ENBIS-15 in Prague

6 – 10 September 2015; Prague, Czech Republic Abstract submission: 1 February – 3 July 2015

My abstracts


The following abstracts have been accepted for this event:

  • Process Improvement in Healthcare

    Authors: Ronald J.M.M. Does (IBIS UvA BV and University of Amsterdam)
    Primary area of focus / application: Process
    Secondary area of focus / application: Business
    Keywords: Operations management, Process improvement, Healthcare, Efficiency, Lean Six Sigma, DMAIC
    Submitted at 15-Apr-2015 13:37 by Ronald J.M.M. Does
    7-Sep-2015 08:50 Opening Keynote by Ronald Does: Process Improvement in Healthcare
    Since 2000, IBIS UvA has been highly involved in improving processes in healthcare. We observe that a number of incidents that occurred in the last couple of years indicate that patient safety in hospitals is not satisfactory. Also, increasing costs, quality problems and long admission times are regular news items. Some people are of the opinion that hospitals are not in control of their operations and that they can learn a lot from car manufacturers, such as Toyota. Others think that hospitals cannot and should not be compared to a car factory. And then there are the numerous methods for improving the operation, that have been offered to hospitals by consultancy agencies, such as Lean Six Sigma and the Theory of Constraints. In this talk, we explain our views of improving operational effectiveness in healthcare.
    The central mission of a healthcare organization is to deliver good healthcare to patients within financial restrictions posed by society. Good healthcare partly consists of effective methods for diagnosis and treatment. The required knowledge to design and apply these methods comes from medical science. We believe that medicine, including medical statistics, is a rather mature science. But in addition, these diagnostic and treatment methods have to be delivered to patients and this is done by various processes: medical processes, medical support processes, and non-medical support processes.
    As medical science acquires knowledge about diagnostics and treatment, the science of operations management acquires the knowledge to design, control and improve processes. Operational effectiveness qualifies how well processes in an organization perform, and operational excellence expresses the ambition of organizations to do this extremely well. We believe that it is in healthcare operations management, rather than in medical science, that the fields of quality and industrial statistics can make a valuable contribution.
    IBIS UvA has supported the implementation of Lean Six Sigma in many hospitals and we have seen tremendous improvements. More than 600 documented projects show that effective healthcare processes lead to more reliable, faster, flexible and cost-efficient healthcare. We have published these findings in leading international journals on subjects like: improvement of the patient’s clinical path; efficiency improvement of resources; measuring healthcare quality; increasing the efficiency of nursing departments; and classification of healthcare improvement projects. In this talk we discuss these issues.
  • Presenting Statistical Information to Non-Statisticians

    Authors: Joel Smith (Minitab Inc.)
    Primary area of focus / application: Consulting
    Secondary area of focus / application: Education & Thinking
    Keywords: Communication, Presentation, Graphical analysis, Management
    Submitted at 15-Apr-2015 14:20 by Joel Smith
    Accepted (view paper)
    7-Sep-2015 11:30 Presenting Statistical Information to Non-Statisticians
    The successful application of statistics not only requires good data collection and analysis, but also communication of those results to those who must make decisions based on the results of that analysis. Too often statistical practitioners think they are clearly demonstrating their results but then soon find themselves trying to teach statistical concepts to their audience rather than focusing on what was learned in the analysis. There are several common traps that the presenter falls into that diverts the audience's attention from the overarching conclusions of the talk to specific numbers being shown to them, but guidelines will be presented to avoid these traps. Likewise, there are visualizations that non-statisticians will reliably understand much more easily than standard output tables. This session will help attendees learn to clearly and concisely communicate results by taking advantage of these natural tendencies while avoiding the common traps that can sabotage their delivery.
  • Control Charts Based on the Exponential Distribution: Adapting Runs Rules for the t-Chart

