Publications by authors named "Matthew Sperrin"

114 Publications

Invited Commentary: Treatment drop-in: making the case for causal prediction.

Am J Epidemiol 2021 Feb 17. Epub 2021 Feb 17.

Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, The Netherlands.

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. The linked article by Xu et al (Am J Epidemiol. XXXX;XXX(XX):XXXX-XXXX) uses 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 illustrates the value of utilising 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 generalisability. Enriching CPMs with causal inference therefore has the potential to add considerable value to the role of prediction in healthcare.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1093/aje/kwab030DOI Listing
February 2021

Estimating the causal effect of BMI on mortality risk in people with heart disease, diabetes and cancer using Mendelian randomization.

Int J Cardiol 2021 Feb 14. Epub 2021 Feb 14.

Division of Endocrinology, Diabetes & Gastroenterology, School of Medical Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Oxford Rd, Manchester M13 9PL, UK; Diabetes, Endocrinology and Metabolism Centre, Peter Mount Building, Manchester Royal Infirmary, Manchester University NHS Foundation Trust, Manchester Academic Health Science Centre, Manchester M13 0HY, UK.

Background: Observational data have reported that being overweight or obese, compared to being normal weight, is associated with a lower risk for death - the "obesity paradox". We used Mendelian randomization (MR) to estimate causal effects of body mass index (BMI) on mortality risks in people with coronary heart disease (CHD), type 2 diabetes mellitus (T2DM) or malignancy in whom this paradox has been often reported.

Methods: We studied 457,746 White British UK Biobank participants including three subgroups with T2DM (n = 19,737), CHD (n = 21,925) or cancer (n = 42,612) at baseline and used multivariable-adjusted Cox models and MR approaches to describe relationships between BMI and mortality risk.

Results: Observational Cox models showed J-shaped relationships between BMI and mortality risk including within disease subgroups in which the BMI values associated with minimum mortality risk were within overweight/obese ranges (26.5-32.5 kg/m). In all participants, MR analyses showed a positive linear causal effect of BMI on mortality risk (HR for mortality per unit higher BMI: 1.05; 95% CI: 1.03-1.08), also evident in people with CHD (HR: 1.08; 95% CI: 1.01-1.14). Point estimates for hazard ratios across all BMI values in participants with T2DM and cancer were consistent with overall positive linear effects but confidence intervals included the null.

Conclusion: These data support the idea that population efforts to promote intentional weight loss towards the normal BMI range would reduce, not enhance, mortality risk in the general population including, importantly, individuals with CHD.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijcard.2021.02.027DOI Listing
February 2021

A scoping review of causal methods enabling predictions under hypothetical interventions.

Diagn Progn Res 2021 Feb 4;5(1). Epub 2021 Feb 4.

Division of Informatics, Imaging and Data Science, Faculty of Biology, Medicine and Health, University of Manchester, Manchester Academic Health Science Centre, Manchester, UK.

Background: The methods with which prediction models are usually developed mean that neither the parameters nor the predictions should be interpreted causally. For many applications, this is perfectly acceptable. However, when prediction models are used to support decision making, there is often a need for predicting outcomes under hypothetical interventions.

Aims: We aimed to identify published methods for developing and validating prediction models that enable risk estimation of outcomes under hypothetical interventions, utilizing causal inference. We aimed to identify the main methodological approaches, their underlying assumptions, targeted estimands, and potential pitfalls and challenges with using the method. Finally, we aimed to highlight unresolved methodological challenges.

Methods: We systematically reviewed literature published by December 2019, considering papers in the health domain that used causal considerations to enable prediction models to be used for predictions under hypothetical interventions. We included both methodologies proposed in statistical/machine learning literature and methodologies used in applied studies.

Results: We identified 4919 papers through database searches and a further 115 papers through manual searches. Of these, 87 papers were retained for full-text screening, of which 13 were selected for inclusion. We found papers from both the statistical and the machine learning literature. Most of the identified methods for causal inference from observational data were based on marginal structural models and g-estimation.

Conclusions: There exist two broad methodological approaches for allowing prediction under hypothetical intervention into clinical prediction models: (1) enriching prediction models derived from observational studies with estimated causal effects from clinical trials and meta-analyses and (2) estimating prediction models and causal effects directly from observational data. These methods require extending to dynamic treatment regimes, and consideration of multiple interventions to operationalise a clinical decision support system. Techniques for validating 'causal prediction models' are still in their infancy.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s41512-021-00092-9DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7860039PMC
February 2021

Infection-related complications after common infection in association with new antibiotic prescribing in primary care: retrospective cohort study using linked electronic health records.

BMJ Open 2021 Jan 15;11(1):e041218. Epub 2021 Jan 15.

Health e-Research Centre, School of Health Sciences, Faculty of Biology, Medicine and Health, The University of Manchester, Manchester, UK.

Objective: Determine the association of incident antibiotic prescribing levels for common infections with infection-related complications and hospitalisations by comparing high with low prescribing general practitioner practices.

Design Retrospective Cohort Study: Retrospective cohort study.

Data Source: UK primary care records from the Clinical Practice Research Datalink (CPRD GOLD) and SAIL Databank (SAIL) linked with Hospital Episode Statistics (HES) data, including 546 CPRD, 346 CPRD-HES and 338 SAIL-HES practices.

Exposures: Initial general practice visit for one of six common infections and the proportion of antibiotic prescribing in each practice.

Main Outcome Measures: Incidence of infection-related complications (as recorded in general practice) or infection-related hospital admission within 30 days after consultation for a common infection.

