London, United Kingdom
London, United Kingdom
SEARCH FILTERS
Time filter
Source Type

Fihn S.D.,Analytics Intelligence | Fihn S.D.,University of Washington | Francis J.,Analytics Intelligence | Clancy C.,VHA | And 7 more authors.
Health Affairs | Year: 2014

Health care has lagged behind other industries in its use of advanced analytics. The Veterans Health Administration (VHA) has three decades of experience collecting data about the veterans it serves nationwide through locally developed information systems that use a common electronic health record. In 2006 the VHA began to build its Corporate Data Warehouse, a repository for patient-level data aggregated from across the VHA's national health system. This article provides a high-level overview of the VHA's evolution toward "big data," defined as the rapid evolution of applying advanced tools and approaches to large, complex, and rapidly changing data sets. It illustrates how advanced analysis is already supporting the VHA's activities, which range from routine clinical care of individual patients-for example, monitoring medication administration and predicting risk of adverse outcomes-to evaluating a systemwide initiative to bring the principles of the patientcentered medical home to all veterans. The article also shares some of the challenges, concerns, insights, and responses that have emerged along the way, such as the need to smoothly integrate new functions into clinical workflow. While the VHA is unique in many ways, its experience may offer important insights for other health care systems nationwide as they venture into the realm of big data. © 2014 by Project HOPE - The People-to-People Health Foundation.


McCarthy J.F.,Serious Mental Illness Treatment Resource and Evaluation Center | Bossarte R.M.,Federal office of Public Health of Fribourg | Katz I.R.,Office of Mental Health Operations | Thompson C.,Mental Health Services | And 3 more authors.
American Journal of Public Health | Year: 2015

Objectives. The Veterans Health Administration (VHA) evaluated the use of predictive modeling to identify patients at risk for suicide and to supplement ongoing care with risk-stratified interventions. Methods. Suicide data came from the National Death Index. Predictors were measures from VHA clinical records incorporating patient-months from October 1, 2008, to September 30, 2011, for all suicide decedents and 1% of living patients, divided randomly into development and validation samples. We used data on all patients alive on September 30, 2010, to evaluate predictions of suicide risk over 1 year. Results. Modeling demonstrated that suicide rates were 82 and 60 times greater than the rate in the overall sample in the highest 0.01% stratum for calculated risk for the development and validation samples, respectively; 39 and 30 times greater in the highest 0.10%; 14 and 12 times greater in the highest 1.00%; and 6.3 and 5.7 times greater in the highest 5.00%. Conclusions. Predictive modeling can identify high-risk patients who were not identified on clinical grounds. VHA is developing modeling to enhance clinical care and to guide the delivery of preventive interventions.


Littman A.J.,Seattle Epidemiologic Research and Information Center | Littman A.J.,University of Washington | Boyko E.J.,Seattle Epidemiologic Research and Information Center | Boyko E.J.,University of Washington | And 3 more authors.
Preventing Chronic Disease | Year: 2012

Introduction To improve the health of overweight and obese veterans, the Department of Veterans Affairs (VA) developed the MOVE! Weight Management Program for Veterans. The aim of this evaluation was to assess its reach and effectiveness. Methods We extracted data on program involvement, demographics, medical conditions, and outcomes from VA administrative databases in 4 Western states. Eligibility criteria for MOVE! were being younger than 70 years and having a body mass index (BMI, in kg/m 2) of at least 30.0, or 25.0 to 29.9 with an obesity-related condition. To evaluate reach, we estimated the percentage of eligible veterans who participated in the program and their representativeness. To evaluate effectiveness, we estimated changes in weight and BMI using multivariable linear regression. Results Less than 5% of eligible veterans participated, of whom half had only a single encounter. Likelihood of participation was greater in women, those with a higher BMI, and those with more primary care visits, sleep apnea, or a mental health condition. Likelihood of participation was lower among those who were younger than 55 (vs 55-64), widowed, current smokers, and residing farther from the medical center (≥30 vs <30 miles). At 6- and 12-month follow-up, participants lost an average of 1.3 lb (95% confidence interval [CI], -2.6 to -0.02 lb) and 0.9 lb (95% CI, -2.0 to 0.1 lb) more than nonparticipants, after covariate adjustment. More intensive treatment (≥6 encounters) was associated with greater weight loss at 12 months (-3.7 lb; 95% CI, -5.1 to -2.3 lb). Conclusion Few eligible patients participated in the program during the study period, and overall estimates of effectiveness were low.


