Search

Modelling clinical pathways from Electronic Health Records (EHRs) can optimize resources and improve patient care, but current methods for generating pathway models using clustering have limitations including scalability and fidelity of the clusters. We propose a novel pathway modelling approach using Maximum Likelihood (ML) data clustering on Markov chain representations of clinical pathways. Our method is calibrated to produce clusters with low inter-cluster variability across the pathways. We use machine learning with Stochastic Radial Basis Functions (SRBF) kernels for surrogate optimization to handle non-convexity and propose an incremental optimization method to improve scalability. We also define a methodology based on novel replayability scores to help analysts compare the fidelity of alternative clustering results. Results show that our ML method produces clusters that have higher fidelity in terms of replayability scores than k-means based clustering and in capturing queueing contention, which is important for bottleneck identification in healthcare.

Efficient anomaly detection and diagnosis in multivariate time-series data is of great importance for modern industrial applications. However, building a system that is able to quickly and accurately pinpoint anomalous observations is a challenging problem. This is due to the lack of anomaly labels, high data volatility and the demands of ultra-low inference times in modern applications. Despite the recent developments of deep learning approaches for anomaly detection, only a few of them can address all of these challenges. In this paper, we propose TranAD, a deep transformer network based anomaly detection and diagnosis model which uses attention-based sequence encoders to swiftly perform inference with the knowledge of the broader temporal trends in the data. TranAD uses focus score-based self-conditioning to enable robust multi-modal feature extraction and adversarial training to gain stability. Additionally, model-agnostic meta learning (MAML) allows us to train the model using limited data. Extensive empirical studies on six publicly available datasets demonstrate that TranAD can outperform state-of-the-art baseline methods in detection and diagnosis performance with data and time-efficient training. Specifically, TranAD increases F1 scores by up to 17%, reducing training times by up to 99% compared to the baselines., Comment: Accepted in VLDB 2022

We overview LN, a novel solver introduced in the LINE soft- ware package to analyze layered queueing network (LQN) models. The novelty of the LN solver lies in its capability to analyze LQNs with a user- defined combination of solution paradigms, including discrete-event and stochastic simulation, continuous-time Markov chain analysis (CTMC), normalizing constant evaluation (NC), matrix analytic methods (MAM), mean-field approximations (FLUID), and mean-value analysis (MVA). Being parametric in the solver used for each LQN layer, LN as a whole enables the efficient computation of advanced performance metrics such as marginal and joint state probabilities, response and passage time distributions, and transient measures, leveraging individual strengths of the supported solution paradigms. We discuss in particular recent developments added to NC, the default layer solver of LN, which significantly improve the solution of queueing network models obtained using loose layering

With constrained resources, what, where, and how to cache at the edge is one of the key challenges for edge computing systems. The cached items include not only the application data contents but also the local caching of edge services that handle incoming requests. However, current systems separate the contents and services without considering the latency interplay of caching and queueing. Therefore, in this paper, we propose a novel class of stochastic models that enable the optimization of content caching and service placement decisions jointly. We first explain how to apply layered queueing networks (LQNs) models for edge service placement and show that combining this with genetic algorithms provides higher accuracy in resource allocation than an established baseline. Next, we extend LQNs with caching components to establish a joint modeling method for content caching and service placement (JCSP) and present analytical methods to analyze the resulting model. Finally, we simulate real-world Azure traces to evaluate the JCSP method and find that JCSP achieves up to 35% improvement in response time and 500MB reduction in memory usage than baseline heuristics for edge caching resource allocation.

In standard generative deep learning models, such as autoencoders or GANs, the size of the parameter set is proportional to the complexity of the generated data distribution. A significant challenge is to deploy resource-hungry deep learning models in devices with limited memory to prevent system upgrade costs. To combat this, we propose a novel framework called generative optimization networks (GON) that is similar to GANs, but does not use a generator, significantly reducing its memory footprint. GONs use a single discriminator network and run optimization in the input space to generate new data samples, achieving an effective compromise between training time and memory consumption. GONs are most suited for data generation problems in limited memory settings. Here we illustrate their use for the problem of anomaly detection in memory-constrained edge devices arising from attacks or intrusion events. Specifically, we use a GON to calculate a reconstruction-based anomaly score for input time-series windows. Experiments on a Raspberry-Pi testbed with two existing and a new suite of datasets show that our framework gives up to 32% higher detection F1 scores and 58% lower memory consumption, with only 5% higher training overheads compared to the state-of-the-art., Comment: Accepted in NeurIPS 2021 - Workshop on ML for Systems