Results: A practice with 10.4% higher antibiotic prescribing (the IQR) was associated with a 5.7% lower rate of infection-related hospital admissions (adjusted analysis, 95% CI 3.3% to 8.0%). The association varied by infection with larger associations in hospital admissions with lower respiratory tract infection (16.1%; 95% CI 12.4% to 19.7%) and urinary tract infection (14.7%; 95% CI 7.6% to 21.1%) and smaller association in hospital admissions for upper respiratory tract infection (6.5%; 95% CI 3.5% to 9.5%) The association of antibiotic prescribing levels and hospital admission was largest in patients aged 18-39 years (8.6%; 95% CI 4.0% to 13.0%) and smallest in the elderly aged 75+ years (0.3%; 95% CI -3.4% to 3.9%).

Conclusions: There is an association between lower levels of practice level antibiotic prescribing and higher infection-related hospital admissions. Indiscriminately reducing antibiotic prescribing may lead to harm. Greater focus is needed to optimise antibiotic use by reducing inappropriate antibiotic prescribing and better targeting antibiotics to patients at high risk of infection-related complications.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1136/bmjopen-2020-041218DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7813359PMC
January 2021

Continual updating and monitoring of clinical prediction models: time for dynamic prediction systems?

Diagn Progn Res 2021 Jan 11;5(1). Epub 2021 Jan 11.

Division of Informatics, Imaging and Data Science, Faculty of Biology, Medicine and Health, The University of Manchester, Manchester Academic Health Science Centre, Manchester, UK.

Clinical prediction models (CPMs) have become fundamental for risk stratification across healthcare. The CPM pipeline (development, validation, deployment, and impact assessment) is commonly viewed as a one-time activity, with model updating rarely considered and done in a somewhat ad hoc manner. This fails to address the fact that the performance of a CPM worsens over time as natural changes in populations and care pathways occur. CPMs need constant surveillance to maintain adequate predictive performance. Rather than reactively updating a developed CPM once evidence of deteriorated performance accumulates, it is possible to proactively adapt CPMs whenever new data becomes available. Approaches for validation then need to be changed accordingly, making validation a continuous rather than a discrete effort. As such, "living" (dynamic) CPMs represent a paradigm shift, where the analytical methods dynamically generate updated versions of a model through time; one then needs to validate the system rather than each subsequent model revision.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s41512-020-00090-3DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7797885PMC
January 2021

Penalization and shrinkage methods produced unreliable clinical prediction models especially when sample size was small.

J Clin Epidemiol 2020 Dec 8;132:88-96. Epub 2020 Dec 8.

Nuffield Department of Orthopaedics, Centre for Statistics in Medicine, Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford, UK, OX3 7LD; NIHR Oxford Biomedical Research Centre, John Radcliffe Hospital, Oxford, OX3 9DU, UK.

Objectives: When developing a clinical prediction model, penalization techniques are recommended to address overfitting, as they shrink predictor effect estimates toward the null and reduce mean-square prediction error in new individuals. However, shrinkage and penalty terms ('tuning parameters') are estimated with uncertainty from the development data set. We examined the magnitude of this uncertainty and the subsequent impact on prediction model performance.

Study Design And Setting: This study comprises applied examples and a simulation study of the following methods: uniform shrinkage (estimated via a closed-form solution or bootstrapping), ridge regression, the lasso, and elastic net.

Results: In a particular model development data set, penalization methods can be unreliable because tuning parameters are estimated with large uncertainty. This is of most concern when development data sets have a small effective sample size and the model's Cox-Snell R is low. The problem can lead to considerable miscalibration of model predictions in new individuals.

Conclusion: Penalization methods are not a 'carte blanche'; they do not guarantee a reliable prediction model is developed. They are more unreliable when needed most (i.e., when overfitting may be large). We recommend they are best applied with large effective sample sizes, as identified from recent sample size calculations that aim to minimize the potential for model overfitting and precisely estimate key parameters.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.jclinepi.2020.12.005DOI Listing
December 2020

Three-dimensional (3D) magnetic resonance volume assessment and loco-regional failure in anal cancer: early evaluation case-control study.

BMC Cancer 2020 Nov 30;20(1):1165. Epub 2020 Nov 30.

Division of Molecular & Clinical Cancer Sciences, School of Medical Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Wilmslow Road, Manchester, M20 4BX, UK.

Background: The primary aim was to test the hypothesis that deriving pre-treatment 3D magnetic resonance tumour volume (mrTV) quantification improves performance characteristics for the prediction of loco-regional failure compared with standard maximal tumour diameter (1D) assessment in patients with squamous cell carcinoma of the anus undergoing chemoradiotherapy.

Methods: We performed an early evaluation case-control study at two UK centres (2007-2014) in 39 patients with loco-regional failure (cases), and 41 patients disease-free at 3 years (controls). mrTV was determined using the summation of areas method (Vol). Reproducibility was assessed using intraclass concordance correlation (ICC) and Bland-Altman limits of agreements. We derived receiver operating curves using logistic regression models and expressed accuracy as area under the curve (ROC).

Results: The median time per patient for Vol quantification was 7.00 (inter-quartile range, IQR: 0.57-12.48) minutes. Intra and inter-observer reproducibilities were generally good (ICCs from 0.79 to 0.89) but with wide limits of agreement (intra-observer: - 28 to 31%; inter-observer: - 28 to 46%). Median mrTVs were greater for cases (32.6 IQR: 21.5-53.1 cm) than controls (9.9 IQR: 5.7-18.1 cm, p < 0.0001). The ROC for mrT-size predicting loco-regional failure was 0.74 (95% CI: 0.63-0.85) improving to 0.82 (95% CI: 0.72-0.92) when replaced with mrTV (test for ROC differences, p = 0.024).

Conclusion: Preliminary results suggest that the replacement of mrTV for mrT-size improves prediction of loco-regional failure after chemoradiotherapy for squamous cell carcinoma of the anus. However, mrTV calculation is time consuming and variation in its reproducibility are drawbacks with the current technology.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s12885-020-07613-7DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7706015PMC
November 2020

Informative presence and observation in routine health data: A review of methodology for clinical risk prediction.

