable prosthesis. Randomized trials including different valve types are required to provide conclusive evidence regarding the utility of predilation before‐TAVI.

Inadequacy of existing clinical prediction models for predicting mortality after transcatheter aortic valve implantation

Background The performance of emerging transcatheter aortic valve implantation (TAVI) clinical prediction models (CPMs) in national TAVI cohorts distinct from those where they have been derived is unknown. This study aimed to investigate the performance of the German Aortic Valve, FRANCE‐2, OBSERVANT and American College of Cardiology (ACC) TAVI CPMs compared with the performance of historic cardiac CPMs such as the EuroSCORE and STS‐PROM, in a large national TAVI registry. Methods The calibration and discrimination of each CPM were analyzed in 6676 patients from the UK TAVI registry, as a whole cohort and across several subgroups. Strata included gender, diabetes status, access route, and valve type. Furthermore, the amount of agreement in risk classification between each of the considered CPMs was analyzed at an individual patient level. Results The observed 30‐day mortality rate was 5.4%. In the whole cohort, the majority of CPMs over‐estimated the risk of 30‐day mortality, although the mean ACC score (5.2%) approximately matched the observed mortality rate. The areas under ROC curve were between 0.57 for OBSERVANT and 0.64 for ACC. Risk classification agreement was low across all models, with Fleisss kappa values between 0.17 and 0.50. Conclusions Although the FRANCE‐2 and ACC models outperformed all other CPMs, the performance of current TAVI‐CPMs was low when applied to an independent cohort of TAVI patients. Hence, TAVI specific CPMs need to be derived outside populations previously used for model derivation, either by adapting existing CPMs or developing new risk scores in large national registries.

Effect of weekend admission on process of care and clinical outcomes for the management of acute coronary syndromes: a retrospective analysis of three UK centres

Objectives The effect of weekend versus weekday admission following acute coronary syndrome (ACS) on process of care and mortality remains controversial. This study aimed to investigate the ‘weekend-effect’ on outcomes using a multicentre dataset of patients with ST elevation myocardial infarction (STEMI) and non-ST elevation myocardial infarction/unstable angina (NSTEMI/UA). Design This retrospective observational study used propensity score (PS) stratification to adjust estimates of weekend effect for observed confounding. Logistic regression was used to estimate odds ratios (ORs) for binary outcomes and time-to-event endpoints were modelled using Cox proportional hazards to estimate hazard ratios (HRs). Setting Three tertiary cardiac centres in England and Wales that contribute to the Myocardial Ischaemia National Audit Project. Participants Between January 2010 and March 2016, 17 705 admissions met the study inclusion criteria, 4327 of which were at a weekend. Primary and secondary outcomes Associations were studied between weekend admissions and the following primary outcome measures: in-hospital mortality, 30-day mortality and long-term survival; secondary outcomes included several processes of care indicators, such as time to coronary angiography. Results After PS stratification adjustment, mortality outcomes were similar between weekend and weekday admission across patients with STEMI and NSTEMI/UA. Weekend admissions were less likely to be discharged within 1 day (HR 0.72, 95% CI 0.66 to 0.78), but after 4 days the length of stay was similar (HR 0.97, 95% CI 0.90 to 1.04). Fewer patients with NSTEMI/UA received angiography between 0 and 24 hours at a weekend (HR 0.71, 95% CI 0.65 to 0.77). Weekend patients with STEMI were less likely to undergo an angiogram within 1 hour, but there was no significant difference after this time point. Conclusion Patients with ACS had similar mortality and processes of care when admitted on a weekend compared with a weekday. There was evidence of a delay to angiography for patients with NSTEMI/UA admitted at the weekend.

Novel United Kingdom prognostic model for 30-day mortality following transcatheter aortic valve implantation

Objective Existing clinical prediction models (CPM) for short-term mortality after transcatheter aortic valve implantation (TAVI) have limited applicability in the UK due to moderate predictive performance and inconsistent recording practices across registries. The aim of this study was to derive a UK-TAVI CPM to predict 30-day mortality risk for benchmarking purposes. Methods A two-step modelling strategy was undertaken: first, data from the UK-TAVI Registry between 2009 and 2014 were used to develop a multivariable logistic regression CPM using backwards stepwise regression. Second, model-updating techniques were applied using the 2013–2014 data, thereby leveraging new approaches to include frailty and to ensure the model was reflective of contemporary practice. Internal validation was performed by bootstrapping to estimate in-sample optimism-corrected performance. Results Between 2009 and 2014, up to 6339 patients were included across 34 centres in the UK-TAVI Registry (mean age, 81.3; 2927 female (46.2%)). The observed 30-day mortality rate was 5.14%. The final UK-TAVI CPM included 15 risk factors, which included two variables associated with frailty. After correction for in-sample optimism, the model was well calibrated, with a calibration intercept of 0.02 (95% CI −0.17 to 0.20) and calibration slope of 0.79 (95% CI 0.55 to 1.03). The area under the receiver operating characteristic curve, after adjustment for in-sample optimism, was 0.66. Conclusion The UK-TAVI CPM demonstrated strong calibration and moderate discrimination in UK-TAVI patients. This model shows potential for benchmarking, but even the inclusion of frailty did not overcome the need for more wide-ranging data and other outcomes might usefully be explored.