Yu S.,Analytics Intelligence | Krishnapuram B.,Analytics Intelligence | Rosales R.,Yahoo! | Bharat Rao R.,Analytics Intelligence
Journal of Machine Learning Research | Year: 2011

Co-training (or more generally, co-regularization) has been a popular algorithm for semi-supervised learning in data with two feature representations (or views), but the fundamental assumptions underlying this type of models are still unclear. In this paper we propose a Bayesian undirected graphical model for co-training, or more generally for semi-supervised multi-view learning. This makes explicit the previously unstated assumptions of a large class of co-training type algorithms, and also clarifies the circumstances under which these assumptions fail. Building upon new insights from this model, we propose an improved method for co-training, which is a novel co-training kernel for Gaussian process classifiers. The resulting approach is convex and avoids local-maxima problems, and it can also automatically estimate how much each view should be trusted to accommodate noisy or unreliable views. The Bayesian co-training approach can also elegantly handle data samples with missing views, that is, some of the views are not available for some data points at learning time. This is further extended to an active sensing framework, in which the missing (sample, view) pairs are actively acquired to improve learning performance. The strength of active sensing model is that one actively sensed (sample, view) pair would improve the joint multi-view classification on all the samples. Experiments on toy data and several real world data sets illustrate the benefits of this approach. © 2011 Shipeng Yu, Balaji Krishnapuram, Romer Rosales and R. Bharat Rao.


Roy G.,Analytics Intelligence
Proceedings of the 2013 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2013 | Year: 2013

To provide better and improved patient care, it's important to know the driver of critical clinical processes and focus areas of improvement. For example, readmission to hospital and death after discharge are adverse patient outcomes that are serious, common and costly. Similarly, prolonged 'Length of Stay' increases the possibility of nosocomial infections, introduces patient safety issues, increases the cost per discharge and consumes considerable hospital resources including beds and staff time. Being able to accurately predict the risk of such outcomes would allow health care professionals to focus on appropriate measures to reduce the risk. This paper demonstrates how, with the help of a predictive analytics tool, we can dynamically build models to predict any patient outcome of our choice. This prediction function calculates the probability of the outcome for a given set of patients and those probabilities are used to derive a risk score index for each patient. The result is finally presented in a dynamic business dashboard for further management actions. © 2013 IEEE.


Patent
Analytics Intelligence | Date: 2011-04-11

A computer-implemented method, system and program for interactive data delivering are described. A method for the interactive data delivering provides an effective way for retrieving, analyzing, processing and presenting business analytics data to a user in a natural, conversational way. The method may comprise receiving a request from the user to provide the analytics data in the natural language format, converting the command in the natural language format into one or more Application Programming Interface (API) calls, retrieving generic data associated with the request of the user based on the API calls, generating a semantic model associated with the generic data and the user request, processing the retrieved generic data to generate analytics data, with the processing being based on the semantic model, communicating the analytics data to a chatbot, and converting, under control of the chatbot, the analytics data into a natural language format for delivering to the user.