J Am Med Inform Assoc 2021 Jan;28(1):155-166

Division of Informatics, Imaging and Data Sciences, School of Health Sciences, University of Manchester, Manchester, United Kingdom.

Objective: Informative presence (IP) is the phenomenon whereby the presence or absence of patient data is potentially informative with respect to their health condition, with informative observation (IO) being the longitudinal equivalent. These phenomena predominantly exist within routinely collected healthcare data, in which data collection is driven by the clinical requirements of patients and clinicians. The extent to which IP and IO are considered when using such data to develop clinical prediction models (CPMs) is unknown, as is the existing methodology aiming at handling these issues. This review aims to synthesize such existing methodology, thereby helping identify an agenda for future methodological work.

Materials And Methods: A systematic literature search was conducted by 2 independent reviewers using prespecified keywords.

Results: Thirty-six articles were included. We categorized the methods presented within as derived predictors (including some representation of the measurement process as a predictor in the model), modeling under IP, and latent structures. Including missing indicators or summary measures as predictors is the most commonly presented approach amongst the included studies (24 of 36 articles).

Discussion: This is the first review to collate the literature in this area under a prediction framework. A considerable body relevant of literature exists, and we present ways in which the described methods could be developed further. Guidance is required for specifying the conditions under which each method should be used to enable applied prediction modelers to use these methods.

Conclusions: A growing recognition of IP and IO exists within the literature, and methodology is increasingly becoming available to leverage these phenomena for prediction purposes. IP and IO should be approached differently in a prediction context than when the primary goal is explanation. The work included in this review has demonstrated theoretical and empirical benefits of incorporating IP and IO, and therefore we recommend that applied health researchers consider incorporating these methods in their work.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1093/jamia/ocaa242DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7810439PMC
January 2021

Consistency of variety of machine learning and statistical models in predicting clinical risks of individual patients: longitudinal cohort study using cardiovascular disease as exemplar.

BMJ 2020 11 4;371:m3919. Epub 2020 Nov 4.

Health e-Research Centre, Health Data Research UK North, School of Health Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Manchester, Manchester M13 9PL, UK.

Objective: To assess the consistency of machine learning and statistical techniques in predicting individual level and population level risks of cardiovascular disease and the effects of censoring on risk predictions.

Design: Longitudinal cohort study from 1 January 1998 to 31 December 2018.

Setting And Participants: 3.6 million patients from the Clinical Practice Research Datalink registered at 391 general practices in England with linked hospital admission and mortality records.

Main Outcome Measures: Model performance including discrimination, calibration, and consistency of individual risk prediction for the same patients among models with comparable model performance. 19 different prediction techniques were applied, including 12 families of machine learning models (grid searched for best models), three Cox proportional hazards models (local fitted, QRISK3, and Framingham), three parametric survival models, and one logistic model.

Results: The various models had similar population level performance (C statistics of about 0.87 and similar calibration). However, the predictions for individual risks of cardiovascular disease varied widely between and within different types of machine learning and statistical models, especially in patients with higher risks. A patient with a risk of 9.5-10.5% predicted by QRISK3 had a risk of 2.9-9.2% in a random forest and 2.4-7.2% in a neural network. The differences in predicted risks between QRISK3 and a neural network ranged between -23.2% and 0.1% (95% range). Models that ignored censoring (that is, assumed censored patients to be event free) substantially underestimated risk of cardiovascular disease. Of the 223 815 patients with a cardiovascular disease risk above 7.5% with QRISK3, 57.8% would be reclassified below 7.5% when using another model.

Conclusions: A variety of models predicted risks for the same patients very differently despite similar model performances. The logistic models and commonly used machine learning models should not be directly applied to the prediction of long term risks without considering censoring. Survival models that consider censoring and that are explainable, such as QRISK3, are preferable. The level of consistency within and between models should be routinely assessed before they are used for clinical decision making.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1136/bmj.m3919DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7610202PMC
November 2020

Clinical prediction models to predict the risk of multiple binary outcomes: a comparison of approaches.

Stat Med 2021 Jan 26;40(2):498-517. Epub 2020 Oct 26.

Centre for Prognosis Research, School of Primary, Community and Social Care, Keele University, Staffordshire, UK.

Clinical prediction models (CPMs) can predict clinically relevant outcomes or events. Typically, prognostic CPMs are derived to predict the risk of a single future outcome. However, there are many medical applications where two or more outcomes are of interest, meaning this should be more widely reflected in CPMs so they can accurately estimate the joint risk of multiple outcomes simultaneously. A potentially naïve approach to multi-outcome risk prediction is to derive a CPM for each outcome separately, then multiply the predicted risks. This approach is only valid if the outcomes are conditionally independent given the covariates, and it fails to exploit the potential relationships between the outcomes. This paper outlines several approaches that could be used to develop CPMs for multiple binary outcomes. We consider four methods, ranging in complexity and conditional independence assumptions: namely, probabilistic classifier chain, multinomial logistic regression, multivariate logistic regression, and a Bayesian probit model. These are compared with methods that rely on conditional independence: separate univariate CPMs and stacked regression. Employing a simulation study and real-world example, we illustrate that CPMs for joint risk prediction of multiple outcomes should only be derived using methods that model the residual correlation between outcomes. In such a situation, our results suggest that probabilistic classification chains, multinomial logistic regression or the Bayesian probit model are all appropriate choices. We call into question the development of CPMs for each outcome in isolation when multiple correlated or structurally related outcomes are of interest and recommend more multivariate approaches to risk prediction.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/sim.8787DOI Listing
January 2021

Prediction models for covid-19 outcomes.