School climate is a topic of increasing importance internationally. The current study investigated the established measurement invariance of an eight-factor school climate scale using a multinational sample of secondary students. School climate factor means across 14 international groups were compared and findings on the association between school climate factors and mental health were also investigated. Findings, from this study, illustrate several cross-national similarities regarding the ways in which secondary students perceive school climate and the influence of school climate on student mental health. These findings can support school psychologists’ efforts to identify strategies and supports that improve the school environment in areas that are most consistently related to student experiences, such as school safety and school connectedness. Implications, limitations, and future directions are discussed. Impact and Implications Statement—This study contributes insights regarding cross-national perceptions of school climate and mental health. The findings from the current study inform school personnel efforts to develop and implement school improvement strategies that data shows have a notable impact on student experiences. Such information can be established and applied globally while also recognizing the additional influence of localized norms when implementing strategies or interventions. (PsycInfo Database Record (c) 2021 APA, all rights reserved) © 2021 American Psychological Association

In Italy, chronic pain affects more than a quarter of the population, whereas the average European prevalence is 21%. This high prevalence might be due to the high percentage of Italian people who do not receive treatment, even after the passing of law 38/2010 (the right to access pain management in Italy), which created a regional network for the diagnosis and treatment of noncancer chronic pain. Italian epidemiologic studies on chronic pain are scanty, and this observational, multicenter, cross-sectional study is the first to investigate the clinical characteristics of patients who attended the pain management clinics in the Latium Region, Italy, for the management of their noncancer chronic pain. A total of 1,606 patients (mean age 56.8 years, standard deviation ± 11.4), 67% women, were analyzed. Severe pain was present in 54% of the sample. Women experienced pain and had it in two or more sites more often than men (57% vs. 50%, p = .02; and 55.2% vs. 45.9%, p .001, respectively). Chronic pain was musculoskeletal (45%), mixed (34%), and neuropathic (21%). In more than 60% of the cases, chronic pain was continuous, and in 20% it had lasted for more than 48 months; long-lasting pain was often neuropathic. Low back (33.4%) and lower limbs (28.2%) were the main locations. Severe intensity of pain was statistically significantly associated with female gender (odds ratio [OR] 1.39; 95% confidence interval [CI] 1.06-1.84); with International Classification of Diseases, Ninth Revision, codes for chronic pain syndrome (OR 2.14; 95% CI 1.55-2.95); and with continuous pain (OR 2.02; 95% CI 1.54-2.66). Neuropathic pain and mixed pain were significantly associated with number of sites, and a trend seemed to be present (OR 2.11 and 3.02 for 2 and 3 + sites; 95% CI 1.59-2.79 and 2.00-4.55, respectively).

Integrating a product of linear forms over the unit simplex can be done in polynomial time if the number of variables n is fixed (V. Baldoni et al., 2011). In this note, we highlight that this problem is equivalent to obtaining the normalizing constant of state probabilities for a popular class of Markov processes used in queueing network theory. In light of this equivalence, we survey existing computational algorithms developed in queueing theory that can be used for exact integration. For example, under some regularity conditions, queueing theory algorithms can exactly integrate a product of linear forms of total degree N by solving N systems of linear equations.

Background We performed a qualitative study to investigate the experiences of participants in a multicentre randomized controlled trial on a home-based palliative approach (HPA) for adults with severe multiple sclerosis (MS) and their caregivers. Our aim was to explore the strengths and challenges of the intervention, and circumstances that may have influenced its efficacy. Methods Participants to the qualitative study were the patients, their caregivers, patient referring physicians, and the teams who delivered the HPA intervention. We performed semi-structured one-on-one interviews with 12 patients and 15 informal caregivers chosen using a maximum variation strategy, two focus group meetings with patient referring physicians (4 participants each), and one with the HPA teams (9 participants). Results From data analysis (framework method) 38 sub-categories emerged, which were grouped into 10 categories and 3 themes: 'expectations,' 'met and unmet needs', and 'barriers’. Intervention benefits were improved control of symptoms and reduced sense of isolation of the patient-caregiver dyads. Limitations were: factors related to experimental design (difficulty of dyads in identifying examiner and team roles, additional burden for caregivers); team issues (insufficient team building /supervision, competing priorities); limitations of the intervention itself (insufficient length, lack of rehabilitation input); and external factors (resource limitations, under-responsive services/professionals). The referring physician focus groups provided little experiential data. Conclusions The HPA reduced patient symptoms and sense of isolation in patients and caregivers. The indirect role of the HPA teams, and insufficient length of the intervention were key limitations. The experimental design imposed additional burdens on the dyads. Key barriers were the paucity of available services, the demanding administrative procedures, and lack of networking facilities. These findings suggest that two major requirements are necessary for home palliative care to be effective in this patient population: HPA teams well-connected with MS rehabilitation services, and care delivered over the long-term, with variable intensity. Trial registration Current Controlled Trials ISRCTN73082124 (Registered 19/06/2014).

