Jennifer Morgan

A toolkit of designs for mixing Discrete Event Simulation and System Dynamics

In recent years there has been significant interest in multimethodology and the mixing of OR/MS methods, including Discrete Event Simulation (DES) with System Dynamics (SD). Several examples of mixing DES and SD are described in the literature but there is no overarching framework which characterises the spectrum of options available to modellers. This paper draws on a sample of published case studies, in conjunction with the theoretical literature on mixing methods, to propose a toolkit of designs for mixing DES and SD which can be implemented as a set of questions which a modeller should ask in order to guide the choice of design and inform the associated project methodology. The impetus for this work was the perceived need to transfer insight from reported practice in order to formalise how the two methods can be and have been mixed.

Designs for the complementary use of system dynamics and discrete-event simulation

Discrete-Event Simulation (DES) and System Dynamics (SD) are popular modeling approaches that have been successfully applied in a wide range of situations for various purposes. The two approaches can be viewed as complementary, and show potential for combination. Examining multimethodology literature allows us to develop a modeling framework that considers possible designs for such a combination. The aim of this paper is to apply, reflect on and develop this framework through an intervention that lends itself to both approaches, and to explore how DES and SD can be combined in practice. Models under development with a radiotherapy center to explore the impact of altering patient treatment regimes in response to the adoption of new, more complex, technology are discussed. The potential to combine DES and SD in a way which is both complementary and synergistic is explored, and this paper reflects on the experience to date with regard to the proposed methodology.

Lessons from mixing OR methods in practice: using DES and SD to explore a radiotherapy treatment planning process

Mixing Operational Research (OR) methods is becoming more commonplace. Discrete-Event Simulation (DES) and System Dynamics (SD) are popular modelling methods previously applied to a range of situations for various purposes, which are starting to be mixed in healthcare. However, the practicalities of mixing DES and SD in practice remain unclear. Radiotherapy treatment is a complex multi-stage process where technology and best practice continue to evolve. This paper describes a project undertaken to explore the treatment planning process using mixed OR methods. It presents insights obtained through mixing OR methods within a real-world project. The model development process, the role of each modelling method and the benefits of undertaking a mixed OR methods project design are described. Lessons for mixing DES and SD, and more generally mixing OR methods, are discussed.

Risk of suicide following an alcohol-related emergency hospital admission: An electronic cohort study of 2.8 million people

Objective Alcohol misuse is a well-known risk factor for suicide however, the relationship between alcohol-related hospital admission and subsequent risk of death from suicide is unknown. We aimed to determine the risk of death from suicide following emergency admission to hospital with an alcohol-related cause. Methods We established an electronic cohort study of all 2,803,457 residents of Wales, UK, aged from 10 to under 100 years on 1 January 2006 with six years’ follow-up. The outcome event was death from suicide defined as intentional self-harm (ICD-10 X60-84) or undetermined intent (Y10-34). The main exposure was an alcohol-related admission defined as a ‘wholly attributable’ ICD-10 alcohol code in the admission record. Admissions were coded for the presence or absence of co-existing psychiatric morbidity. The analysis was by Cox regression with adjustments for confounding variables within the dataset. Results During the study follow-up period, there were 15,546,355 person years at risk with 28,425 alcohol-related emergency admissions and 1562 suicides. 125 suicides followed an admission (144.6 per 100,000 person years), of which 11 (9%) occurred within 4 weeks of discharge. The overall adjusted hazard ratio (HR) for suicide following admission was 26.8 (95% confidence interval (CI) 18.8 to 38.3), in men HR 9.83 (95% CI 7.91 to 12.2) and women HR 28.5 (95% CI 19.9 to 41.0). The risk of suicide remained substantial in subjects without known co-existing psychiatric morbidity: HR men 8.11 (95% CI 6.30 to 10.4) and women HR 24.0 (95% CI 15.5 to 37.3). The analysis was limited by the absence in datasets of potentially important confounding variables and the lack of information on alcohol-related harm and psychiatric morbidity in subjects not admitted to hospital. Conclusion Emergency alcohol-related hospital admission is associated with an increased risk of suicide. Identifying individuals in hospital provides an opportunity for psychosocial assessment and suicide prevention of a targeted at-risk group before their discharge to the community.

Sparse generalised principal component analysis

Abstract In this paper, we develop a sparse method for unsupervised dimension reduction for data from an exponential-family distribution. Our idea extends previous work on Generalised Principal Component Analysis by adding L1 and SCAD penalties to introduce sparsity. We demonstrate the significance and advantages of our method with synthetic and real data examples. We focus on the application to text data which is high-dimensional and non-Gaussian by nature and discuss the potential advantages of our methodology in achieving dimension reduction.

A best practice framework to measure spatial variation in alcohol availability

Alcohol outlet density and alcohol-related harms are an internationally reported phenomenon. There are multiple methods described in the literature to measure alcohol outlet density, but with very little commentary on the geographical underpinnings of the methods. In this paper, we present a framework to help practitioners and researchers choose the most appropriate spatial method of measuring alcohol outlet density. The framework includes components on theoretical geography, statistical implications and practical considerations, with an emphasis on population-level exposure. We describe the CHALICE alcohol outlet density measurement method that was developed to investigate the relationships between alcohol outlet density and population harm. The CHALICE method is compared to four other methods found in the published literature. We demonstrate the impact of methodological choices (e.g. network vs. Euclidean distances) on resulting alcohol outlet density scores. We conclude that wherever possible the best practice approach to modelling alcohol outlet density should be used to facilitate flexibility in subsequent statistical analysis and improve the transparency of the results.

Alcohol outlet density and hospital admissions for alcohol-related injury: an electronic record-linked cohort study

Background Little is known on the longitudinal relationship between alcohol availability and the risk of admission to hospital for an alcohol-related injury. We aimed to quantify this relationship using observational record-linked data in a total adult population cohort in Wales, UK; population 2.5 million aged 16 years and over. Methods Data sources included licensed outlets held by the 22 local authorities in Wales under The Licensing Act 2003 for each quarter between 2006 and 2011. Alcohol outlet density was estimated for Census geography small areas (n = 1896 divided into quintiles of equal counts) based on the mean network walking distances between each household and alcohol outlet within a 10 minute walk. Hospital admissions 2006–2011 from the Patient Episode Database for Wales (PEDW) for wholly alcohol-related conditions were anonymously record-linked to the Welsh Demographic Service age-sex register within the Secure Anonymised Information Linkage Databank. Injury admissions were defined by ICD-10 codes S00-99 and T00-19 linked to an alcohol code. Longitudinal statistical analysis used Cox regression models of hospital admissions as a function of outlet density at baseline, adjusting for confounding variables of age, sex and small area deprivation and settlement type, and censoring for death, migration and other cause admission. Results Alcohol outlet density at baseline was significantly associated with emergency hospital admission; the hazard ratio (HR) of an alcohol-related admission (n = 25,722) for living in the highest compared with the lowest quintile of outlet density was 1.17 (95% CI: 1.11, 1.23). The risk was higher for alcohol-related admissions with injury (n = 4,308; HR 1.27, 95% CI: 1.13, 1.42). Conclusions Higher alcohol outlet availability was associated with more emergency hospital admissions for alcohol-related injury. This suggests that restricting the density of alcohol outlets within walking distance from home may improve population health.