BMJ 2020 10 20;371:m3777. Epub 2020 Oct 20.

Centre for Primary Care and Health Services Research, Division of Population Health, Health Services Research and Primary Care, School of Health Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Manchester, UK.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1136/bmj.m3777DOI Listing
October 2020

Performance of prediction models for COVID-19: the Caudine Forks of the external validation.

Eur Respir J 2020 12 24;56(6). Epub 2020 Dec 24.

Clinical Epidemiology and Medical Statistics Unit, Dept of Medical, Surgical and Experimental Sciences, University of Sassari, Sassari, Italy.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1183/13993003.03728-2020DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7562696PMC
December 2020

Adaptive Symptom Monitoring using Hidden Markov - an Application in Ecological Momentary Assessment.

IEEE J Biomed Health Inform 2020 Oct 15;PP. Epub 2020 Oct 15.

Wearable and mobile technology provides new opportunities to manage health conditions remotely and unobtrusively. For example, healthcare providers can repeatedly sample a person's condition to monitor progression of symptoms and intervene if necessary. There is usually a utility-tolerability trade-off between collecting information at sufficient frequencies and quantities to be useful, and over-burdening the user or the underlying technology, particularly when active input is required from the user. Selecting the next sampling time adaptively using previous responses, so that people are only sampled at high frequency when necessary, can help to manage this trade-off. We present a novel approach to adaptive sampling using clustered continuous-time hidden Markov models. The model predicts, at any given sampling time, the probability of moving to an "alert" state, and the next sample time is scheduled when this probability has exceeded a given threshold. The clusters, each representing a distinct sub-model, allow heterogeneity in states and state transitions. The work is illustrated using longitudinal mentalhealth symptom data in 49 people collected using ClinTouch, a mobile app designed to monitor people with a diagnosis of schizophrenia. Using these data, we show how the adaptive sampling scheme behaves under different model parameters and risk thresholds, and how the average sampling can be substantially reduced whilst maintaining a high sampling frequency during high-risk periods.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/JBHI.2020.3031263DOI Listing
October 2020

Impact of sample size on the stability of risk scores from clinical prediction models: a case study in cardiovascular disease.

Diagn Progn Res 2020 9;4:14. Epub 2020 Sep 9.

Centre for Health Informatics, School of Health Sciences, Faculty of Biology, Medicine and Health, The University of Manchester, Oxford Road, Manchester, M13 9PL UK.

Background: Stability of risk estimates from prediction models may be highly dependent on the sample size of the dataset available for model derivation. In this paper, we evaluate the stability of cardiovascular disease risk scores for individual patients when using different sample sizes for model derivation; such sample sizes include those similar to models recommended in the national guidelines, and those based on recently published sample size formula for prediction models.

Methods: We mimicked the process of sampling patients from a population to develop a risk prediction model by sampling patients from the Clinical Practice Research Datalink. A cardiovascular disease risk prediction model was developed on this sample and used to generate risk scores for an independent cohort of patients. This process was repeated 1000 times, giving a distribution of risks for each patient. = 100,000, 50,000, 10,000, (derived from sample size formula) and (meets 10 events per predictor rule) were considered. The 5-95th percentile range of risks across these models was used to evaluate instability. Patients were grouped by a risk derived from a model developed on the entire population (population-derived risk) to summarise results.

Results: For a sample size of 100,000, the median 5-95th percentile range of risks for patients across the 1000 models was 0.77%, 1.60%, 2.42% and 3.22% for patients with population-derived risks of 4-5%, 9-10%, 14-15% and 19-20% respectively; for = 10,000, it was 2.49%, 5.23%, 7.92% and 10.59%, and for using the formula-derived sample size, it was 6.79%, 14.41%, 21.89% and 29.21%. Restricting this analysis to models with high discrimination, good calibration or small mean absolute prediction error reduced the percentile range, but high levels of instability remained.

Conclusions: Widely used cardiovascular disease risk prediction models suffer from high levels of instability induced by sampling variation. Many models will also suffer from overfitting (a closely linked concept), but at acceptable levels of overfitting, there may still be high levels of instability in individual risk. Stability of risk estimates should be a criterion when determining the minimum sample size to develop models.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s41512-020-00082-3DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7487849PMC
September 2020

Toward a framework for the design, implementation, and reporting of methodology scoping reviews.

J Clin Epidemiol 2020 11 26;127:191-197. Epub 2020 Jul 26.

Division of Informatics, Imaging and Data Science, Faculty of Biology, Medicine and Health, University of Manchester, Manchester Academic Health Science Centre, Manchester, UK; NIHR Greater Manchester Patient Safety Translational Research Centre, University of Manchester, Manchester, UK.

Background And Objective: In view of the growth of published articles, there is an increasing need for studies that summarize scientific research. An increasingly common review is a "methodology scoping review," which provides a summary of existing analytical methods, techniques and software that have been proposed or applied in research articles to address an analytical problem or further an analytical approach. However, guidelines for their design, implementation, and reporting are limited.

Methods: Drawing on the experiences of the authors, which were consolidated through a series of face-to-face workshops, we summarize the challenges inherent in conducting a methodology scoping review and offer suggestions of best practice to promote future guideline development.

Results: We identified three challenges of conducting a methodology scoping review. First, identification of search terms; one cannot usually define the search terms a priori, and the language used for a particular method can vary across the literature. Second, the scope of the review requires careful consideration because new methodology is often not described (in full) within abstracts. Third, many new methods are motivated by a specific clinical question, where the methodology may only be documented in supplementary materials. We formulated several recommendations that build upon existing review guidelines. These recommendations ranged from an iterative approach to defining search terms through to screening and data extraction processes.