    Authors: Joel Smith (Minitab Inc.)
    Primary area of focus / application: Process
    Keywords: t-Chart, Exponential distribution, Process control, Poisson process, Rare events
    Submitted at 15-Apr-2015 14:37 by Joel Smith
    Accepted (view paper)
    7-Sep-2015 10:00 Control Charts Based on the Exponential Distribution: Adapting Runs Rules for the t-Chart
    In this session a probabilistic-based method to construct a t-chart to monitor the stability of certain processes is presented. Assuming that the time between events can be modeled with an exponential distribution (a Poisson process), it will be demonstrated how to apply the supplementary runs rules to identify whether the process is out of control. The exact average run length (ARL) until a signal when the process is stable and under control is calculated, as well as the ability of this chart to identify a shift in the process; both are compared to respective behaviors of other charts currently used to monitor the same type of processes. These charts are applied in health care to monitor the rate of infections and other adverse events as well as in various other applications such as workplace accidents and injuries. However, some existing methods provide undesirable behavior when attempting to detect shifts and may hide or incorrectly demonstrate the nature of such shifts and the method shown provides much more desirable behavior. This method was originally published in the journal Quality Engineering by the submitting author and Dr. Eduardo Santiago.
  • Can We Predict a Heart Attack (or a Design Fault)? The Levy Generator Process – A Discussion

    Authors: Chris McCollin (Nottingham Trent University), Rainer Göb (University of Würzburg)
    Primary area of focus / application: Modelling
    Secondary area of focus / application: Reliability
    Keywords: Levy process, RLC circuit, Parsum residuals, Resonance
    Submitted at 17-Apr-2015 11:02 by Chris McCollin
    8-Sep-2015 10:10 Can We Predict a Heart Attack (or a Design Fault)? The Levy Generator Process – A Discussion
    By analysing various repairable systems failure times by assuming a Levy process, the cusum residuals are plotted and a standard model is hypothesised. The parameters of the model are compared with those of a standard RLC circuit to determine what they might mean in a system failure context. The issue of relevant data is raised and considered. Various other data sets are analysed to determine residual structures and where the standard model differs and alternative models are considered. Common residual structures are discussed: critical events at energy extremes, change points at zero energy and zero energy at resonance. The paper discusses further work to be considered.
  • From a MOOC for Students to a MOOC for Life-Long Training

    Authors: Avner Bar-Hen (University Paris Descartes)
    Primary area of focus / application: Education & Thinking
    Keywords: Mooc, Statistics, Typology of students, Life-long training
    Submitted at 19-Apr-2015 17:05 by Avner Bar-Hen
    The aim of this talk is to analyze the typology of students, especially in terms of age and diploma, of a five-weeks Mooc entitled "Fundamentals of statistics". This Mooc was proposed in January 2014 on the France Université Numérique platform. We show that university students are a minority of the participants and that most of them are registering for lifelong training.

    We stress some of the difficulties to fulfill the hope of Mooc in terms of lifelong training such as motivation, synchronicity. Finally we propose some practical issues adapted for a Mooc dedicated to statistics. Some of these proposals were implemented in a second edition of this Mooc in fundamentals of statistics.
  • Flexible Post-Hoc Predictions in R

    Authors: Russell Lenth (University of Iowa)
    Primary area of focus / application: Other: Invited US session
    Secondary area of focus / application: Modelling
    Keywords: Least-squares means, Post-hoc analysis, Linear models, Mixed models, Bayesian analysis
    Submitted at 21-Apr-2015 18:38 by Russell Lenth
    Accepted (view paper)
    7-Sep-2015 12:00 Flexible Post-Hoc Predictions in R
    Once a statistical model is fitted to experimental or observational data, it is important to communicate its implications effectively. Frequently, this is done by showing what the model predicts. We define a "reference grid" as a regular grid of predictor values -- the levels of factors along with one or more representative values of each covariate. We make predictions on this grid; and averages of these predictions over zero or more predictors are called "least-squares means." The talk discusses the lsmeans package for R, which provides an easy way of obtaining quantitative and graphical summaries of least-squares means (and contrasts thereof) for a large variety of linear and mixed models fitted in R. Emphasis is on specifying the desired results in complex models, and newer additions to the package such as support for MCMC simulations.