Discharge against medical advice after hospitalisation for acute myocardial infarction

Background Discharge against medical advice (AMA) occurs infrequently but is associated with poor outcomes. There are limited descriptions of discharges AMA in national cohorts of patients with acute myocardial infarction (AMI). This study aims to evaluate discharge AMA in AMI and how it affects readmissions. Methods We conducted a cohort study of patients with AMI in USA in the Nationwide Readmission Database who were admitted between the years 2010 and 2014. Descriptive statistics were presented for variables according to discharge home or AMA. The primary end point was all-cause 30-day unplanned readmissions and their causes. Results 2663 019 patients were admitted with AMI of which 10.3% (n=162 070) of 1569 325 patients had an unplanned readmission within 30 days. The crude rate of discharge AMA remained stable between 2010 and 2014 at 1.5%. Discharge AMA was an independent predictor of unplanned all-cause readmissions (OR 2.27 95% CI 2.14 to 2.40); patients who discharged AMA had >twofold increased crude rate of readmission for AMI (30.4% vs 13.4%) and higher crude rate of admissions for neuropsychiatric reasons (3.2% vs 1.3%). After adjustment, discharge AMA was associated with increased odds of readmissions for AMI (OR 3.65 95% CI 3.31 to 4.03, p<0.001). We estimate that there are 1420 excess cases of AMI among patients who discharged AMA. Conclusions Discharge AMA occurs in 1.5% of the population with AMI and these patients are at higher risk of early readmissions for re-infarction. Interventions should be developed to reduce discharge AMA in high-risk groups and initiate interventions to avoid adverse outcomes and readmission.

Informatics for Health 2017: Advancing both science and practice

Introduction The Informatics for Health congress, 24-26 April 2017, in Manchester, UK, brought together the Medical Informatics Europe (MIE) conference and the Farr Institute International Conference. This special issue of the Journal of Innovation in Health Informatics contains 113 presentation abstracts and 149 poster abstracts from the congress. Discussion The twin programmes of “Big Data” and “Digital Health” are not always joined up by coherent policy and investment priorities. Substantial global investment in health IT and data science has led to sound progress but highly variable outcomes. Society needs an approach that brings together the science and the practice of health informatics. The goal is multi-level Learning Health Systems that consume and intelligently act upon both patient data and organizational intervention outcomes. Conclusions Informatics for Health 2017 demonstrated the art of the possible, seen in the breadth and depth of our contributions. We call upon policy makers, research funders and programme leaders to learn from this joined-up approach.

Clinical prediction in defined populations: a simulation study investigating when and how to aggregate existing models

BackgroundClinical prediction models (CPMs) are increasingly deployed to support healthcare decisions but they are derived inconsistently, in part due to limited data. An emerging alternative is to aggregate existing CPMs developed for similar settings and outcomes. This simulation study aimed to investigate the impact of between-population-heterogeneity and sample size on aggregating existing CPMs in a defined population, compared with developing a model de novo.MethodsSimulations were designed to mimic a scenario in which multiple CPMs for a binary outcome had been derived in distinct, heterogeneous populations, with potentially different predictors available in each. We then generated a new ‘local’ population and compared the performance of CPMs developed for this population by aggregation, using stacked regression, principal component analysis or partial least squares, with redevelopment from scratch using backwards selection and penalised regression.ResultsWhile redevelopment approaches resulted in models that were miscalibrated for local datasets of less than 500 observations, model aggregation methods were well calibrated across all simulation scenarios. When the size of local data was less than 1000 observations and between-population-heterogeneity was small, aggregating existing CPMs gave better discrimination and had the lowest mean square error in the predicted risks compared with deriving a new model. Conversely, given greater than 1000 observations and significant between-population-heterogeneity, then redevelopment outperformed the aggregation approaches. In all other scenarios, both aggregation and de novo derivation resulted in similar predictive performance.ConclusionThis study demonstrates a pragmatic approach to contextualising CPMs to defined populations. When aiming to develop models in defined populations, modellers should consider existing CPMs, with aggregation approaches being a suitable modelling strategy particularly with sparse data on the local population.