Conclusion: Although methodology scoping reviews are an important aspect of research, there is currently a lack of guidelines to standardize their design, implementation, and reporting. We recommend a wider discussion on this topic.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.jclinepi.2020.07.014DOI Listing
November 2020

Multiple imputation with missing indicators as proxies for unmeasured variables: simulation study.

BMC Med Res Methodol 2020 07 8;20(1):185. Epub 2020 Jul 8.

Faculty of Biology, Medicine and Health, Vaughan House, University of Manchester, Manchester, M13 9PL, UK.

Background: Within routinely collected health data, missing data for an individual might provide useful information in itself. This occurs, for example, in the case of electronic health records, where the presence or absence of data is informative. While the naive use of missing indicators to try to exploit such information can introduce bias, its use in conjunction with multiple imputation may unlock the potential value of missingness to reduce bias in causal effect estimation, particularly in missing not at random scenarios and where missingness might be associated with unmeasured confounders.

Methods: We conducted a simulation study to determine when the use of a missing indicator, combined with multiple imputation, would reduce bias for causal effect estimation, under a range of scenarios including unmeasured variables, missing not at random, and missing at random mechanisms. We use directed acyclic graphs and structural models to elucidate a variety of causal structures of interest. We handled missing data using complete case analysis, and multiple imputation with and without missing indicator terms.

Results: We find that multiple imputation combined with a missing indicator gives minimal bias for causal effect estimation in most scenarios. In particular the approach: 1) does not introduce bias in missing (completely) at random scenarios; 2) reduces bias in missing not at random scenarios where the missing mechanism depends on the missing variable itself; and 3) may reduce or increase bias when unmeasured confounding is present.

Conclusion: In the presence of missing data, careful use of missing indicators, combined with multiple imputation, can improve causal effect estimation when missingness is informative, and is not detrimental when missingness is at random.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s12874-020-01068-xDOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7346454PMC
July 2020

Missing data should be handled differently for prediction than for description or causal explanation.

J Clin Epidemiol 2020 09 12;125:183-187. Epub 2020 Jun 12.

Faculty of Biology, Medicine and Health, University of Manchester, Manchester, UK.

Missing data are much studied in epidemiology and statistics. Theoretical development and application of methods for handling missing data have mostly been conducted in the context of prospective research data and with a goal of description or causal explanation. However, it is now common to build predictive models using routinely collected data, where missing patterns may convey important information, and one might take a pragmatic approach to optimizing prediction. Therefore, different methods to handle missing data may be preferred. Furthermore, an underappreciated issue in prediction modeling is that the missing data method used in model development may not match the method used when a model is deployed. This may lead to overoptimistic assessments of model performance. For prediction, particularly with routinely collected data, methods for handling missing data that incorporate information within the missingness pattern should be explored and further developed. Where missing data methods differ between model development and model deployment, the implications of this must be explicitly evaluated. The trade-off between building a prediction model that is causally principled, and building a prediction model that maximizes the use of all available information, should be carefully considered and will depend on the intended use of the model.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.jclinepi.2020.03.028DOI Listing
September 2020

Prediction models for diagnosis and prognosis in Covid-19.

BMJ 2020 04 14;369:m1464. Epub 2020 Apr 14.

Faculty of Biology, Medicine and Health, University of Manchester, Manchester M13 9PL, UK.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1136/bmj.m1464DOI Listing
April 2020

Prediction models for diagnosis and prognosis of covid-19 infection: systematic review and critical appraisal.

BMJ 2020 04 7;369:m1328. Epub 2020 Apr 7.

Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht, Utrecht University, Utrecht, Netherlands

Objective: To review and critically appraise published and preprint reports of prediction models for diagnosing coronavirus disease 2019 (covid-19) in patients with suspected infection, for prognosis of patients with covid-19, and for detecting people in the general population at increased risk of becoming infected with covid-19 or being admitted to hospital with the disease.

Design: Living systematic review and critical appraisal by the COVID-PRECISE (Precise Risk Estimation to optimise covid-19 Care for Infected or Suspected patients in diverse sEttings) group.

Data Sources: PubMed and Embase through Ovid, up to 1 July 2020, supplemented with arXiv, medRxiv, and bioRxiv up to 5 May 2020.

Study Selection: Studies that developed or validated a multivariable covid-19 related prediction model.

Data Extraction: At least two authors independently extracted data using the CHARMS (critical appraisal and data extraction for systematic reviews of prediction modelling studies) checklist; risk of bias was assessed using PROBAST (prediction model risk of bias assessment tool).

Results: 37 421 titles were screened, and 169 studies describing 232 prediction models were included. The review identified seven models for identifying people at risk in the general population; 118 diagnostic models for detecting covid-19 (75 were based on medical imaging, 10 to diagnose disease severity); and 107 prognostic models for predicting mortality risk, progression to severe disease, intensive care unit admission, ventilation, intubation, or length of hospital stay. The most frequent types of predictors included in the covid-19 prediction models are vital signs, age, comorbidities, and image features. Flu-like symptoms are frequently predictive in diagnostic models, while sex, C reactive protein, and lymphocyte counts are frequent prognostic factors. Reported C index estimates from the strongest form of validation available per model ranged from 0.71 to 0.99 in prediction models for the general population, from 0.65 to more than 0.99 in diagnostic models, and from 0.54 to 0.99 in prognostic models. All models were rated at high or unclear risk of bias, mostly because of non-representative selection of control patients, exclusion of patients who had not experienced the event of interest by the end of the study, high risk of model overfitting, and unclear reporting. Many models did not include a description of the target population (n=27, 12%) or care setting (n=75, 32%), and only 11 (5%) were externally validated by a calibration plot. The Jehi diagnostic model and the 4C mortality score were identified as promising models.