Grant
Agency: European Commission | Branch: H2020 | Program: RIA | Phase: FCT-14-2014 | Award Amount: 5.00M | Year: 2015

The challenges of international police reform assistance are formidable. Conventional top-down institutional reform has proven neither effective nor sustainable. Community-based policing (COP) holds promise, however evaluations have pointed to a lack of in-depth understanding of police-community relations in police reform assistance. This project will conduct integrated social and technical research on COP in post-conflict countries in S.E. Europe, Asia, Africa and Central America. New knowledge, reflection on lessons learnt and best practices will support both national police and EU/International police reform assistance. The project will lead to a better understanding of police-community relations, and innovation in information and communication technology (ICT) for enhancing these relations in post-conflict countries undergoing serious security reform. Linking social and technological research, the project will study social, cultural, human security, legal and ethical dimensions of COP to understand how citizens and police can develop sustainable relations with the use of ICTs. We will explore how technological innovation can support COP in crime reporting and prevention. The project will explore ICT solutions to facilitate, strengthen and accelerate positive COP efforts and police-citizen interactions where trust levels are weak. Solutions will depend on the context and identified needs of end-users: communities, local police, national and international police (EU/UN), and policymakers, and may include citizen reporting, information monitoring, mobile value transfer, or improved organizational systems. The project includes a Policing Experts Network whose role is to support research planning, and dissemination and exploitation of findings, grounding the research in police practice. This will ensure findings are communicated by engaged police practitioners, and directly applied in COP education and training curricula in Europe and case countries.


Whitehead A.M.,Womens Health Services | Czarnogorski M.,Womens Health Services | Wright S.M.,Analytics Intelligence | Hayes P.M.,Womens Health Services | And 2 more authors.
American Journal of Public Health | Year: 2014

Increasing numbers of women veterans using Department of Veterans Affairs (VA) services has contributed to the need for equitable, high-quality care for women. The VA has evaluated performance measure data by gender since 2006. In 2008, the VA launched a 5-year women's health redesign, and, in 2011, gender disparity improvement was included on leadership performance plans. We examined data from VA Office of Analytics and Business Intelligence quarterly gender reports for trends in gender disparities in gender-neutral performance measures from 2008 to 2013. Through reporting of data by gender, leadership involvement, electronic reminders, and population management dashboards, VA has seen a decreasing trend in gender inequities on most Health Effectiveness Data and Information Set performance measures.


Grant
Agency: European Commission | Branch: H2020 | Program: RIA | Phase: DRS-14-2015 | Award Amount: 4.96M | Year: 2016

Modern critical infrastructures are becoming increasingly smarter (e.g. cities). Making the infrastructures smarter usually means making them smarter in normal operation and use: more adaptive, more intelligent But will these smart critical infrastructures (SCIs) behave equally smartly and be smartly resilient also when exposed to extreme threats, such as extreme weather disasters or terrorist attacks? If making existing infrastructure smarter is achieved by making it more complex, would it also make it more vulnerable? Would this affect resilience of an SCI as its ability to anticipate, prepare for, adapt and withstand, respond to, and recover? These are the main questions tackled by this proposal. The proposal envisages answering the above questions in several steps. (#1) By identifying existing indicators suitable for assessing resilience of SCIs. (#2) By identifying new smart resilience indicators (RIs) including those from Big Data. (#3) By developing a new advanced resilience assessment methodology (TRL4) based on smart RIs (resilience indicators cube, including the resilience matrix). (#4) By developing the interactive SCI Dashboard tool. (#5) By applying the methodology/tools in 8 case studies, integrated under one virtual, smart-city-like, European case study. The SCIs considered (in 8 European countries!) deal with energy, transportation, health, water Results #2, #3, #4 and #5 are a breakthrough innovation. This approach will allow benchmarking the best-practice solutions and identifying the early warnings, improving resilience of SCIs against new threats and cascading and ripple effects. The benefits/savings to be achieved by the project will be assessed by the reinsurance company participant. The consortium involves 7 leading end-users/industries in the area, 7 leading research organizations, supported by academia and lead by a dedicated European organization. External world leading resilience experts will be included in the CIRAB.


Loading Analytics Intelligence collaborators
Loading Analytics Intelligence collaborators