Background: Evidence on the efficacy of palliative care in persons with severe multiple sclerosis (MS) is scarce. Objective: To assess the efficacy of a home-based palliative approach (HPA) for adults with severe MS and their carers. Methods: Adults with severe MS-carer dyads were assigned (2:1 ratio) to either HPA or usual care (UC). At each center, a multi-professional team delivered the 6-month intervention. A blind examiner assessed dyads at baseline, 3 months, and 6 months. Primary outcome measures were Palliative care Outcome Scale-Symptoms-MS (POS-S-MS) and Schedule for the Evaluation of Individual Quality of Life-Direct Weighting (SEIQoL-DW, not assessed in severely cognitively compromised patients). Results: Of 78 dyads randomized, 76 (50 HPA, 26 UC) were analyzed. Symptom burden (POS-S-MS) significantly reduced in HPA group compared to UC ( p = 0.047). Effect size was 0.20 at 3 months and 0.32 at 6 months, and statistical significance was borderline in per-protocol analysis ( p = 0.062). Changes in SEIQoL-DW index did not differ in the two groups, as changes in secondary patient and carer outcomes. Conclusion: HPA slightly reduced symptoms burden. We found no evidence of HPA efficacy on patient quality of life and on secondary outcomes.

Introduzione: Le migliori evidenze sulla gestione del dolore raccomandano l’uso di programmi multidisciplinari al fine di ottenere significativi risultati terapeutici. Questi possono essere garantiti dai Centri di Terapia del Dolore (CPM), in cui team multidisciplinari sono in grado di fornire attività assistenziali avanzate e specializzate per la valutazione, la diagnosi e il trattamento del dolore cronico benigno. Ad oggi, le informazioni relative alla rete sanitaria e la struttura organizzativa di questi centri in Italia sono esigue. Lo scopo di questo studio è quello di descrivere l’offerta sanitaria e organizzativa dei CPM in una regione italiana (Lazio).Metodi:. Studio descrittivo che ha coinvolto tutti i 37 CPM esistenti nella regione Lazio nel 2011, e che ha focalizzato l’attenzione sui 28 CPM che hanno partecipato all’indagine. Risultati: I CPM sono generalmente situati in policlinici universitarie e in strutture ospedaliere pubbliche o private. La loro struttura organizzativa prevede la presenza di ambulatori, Day Surgery, Day Hospital e posti letto. I CPM sono gestiti da medici anestesisti che, nella maggior parte dei casi, non lavorano in team multidisciplinari. Infatti, il numero di altri operatori sanitari, quali infermieri, psicologi e fisioterapisti è esiguo. I centri forniscono principalmente trattamenti di natura farmacologica, seguiti da terapie complementari (CAM ) e trattamenti interventistici invasivi di diversa complessità. Il tempo di attesa mediano per una visita ambulatoriale è di circa 30 giorni. I centri non sono omogeneamente distribuiti sul territorio regionale, e sono maggiormente concentrati nella città di Roma (56%), seguiti da quelli presenti nelle altre province della regione Lazio (26%) e infine nella provincia di Roma (18%). Chiaramente, Roma è la città offre la più ampia gamma di servizi ambulatoriali per il trattamento del dolore cronico benigno e il più alto numero di prestazioni di consulenze ambulatoriali ai pazienti, con differenze significative tra le tre diverse aree (χ2 = 19,6 p

© Springer International Publishing Switzerland 2016.Model-driven engineering (MDE) has been extended in recent years to account for reliability and performance requirements since the early design stages of an application. While this quality-aware MDE exists for both enterprise and cloud applications, it does not exist yet for Big Data systems. DICE is a novel Horizon2020 project that aims at filling this gap by defining the first quality-driven MDE methodology for Big Data applications. Concrete outputs of the project will include a data-aware UML profile capable of describing Big Data technologies and architecture styles, data-aware quality prediction methods, and continuous delivery tools.