Conclusion: Prediction models for covid-19 are quickly entering the academic literature to support medical decision making at a time when they are urgently needed. This review indicates that almost all pubished prediction models are poorly reported, and at high risk of bias such that their reported predictive performance is probably optimistic. However, we have identified two (one diagnostic and one prognostic) promising models that should soon be validated in multiple cohorts, preferably through collaborative efforts and data sharing to also allow an investigation of the stability and heterogeneity in their performance across populations and settings. Details on all reviewed models are publicly available at https://www.covprecise.org/. Methodological guidance as provided in this paper should be followed because unreliable predictions could cause more harm than benefit in guiding clinical decisions. Finally, prediction model authors should adhere to the TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) reporting guideline.

Systematic Review Registration: Protocol https://osf.io/ehc47/, registration https://osf.io/wy245.

Readers' Note: This article is a living systematic review that will be updated to reflect emerging evidence. Updates may occur for up to two years from the date of original publication. This version is update 3 of the original article published on 7 April 2020 (BMJ 2020;369:m1328). Previous updates can be found as data supplements (https://www.bmj.com/content/369/bmj.m1328/related#datasupp). When citing this paper please consider adding the update number and date of access for clarity..
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1136/bmj.m1328DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7222643PMC
April 2020

Young adulthood body mass index, adult weight gain and breast cancer risk: the PROCAS Study (United Kingdom).

Br J Cancer 2020 05 23;122(10):1552-1561. Epub 2020 Mar 23.

Manchester Cancer Research Centre and NIHR Manchester Biomedical Research Centre, Manchester, UK.

Background: We tested the hypothesis that body mass index (BMI) aged 20 years modifies the association of adult weight gain and breast cancer risk.

Methods: We recruited women (aged 47-73 years) into the PROCAS (Predicting Risk Of Cancer At Screening; Manchester, UK: 2009-2013) Study. In 47,042 women, we determined BMI at baseline and (by recall) at age 20 years, and derived weight changes. We estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for new breast cancer using Cox models and explored relationships between BMI aged 20 years, subsequent weight changes and breast cancer risk.

Results: With median follow-up of 5.6 years, 1142 breast cancers (post-menopausal at entry: 829) occurred. Among post-menopausal women at entry, BMI aged 20 years was inversely associated [HR per SD: 0.87 (95% CI: 0.79-0.95)], while absolute weight gain was associated with breast cancer [HR per SD:1.23 (95% CI: 1.14-1.32)]. For post-menopausal women who had a recall BMI aged 20 years <23.4 kg/m (75th percentile), absolute weight gain was associated with breast cancer [HR per SD: 1.31 (95% CIs: 1.21-1.42)], but there were no associations for women with a recall BMI aged 20 years of >23.4 kg/m (P values <0.05).

Conclusions: Adult weight gain increased post-menopausal breast cancer risk only among women who were <23.4 kg/m aged 20 years.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41416-020-0807-9DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7217761PMC
May 2020

Temporal improvements in loco-regional failure and survival in patients with anal cancer treated with chemo-radiotherapy: treatment cohort study (1990-2014).

Br J Cancer 2020 03 14;122(6):749-758. Epub 2020 Jan 14.

Division of Cancer Sciences, School of Medical Sciences, Faculty of Biological, Medicine and Health, University of Manchester, Manchester, UK.

Background: We evaluated oncological changes in patients with squamous cell carcinoma of the anus (SCCA) treated by chemoradiotherapy (CRT) from a large UK institute, to derive estimates of contemporary outcomes.

Methods: We performed a treatment-cohort analysis in 560 patients with non-metastatic SCCA treated with CRT over 25 years. The primary outcomes were 3-year loco-regional failure (LRF), 5-year overall survival (OS), and 5-year cancer-specific survival (CSS). We developed prediction models; and overlaid estimates on published results from historic trials.

Results: Age distributions, proportions by gender and cT stage remained stable over time. The median follow-up was 61 (IQR: 36-79) months. Comparing the first period (1990-1994) with the last period (2010-2014), 3-year LRF declined from 33 to 16% (P < 0.001); 5-year OS increased from 60% to 76% (P = 0.001); and 5-year CCS increased from 62% in to 80% (P = 0.001). For 2020, the models predicted a 3-year LRF of 14.7% (95% CIs: 0-31.3); 5-year OS of 74.7% (95% CIs: 54.6-94.9); and 5-year CSS of 85.7% (95% CIs: 75.3-96.0). Reported oncological outcomes from historic trials generally underestimated contemporary outcomes.

Conclusions: Current and predicted rates for 3-year LRF and 5-year survivals are considerably improved compared with those in historic trials.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41416-019-0689-xDOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7078229PMC
March 2020

Longitudinal trajectories of severe wheeze exacerbations from infancy to school age and their association with early-life risk factors and late asthma outcomes.

Clin Exp Allergy 2020 03 21;50(3):315-324. Epub 2020 Jan 21.

National Heart and Lung Institute, Imperial College of Science, Technology, and Medicine, London, UK.

Introduction: Exacerbation-prone asthma subtype has been reported in studies using data-driven methodologies. However, patterns of severe exacerbations have not been studied.

Objective: To investigate longitudinal trajectories of severe wheeze exacerbations from infancy to school age.

Methods: We applied longitudinal k-means clustering to derive exacerbation trajectories among 887 participants from a population-based birth cohort with severe wheeze exacerbations confirmed in healthcare records. We examined early-life risk factors of the derived trajectories, and their asthma-related outcomes and lung function in adolescence.