A Multiple-Model Generalisation of Updating Clinical Prediction Models

There is growing interest in developing clinical prediction models (CPMs) to aid local healthcare decision‐making. Frequently, these CPMs are developed in isolation across different populations, with repetitive de novo derivation a common modelling strategy. However, this fails to utilise all available information and does not respond to changes in health processes through time and space. Alternatively, model updating techniques have previously been proposed that adjust an existing CPM to suit the new population, but these techniques are restricted to a single model. Therefore, we aimed to develop a generalised method for updating and aggregating multiple CPMs. The proposed “hybrid method” re‐calibrates multiple CPMs using stacked regression while concurrently revising specific covariates using individual participant data (IPD) under a penalised likelihood. The performance of the hybrid method was compared with existing methods in a clinical example of mortality risk prediction after transcatheter aortic valve implantation, and in 2 simulation studies. The simulation studies explored the effect of sample size and between‐population‐heterogeneity on the method, with each representing a situation of having multiple distinct CPMs and 1 set of IPD. When the sample size of the IPD was small, stacked regression and the hybrid method had comparable but highest performance across modelling methods. Conversely, in large IPD samples, development of a new model and the hybrid method gave the highest performance. Hence, the proposed strategy can inform the choice between utilising existing CPMs or developing a model de novo, thereby incorporating IPD, existing research, and prior (clinical) knowledge into the modelling strategy.

Using marginal structural models to adjust for treatment drop-in when developing clinical prediction models

Clinical prediction models (CPMs) can inform decision making about treatment initiation, which requires predicted risks assuming no treatment is given. However, this is challenging since CPMs are usually derived using data sets where patients received treatment, often initiated postbaseline as “treatment drop‐ins.” This study proposes the use of marginal structural models (MSMs) to adjust for treatment drop‐in. We illustrate the use of MSMs in the CPM framework through simulation studies that represent randomized controlled trials and real‐world observational data and the example of statin initiation for cardiovascular disease prevention. The simulations include a binary treatment and a covariate, each recorded at two timepoints and having a prognostic effect on a binary outcome. The bias in predicted risk was examined in a model ignoring treatment, a model fitted on treatment‐naïve patients (at baseline), a model including baseline treatment, and the MSM. In all simulation scenarios, all models except the MSM underestimated the risk of outcome given absence of treatment. These results were supported in the statin initiation example, which showed that ignoring statin initiation postbaseline resulted in models that significantly underestimated the risk of a cardiovascular disease event occurring within 10 years. Consequently, CPMs that do not acknowledge treatment drop‐in can lead to underallocation of treatment. In conclusion, when developing CPMs to predict treatment‐naïve risk, researchers should consider using MSMs to adjust for treatment drop‐in, and also seek to exploit the ability of MSMs to allow estimation of individual treatment effects.

Association of comorbid burden with clinical outcomes after transcatheter aortic valve implantation

Objectives To investigate the association of the CharlsonComorbidity Index (CCI) with clinical outcomes after transcatheter aortic valve implantation (TAVI). Background Patients undergoing TAVI have high comorbid burden; however, there is limited evidence of its impact on clinical outcomes. Methods Data from 1887 patients from the UK, Canada, Spain, Switzerland and Italy were collected between 2007 and 2016. The association of CCI with 30-day mortality, Valve Academic Research Consortium-2 (VARC-2) composite early safety, long-term survival and length of stay (LoS) was calculated using logistic regression and Cox proportional hazard models, as a whole cohort and at a country level, through a two-stage individual participant data (IPD) random effect meta-analysis. Results Most (60%) of patients had a CCI ≥3. A weak correlation was found between the total CCI and four different preoperative risks scores (ρ=0.16 to 0.29), and approximately 50% of patients classed as low risk from four risk prediction models still presented with a CCI ≥3. Per-unit increases in total CCI were not associated with increased odds of 30-day mortality (OR 1.09, 95% CI 0.96 to 1.24) or VARC-2 early safety (OR 1.04, 95% CI 0.96 to 1.14) but were associated with increased hazard of long-term mortality (HR 1.10, 95% CI 1.05 to 1.16). The two-stage IPD meta-analysis indicated that CCI was not associated with LoS (HR 0.97, 95% CI 0.93 to 1.02). Conclusion In this multicentre international study, patients undergoing TAVI had significant comorbid burden. We found a weak correlation between the CCI and well-established preoperative risks scores. The CCI had a moderate association with long-term mortality up to 5 years post-TAVI.