© 2015 IEEE.Model-driven engineering (MDE) often features quality assurance (QA) techniques to help developers creating software that meets reliability, efficiency, and safety requirements. In this paper, we consider the question of how quality-aware MDE should support data-intensive software systems. This is a difficult challenge, since existing models and QA techniques largely ignore properties of data such as volumes, velocities, or data location. Furthermore, QA requires the ability to characterize the behavior of technologies such as Hadoop/MapReduce, NoSQL, and stream-based processing, which are poorly understood from a modeling standpoint. To foster a community response to these challenges, we present the research agenda of DICE, a quality-aware MDE methodology for data-intensive cloud applications. DICE aims at developing a quality engineering tool chain offering simulation, verification, and architectural optimization for Big Data applications. We overview some key challenges involved in developing these tools and the underpinning models.

Introduction Liquid targets using top-of-foil loading concept have been succesfully employed for routine high current production of 18F and 13N at Cyclotope (Houston,TX), over the past ten years1,2. These targets are typically filled with 3.5 ml of water, then pressurized with helium gas at 22 bar and bombarded with 18MeV protons (70–100 µA). Average calculated saturation yield for produc-tion of 18F is ~7.8 GBq/µA (210 mCi/µA) using in-house recycled [18O]-water at approximately 93% enrichment. Reduction of beam power per unit of area is one of the advantages of a tilted entrance-foil geo-metry. Implementation of this target geometry on the ACSI TR19 cyclotron 25degrees upwards irradiation port results in an almost horizontal target entrance foil. A 6ml total cavity volume target allows variable liquid fill volumes of 1.2–4.5 ml for beam current operation from 30–120 µA, resulting in a very efficient use of the costly 18O-water. In a near horizontal installation as in the mayority of cyclotrons, the fill volume flexibility is drastically reduced, having a minimum fill volume of 3.3 ml. At the requirement of Laboratorios Bacon, Cyc-lotope modified the target design with a front mounted collimator compatible with the IBA Cyclone 18/9 cyclotron. A second requirement was to reduce the minimum fill volume for horizontally mounted targets to 2.5 ml or less, while maintaining saturation yield performance. To preserve compatibility with existing IBA targets, the target hardware was modified to operate in self-pressurization mode. This paper presents the results obtained with high and low volume Niobium target inserts (6ml and 4 ml) mounted near horizontally on the IBA Cyclone 18/9 cyclotron and operated in self-pressurization mode. We present pressure/current characteristics, target performance (saturation yield, produced activities, maintenance frequency, FDG yields, etc.). Material and Methods The following targets manufactured by Cyclotope were tested and routinely used for production at Laboratorios Bacon: 1-High Volume Target CY2 model (“American Standard”), 6ml Niobium cavity. 2-Low Volume Target, CY3a model (“Traful”), 4ml Niobium cavity. 3- Low volume Target, CY3b model (“Ferrum”), 4.1ml Niobium cavity. Results and Conclusion The advantages of self-pressurization mode (Laboratorios Bacon setup) are: - Using the vapor pressure as a performance parameter - heat removal by boiling/condensation cycle starts at lower temperature (beam cur-rent) . While, the advantages of the pre-pressurized targets (Cyclotope setup) are: - reduced pressure fluctuations - performance is basically unaffected by plumbing dead volume - flexibility to locate instrumentation farther away from radiation fields - less dependence on fill volume - potential target leaks can be detected before starting an irradiation No significant differences were found in target performance when operated in either pressu-rization mode. The self-pressurizing setup seems to require a sligthly lower fill volume (approxi-mately 5%). The maximum beam current was limited by the foil rupture pressure (~ 40 bar). Safe maximum operating pressure was determined as 30 bar. No foil rupture was experienced during nine months of daily irradiation of these targets in self-pressurizing mode at Laboratorios Bacon. The irradiation parameters and target performance for the different targets are shown in Tables 1 and 2 below. The low volume Traful and Ferrum targets have the best saturation activity vs. fill volume, A(sat)/V, relation. Both targets produce 310 ± 31GBq (8.4 ± 0.8 Ci) of high quali-ty fluoride (F-18) in two hours of irradiation at 70 µA. The low volume targets have a low operation pressure (20bar @ 70µA) when compared to the IBA (NIRTA XL) targets. The typical saturation activity for the low volume targets was 592 ± 59 GBq (16 ± 1.6 Ci) of F-18 at 70 µA, 8.5 GBq/µA (228 mCi/µA) using 2.7ml enriched O-18 water (98 % +). The maintenance interval (> 10 mA.h) is very conveniente to reduce personnel radiation dose. No reduction in FDG yields was observed during that operation interval. In contrast, operation of the high volume targets in pre-presurization mode at the Cyclotope facility results in a higher maximum beam current limit (135 µA) for the same operating pressure (25 bar). Nevertheless, more O-18 water will be required to irradiate at this high current (4.5 ml vs. 3.0 ml). In self-pressurizing mode, a higher filling volume will reduce the expansion volume and, in consequence, the maximum beam current.