Results: 498/887 children (56%) had physician-confirmed wheeze by age 8 years, of whom 160 had at least one severe exacerbation. A two-cluster model provided the optimal solution for severe exacerbation trajectories among these 160 children: "Infrequent exacerbations (IE)" (n = 150, 93.7%) and "Early-onset frequent exacerbations (FE)" (n = 10, 6.3%). Shorter duration of breastfeeding was the strongest early-life risk factor for FE (weeks, median [IQR]: FE, 0 [0-1.75] vs. IE, 6 [0-20], P < .001). Specific airway resistance (sR ) was significantly higher in FE compared with IE trajectory throughout childhood. We then compared children in the two exacerbation trajectories with those who have never wheezed (NW, n = 389) or have wheezed but had no severe exacerbations (WNE, n = 338). At age 8 years, FEV /FVC was significantly lower and FeNO significantly higher among FE children compared with all other groups. By adolescence (age 16), subjects in FE trajectory were significantly more likely to have current asthma (67% FE vs. 30% IE vs. 13% WNE, P < .001) and use inhaled corticosteroids (77% FE vs. 15% IE vs. 18% WNE, P < .001). Lung function was significantly diminished in the FE trajectory (FEV /FVC, mean [95%CI]: 89.9% [89.3-90.5] vs. 88.1% [87.3-88.8] vs. 85.1% [83.4-86.7] vs. 74.7% [61.5-87.8], NW, WNE, IE, FE respectively, P < .001).

Conclusion: We have identified two distinct trajectories of severe exacerbations during childhood with different early-life risk factors and asthma-related outcomes in adolescence.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1111/cea.13553DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7065181PMC
March 2020

Examining the impact of data quality and completeness of electronic health records on predictions of patients' risks of cardiovascular disease.

Int J Med Inform 2020 01 11;133:104033. Epub 2019 Nov 11.

Health e-Research Centre, Farr Institute, School of Health Sciences, Faculty of Biology, Medicine and Health, The University of Manchester, Manchester Academic Health Sciences Centre (MAHSC), Oxford Road, Manchester, M13 9PL, UK; Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht, the Netherlands; Alan Turing Institute, Headquartered at the British Library, London, UK. Electronic address:

Objective: To assess the extent of variation of data quality and completeness of electronic health records and impact on the robustness of risk predictions of incident cardiovascular disease (CVD) using a risk prediction tool that is based on routinely collected data (QRISK3).

Design: Longitudinal cohort study.

Settings: 392 general practices (including 3.6 million patients) linked to hospital admission data.

Methods: Variation in data quality was assessed using Sáez's stability metrics quantifying outlyingness of each practice. Statistical frailty models evaluated whether accuracy of QRISK3 predictions on individual predictions and effects of overall risk factors (linear predictor) varied between practices.

Results: There was substantial heterogeneity between practices in CVD incidence unaccounted for by QRISK3. In the lowest quintile of statistical frailty, a QRISK3 predicted risk of 10 % for female was in a range between 7.1 % and 9.0 % when incorporating practice variability into the statistical frailty models; for the highest quintile, this was 10.9%-16.4%. Data quality (using Saez metrics) and completeness were comparable across different levels of statistical frailty. For example, recording of missing information on ethnicity was 55.7 %, 62.7 %, 57.8 %, 64.8 % and 62.1 % for practices from lowest to highest quintiles of statistical frailty respectively. The effects of risk factors did not vary between practices with little statistical variation of beta coefficients.

Conclusions: The considerable unmeasured heterogeneity in CVD incidence between practices was not explained by variations in data quality or effects of risk factors. QRISK3 risk prediction should be supplemented with clinical judgement and evidence of additional risk factors.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ijmedinf.2019.104033DOI Listing
January 2020

Explicit causal reasoning is needed to prevent prognostic models being victims of their own success.

J Am Med Inform Assoc 2019 12;26(12):1675-1676

Division of Informatics, Imaging and Data Science, Faculty of Biology, Medicine and Health, University of Manchester, Manchester, United Kingdom.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1093/jamia/ocz197DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6857504PMC
December 2019

Interaction between co-morbidities and cancer survival.

Eur J Epidemiol 2019 Nov 24;34(11):1103-1105. Epub 2019 Aug 24.

Health eResearch Centre, School of Health Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Manchester, UK.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1007/s10654-019-00547-wDOI Listing
November 2019

Same-Day Discharge After Elective Percutaneous Coronary Intervention: Insights From the British Cardiovascular Intervention Society.

JACC Cardiovasc Interv 2019 08;12(15):1479-1494

Division of Population Health, Health Services Research & Primary Care, University of Manchester, Manchester, United Kingdom; Keele Cardiovascular Research Group, Institute of Primary Care and Health Sciences, University of Keele and Academic Department of Cardiology, Royal Stoke Hospital, Stoke-on-Trent, United Kingdom. Electronic address:

Objectives: The aim of this study was to evaluate national temporal trends in same-day discharge (SDD) and compare clinical outcomes with those among patients admitted for overnight stay undergoing elective percutaneous coronary intervention (PCI) for stable angina.

Background: Overnight observation has been the standard of care following PCI, with no previous national analyses around changes in practice or clinical outcomes from health care systems in which SDD is the predominant practice for elective PCI.

Methods: Data from 169,623 patients undergoing elective PCI between 2007 and 2014 were obtained from the British Cardiovascular Intervention Society registry. Multiple logistic regressions and the British Cardiovascular Intervention Society risk model were used to study the association between SDD and 30-day mortality.

Results: The rate of SDD increased from 23.5% in 2007 to 57.2% in 2014, with center SDD median prevalence varying from 17% (interquartile range: 6% to 39%) in 2007 to 66% (interquartile range: 45% to 77%) in 2014. The largest independent association with SDD was observed for radial access (odds ratio: 1.69; 95% confidence interval: 1.65 to 1.74; p < 0.001). An increase in 30-day mortality rate over time for the SDD cases was observed, without exceeding the predicted mortality risk. According to the difference-in-differences analysis, observed 30-day mortality temporal changes did not differ between SDD and overnight stay (odds ratio: 1.15; 95% confidence interval: 0.294 to 4.475; p = 0.884).

