Search

Objective: To assess the availability and source of guidelines for common infections in European paediatric hospitals and determine their content and characteristics. Design: Participating hospitals completed an online questionnaire on the availability and characteristics of antibiotic prescribing guidelines and on empirical antibiotic treatment including duration of therapy for 5 common infection syndromes: respiratory tract, urinary tract, skin and soft tissue, osteoarticular and sepsis in neonates and children. Results: 84 hospitals from 19 European countries participated in the survey of which 74 confirmed the existence of guidelines. Complete guidelines (existing guidelines for all requested infection syndromes) were reported by 20% of hospitals and the majority (71%) used a range of different sources. Guidelines most commonly available were those for urinary tract infection (UTI) (74%), neonatal sepsis (71%) and sepsis in children (65%). Penicillin and amoxicillin were the antibiotics most commonly recommended for respiratory tract infections (RTIs) (up to 76%), cephalosporin for UTI (up to 50%) and for skin and soft tissue infection (SSTI) and bone infection (20% and 30%, respectively). Antistaphylococcal penicillins were recommended for SSTIs and bone infections in 43% and 36%, respectively. Recommendations for neonatal sepsis included 20 different antibiotic combinations. Duration of therapy guidelines was mostly available for RTI and UTI (82%). A third of hospitals with guidelines for sepsis provided recommendations for length of therapy. Conclusions: Comprehensive antibiotic guideline recommendations are generally lacking from European paediatric hospitals. We documented multiple antibiotics and combinations for most infections. Considerable improvement in the quality of guidelines and their evidence base is required, linking empirical therapy to resistance rates.

In this paper we study a challenging call center operation problem. The goal of our analysis is to identify an optimal policy for allocating tasks to agents. As a first step, we discuss promising randomized policies and use stochastic approximation for finding the optimal randomized policy when implemented via a Bernoulli scheme. As we will show in this paper, the performance of the call center can be improved if the randomized policy is implemented by a deterministic sequence satisfying some regularity conditions. Such sequences are called balanced and we will show that implementing randomized policies by balanced sequences provide an additional step in optimization and control. This motivates our new approach where we avoid the intermediate step of first finding an optimal randomized control and directly find the optimal balanced sequence for control of the call center via stochastic approximation.

A version of the classical secretary problem is studied, in which one is interested in selecting one of the b best out of a group of n differently ranked persons who are presented one by one in a random order. It is assumed that b is bigger than or equal to 1 is a preassigned number. It is known, already for a long time, that for the optimal policy one needs to compute b position thresholds, for instance via backwards induction. In this paper we study approximate policies, that use just a single or a double position threshold, albeit in conjunction with a level rank. We give exact and asymptotic (as n goes to infinity) results, which show that the double-level policy is an extremely accurate approximation.

Welke kansen liggen er om straks infrastructuur met voldoende kwaliteit beschikbaar te hebben? Hoe organiseer je dat? Dit boek is een zoektocht naar mogelijkheden om met infrastructuur waarde te creëren. Door meer professionele ketenintegratie en door bredere samenwerking. Tussen overheden en marktpartijen, maar ook tussen bijvoorbeeld verschillende wegbeheerders. Voor de gebruiker en de belastingbetaler maakt het niet uit wie wat regelt. Als die maar van A naar B kan en weet dat overstromingen worden voorkomen. De rol van een adequaat publiek opdrachtgeverschap is daarbij cruciaal. Dit boek verkent hoe verschillende nutssectoren zich ontwikkelen en trekt daar lessen uit. Door middel van interviews met sleutelpersonen (telecom, logistiek, energie, wonen en infra). Er zijn tien cases, uit binnen- en buitenland, voorbeelden van een slimmere voortbrengingsketen of het verbinden van één vraagstuk met andere maatschappelijke urgenties. Perspectieven voor de Nederlandse Rijkswaterstaat worden geschetst.

Met lit. opg

Neural Networks trained with error back propagation Levenberg-Marquardt training (LM), Neural Networks trained with an algebraic method and Support Vector Machines (SVM) were evaluated to extract the position information from measured light distributions generated by the interactions of 511 keV photons in monolithic scintillator blocks. All three algorithms can achieve a similar average resolution (similar to 1.6 mm FWHM in a 20 x 10 x 10 mm LSO block) but the LM trained neural networks do so most efficiently. When the incidence angle of the photons increases to 30 degrees, the resolution degrades slightly to 2.0 mm FWHM. A small mismatch (

Unique-lowest sealed-bid auctions are auctions in which participation is endogenous and the winning bid is the lowest bid among all unique bids. Such auctions admit very many Nash equilibria (NEs) in pure and mixed strategies. The two-bidders' auction is similar to the Hawk-Dove game, which motivates to study symmetric NEs: Properties and comparative statics are derived and the symmetric NE with the lowest expected gains is the maximin in symmetric strategies, which allows computation through a mathematical program. Simulations provide numerical evidence that the symmetric NE with the lowest expected gains is the unique limit point of the replicator dynamics.