26th European Photovoltaic Solar Energy Conference and Exhibition; 676-680, In this paper we report the experimental results obtained from the manufacturing and characterization of III-V solar cells based on single junction, dual junction and triple junction structures. The main solar cell materials are GaAs and (Al)InGaP compounds. A test plan for qualification of “bare” solar cells for concentration systems is presented with its results. The test plan was derived from the space experience and it is based on the European Standards ECSS-E ST20-08, modified keeping into account the terrestrial standard for CPV modules (IEC 62108). The solar cells performances and their reliability is discussed with emphasis to the particular operation conditions induced by the CPV approach. These solar cells were then used and tested in two different CPV approaches: one based on standard concentrators and the other based on spectrum splitting concentrators. The systems, designed for standard concentration, need very high efficiency solar cells, therefore triple junction solar cells InGaP/GaAs/Ge are preferable. HCPV (High Concentration Photovoltaic) systems, based on this approach, have demonstrated efficiencies as high as 29%. The limit of this approach is that the multi junction solar cell structure is optimised for a specified incident spectrum (generally AM1.5D). During the day the spectrum is changing and the efficiency cannot be maintained, especially in the early morning and in the late afternoon. Spectrum splitting systems divide the light in several wavelength regions and use single or dual junction solar cells, optimised on each wavelength region. In this case a very high efficient system can be obtained that is less affected by spectrum variation. The use of standard concentrator and spectrum splitting systems is compared from a theoretical, experimental and economical point of view. The first results of outdoor performances are presented and discussed.

Evidence on pain management highlights the importance of a multidisciplinary approach in order to achieve optimal therapeutic results. Such programs can be guaranteed by the Centers for Pain Management (CPMs), in which multidisciplinary teams are able to provide advanced and specialized activities for the assessment, diagnosis and treatment of chronic benign pain. To date, information related to healthcare supply and the organizational structure of these centers in Italy is incomplete. The aim of this paper was to provide an overview of the healthcare network of the CPMs in the Lazio region.A descriptive survey was conducted in all the 37 CPMs existing in the Lazio region in 2011 of which 28 participated.CPMs were located either in Universities or in public or private hospital facilities. They included a clinic, a Day Hospital service, Day surgery and day-beds. CPMs were managed by anaesthesiologists who, in most instances, did not work in a multidisciplinary team. The number of other health professionals available, such as nurses, psychologists and physiotherapists, was limited. CPMs mainly provided drug therapy, Complementary Alternative Medicine (CAM) and complex interventional treatments. The median waiting time was 30 days. The clinics were not homogeneously distributed in the region with a higher concentration in Rome (56%), followed by other provinces of the Lazio region (26%) and the province of Rome (18%). Clearly, Rome was the city which offered the greatest range of healthcare services and the highest number of consultations with patients, which significantly differed from those of the other areas (χ²=19.6 p0.01).In 2011, the availability of CPMs was not equally distributed throughout the territory, and there was an over-utilization of the facilities in Rome and an under-utilization in the provincial areas. Moreover, this study showed a lack of a multi-professional approach to chronic pain management.

Since 2001, the region of Lombardy has accreditated nursing homes named R.S.A. "Residenza Sanitario-Assistenziale", which make up Italy's principal RSA network with the annual turnover of approximately 60,000 people. The most noteworthy element of the reform introduced is a concentration on resident's frailty rather than disability. This is assessed by using SOSIA, a form for intermediate observation of assistance. Residents are classified into eight classes of frailty, called isofrailty classes of SOSIA, which are differentiated by how compromised their motor and cognitive skills are, and by the presence of comorbidity. The study presents the methodology used to identify and estimate cut-offs of three indicators employed in SOSIA classification. It also discusses their characteristics versus other evaluation systems, such as Resource Utilization Group RUG-III and Autonomie Gerontologique--Groupes Iso-Resources AGGIR.