Conclusions: SDD has become the predominant model of care among elective PCI cases in the United Kingdom, in increasingly complex patients. SDD appears to be safe, with 30-day mortality rates in line with those calculated using the national risk prediction score used for public reporting. Changes toward SDD practice have important economic implications for health care systems worldwide.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.jcin.2019.03.030DOI Listing
August 2019

Chronic obstructive pulmonary disease exacerbation episodes derived from electronic health record data validated using clinical trial data.

Pharmacoepidemiol Drug Saf 2019 10 5;28(10):1369-1376. Epub 2019 Aug 5.

Epidemiology, Global Medical, GlaxoSmithKline plc., Uxbridge, UK.

Purpose: To validate an algorithm for acute exacerbations of chronic obstructive pulmonary disease (AECOPD) episodes derived in an electronic health record (EHR) database, against AECOPD episodes collected in a randomized clinical trial using an electronic case report form (eCRF).

Methods: We analyzed two data sources from the Salford Lung Study in COPD: trial eCRF and the Salford Integrated Record, a linked primary-secondary routine care EHR database of all patients in Salford. For trial participants, AECOPD episodes reported in eCRF were compared with algorithmically derived moderate/severe AECOPD episodes identified in EHR. Episode characteristics (frequency, duration), sensitivity, and positive predictive value (PPV) were calculated. A match between eCRF and EHR episodes was defined as at least 1-day overlap.

Results: In the primary effectiveness analysis population (n = 2269), 3791 EHR episodes (mean [SD] length: 15.1 [3.59] days; range: 14-54) and 4403 moderate/severe AECOPD eCRF episodes (mean length: 13.8 [16.20] days; range: 1-372) were identified. eCRF episodes exceeding 28 days were usually broken up into shorter episodes in the EHR. Sensitivity was 63.6% and PPV 71.1%, where concordance was defined as at least 1-day overlap.

Conclusions: The EHR algorithm performance was acceptable, indicating that EHR-derived AECOPD episodes may provide an efficient, valid method of data collection. Comparing EHR-derived AECOPD episodes with those collected by eCRF resulted in slightly fewer episodes, and eCRF episodes of extreme lengths were poorly captured in EHR. Analysis of routinely collected EHR data may be reasonable when relative, rather than absolute, rates of AECOPD are relevant for stakeholders' decision making.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/pds.4883DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7028141PMC
October 2019

Do population-level risk prediction models that use routinely collected health data reliably predict individual risks?

Sci Rep 2019 08 2;9(1):11222. Epub 2019 Aug 2.

Health e-Research Centre, School of Health Sciences, Faculty of Biology, Medicine and Health, The University of Manchester, Manchester Academic Health Sciences Centre (MAHSC), Oxford Road, Manchester, M13 9PL, UK.

The objective of this study was to assess the reliability of individual risk predictions based on routinely collected data considering the heterogeneity between clinical sites in data and populations. Cardiovascular disease (CVD) risk prediction with QRISK3 was used as exemplar. The study included 3.6 million patients in 392 sites from the Clinical Practice Research Datalink. Cox models with QRISK3 predictors and a frailty (random effect) term for each site were used to incorporate unmeasured site variability. There was considerable variation in data recording between general practices (missingness of body mass index ranged from 18.7% to 60.1%). Incidence rates varied considerably between practices (from 0.4 to 1.3 CVD events per 100 patient-years). Individual CVD risk predictions with the random effect model were inconsistent with the QRISK3 predictions. For patients with QRISK3 predicted risk of 10%, the 95% range of predicted risks were between 7.2% and 13.7% with the random effects model. Random variability only explained a small part of this. The random effects model was equivalent to QRISK3 for discrimination and calibration. Risk prediction models based on routinely collected health data perform well for populations but with great uncertainty for individuals. Clinicians and patients need to understand this uncertainty.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1038/s41598-019-47712-5DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6677736PMC
August 2019

How unmeasured confounding in a competing risks setting can affect treatment effect estimates in observational studies.

BMC Med Res Methodol 2019 07 31;19(1):166. Epub 2019 Jul 31.

University of Manchester, Vaughan House, Portsmouth Street, Manchester, M13 9GB, UK.

Background: Analysis of competing risks is commonly achieved through a cause specific or a subdistribution framework using Cox or Fine & Gray models, respectively. The estimation of treatment effects in observational data is prone to unmeasured confounding which causes bias. There has been limited research into such biases in a competing risks framework.

Methods: We designed simulations to examine bias in the estimated treatment effect under Cox and Fine & Gray models with unmeasured confounding present. We varied the strength of the unmeasured confounding (i.e. the unmeasured variable's effect on the probability of treatment and both outcome events) in different scenarios.

Results: In both the Cox and Fine & Gray models, correlation between the unmeasured confounder and the probability of treatment created biases in the same direction (upward/downward) as the effect of the unmeasured confounder on the event-of-interest. The association between correlation and bias is reversed if the unmeasured confounder affects the competing event. These effects are reversed for the bias on the treatment effect of the competing event and are amplified when there are uneven treatment arms.

Conclusion: The effect of unmeasured confounding on an event-of-interest or a competing event should not be overlooked in observational studies as strong correlations can lead to bias in treatment effect estimates and therefore cause inaccurate results to lead to false conclusions. This is true for cause specific perspective, but moreso for a subdistribution perspective. This can have ramifications if real-world treatment decisions rely on conclusions from these biased results. Graphical visualisation to aid in understanding the systems involved and potential confounders/events leading to sensitivity analyses that assumes unmeasured confounders exists should be performed to assess the robustness of results.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1186/s12874-019-0808-7DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6668192PMC
July 2019