Takeru Igusa

Influenza Forecasting with Google Flu Trends

Objective We sought to develop a practical influenza forecast model, based on real-time, geographically focused, and easy to access data, to provide individual medical centers with advanced warning of the number of influenza cases, thus allowing sufficient time to implement an intervention. Secondly, we evaluated how the addition of a real-time influenza surveillance system, Google Flu Trends, would impact the forecasting capabilities of this model. Introduction Each year, influenza results in increased Emergency Department crowding which can be mitigated through early detection linked to an appropriate response. Although current surveillance systems, such as Google Flu Trends, yield near real-time influenza surveillance, few demonstrate ability to forecast impending influenza cases. Methods Forecasting models designed to predict one week in advance were developed from weekly counts of confirmed influenza cases over seven seasons (2004 – 2011) divided into training and out-of-sample verification sets. Forecasting procedures using classical Box-Jenkins, generalized linear, and autoregressive methods were employed to develop the final model and assess the relative contribution of external variables such as, Google Flu Trends, meteorological data, and temporal information. Models were developed and evaluated through statistical measures of global deviance and log-likelihood ratio tests. An additional measure of forecast confidence, defined as the percentage of forecast values, during an influenza peak, that are within 7 influenza cases of the actual data, was examined to demonstrate practical utility of the model. Results A generalized autoregressive Poisson (GARMA) forecast model integrating previous influenza cases with Google Flu Trends information provided the most accurate influenza case predictions. Google Flu Trend data was the only source of external information providing significant forecast improvements (p = 0.00002). The final model, a GARMA intercept model with the addition of Google Flu Trends, predicted weekly influenza cases during 4 out-of-sample outbreaks within 7 cases for 80% of estimates (Figure 1). Conclusions Integer-valued autoregression of influenza cases provides a strong base forecast model, which is enhanced by the addition of Google Flu Trends confirming the predictive capabilities of search query based syndromic surveillance. This accessible and flexible forecast model can be used by individual medical centers to provide advanced warning of future influenza cases.

Bayesian analysis of uncertainty for structural engineering applications

There has been recent interest in differentiating aleatory and epistemic uncertainties within the structural engineering context. Aleatory uncertainty, which is related to the inherent physical randomness of a system, has substantially different effects on the analysis and design of structures as compared with epistemic uncertainty, which is knowledge based. Bayesian techniques provide powerful tools for integrating, in a rigorous manner, the two types of uncertainties. In a purely probabilistic viewpoint, the uncertainties merge, resulting in widened probability densities. From the viewpoint of design or experimentation, however, the two types of uncertainties have widely different effects. The purpose of this paper is to develop insight into these effects, using Bayesian-based analytical expressions for the aleatory and epistemic uncertainties. The paper goes beyond standard Bayesian conjugate distributions by incorporating the effects of model uncertainty, where the applicability of two or more analytical models are used to describe the structure of interest. The influence of multiple model uncertainties is explored for two problems: the Bayesian updating process as data is acquired, and the design of simple parallel systems.

PDFs of Tropical Tropospheric Humidity: Measurements and Theory

The spatial variations in the probability density functions (PDFs) of relative humidity (RH) in the tropical and subtropical troposphere are examined using observations from the Atmospheric Infrared Sounder (AIRS) and the Microwave Limb Sounder (MLS) instruments together with a simple statistical model. The model, a generalization of that proposed by Sherwood et al., assumes the RH is determined by a combination of drying by uniform subsidence and random moistening events and has two parameters: r, the ratio of the drying time by subsidence to the time between moistening events, and k, a measure of the variability of the moistening events. The observations show that the characteristics of the PDFs vary between the tropics and subtropics, within the tropics or subtropics, and with altitude. The model fits the observed PDFs well, and the model parameters concisely characterize variations in the PDFs and provide information on the processes controlling the RH distributions. In tropical convective regions, the model PDFs that match the observations have large r and small k, indicating rapid random remoistening, which is consistent with direct remoistening in convection. In contrast, in the nonconvective regions there are small r and large k, indicating slower, less random remoistening, consistent with remoistening by slower, quasi-horizontal transport. The statistical model derived will be useful for quantifying differences between, or temporal changes in, RH distributions from different datasets or models, and for examining how changes in physical processes could alter the RH distribution.

Nonaxisymmetric vibration and acoustic radiation of a submerged cylindrical shell of finite length containing internal substructures

A modal‐based method is developed to analyze the acoustic radiation of axisymmetric submerged shells of finite length with internal substructures, subjected to nonaxisymmetric time‐harmonic loads. In this method, a variational principle is used to determine the relationship between the surface pressure and displacement of the shell, and Lagrange multipliers are used to account for the connections between the shell and the substructures. Fourier series expansions are used to represent the circumferential dependence of the surface pressure and displacement. The method is demonstrated by an extended analysis of a cylindrical shell with hemispherical elastic endcaps containing circular bulkheads. The dominant flexural wave numbers are identified from helical wave spectra. It is shown that the wave numbers of the dominant flexural waves are the same as those found in an infinite, fluid‐loaded cylindrical shell. The locus of wave numbers of the dominant flexural waves lies outside of the sonic cone for low to mid frequencies. However, it is shown that bulkheads can create amplification of the flexural waves with wave numbers within the sonic cone. This is confirmed by a computation of the net power radiated by the shell: The shell with bulkheads radiates more energy than the empty shell for ka≳0.5, where k is the acoustic wave number and a is the radius of the cylindrical shell.

Acoustic radiation from a cylindrical shell with an internal plate

Abstract The presence of a substructure in a submerged cylindrical shell gives rise to dynamic interactions which influence the radiated acoustic field. The lowest modes of the substructure may resonate vigorously with a higher order mode of the shell. In this paper, a Lagranges equations formulation is used to analyze the forced harmonic response of a submerged shell/substructure system. An infinitely long cylindrical shell is considered and the substructure is a plate with various types of connections to the shell. Harmonic forcing is applied to the structure and expressions are developed for the far-field radiated pressure. The frequency window method is used to simplify the analysis and to provide approximate solutions in a frequency range of interest.

Effect of a Vibrating Substructure on Acoustic Radiation from a Cylindrical Shell

The presence of a substructure in a cylindrical shell gives rise to dynamic interactions, including resonance effects, which have an influence on the radiated acoustic field. The lowest mode of the substructure may interact vigorously with a higher order mode of the shell. In this paper a Lagrangian formulation is used to analyze the forced motion of the immersed shell/substructure system. In addition, the frequency window method is used to reduce the complexity of the resulting expressions and to obtain approximate solutions in a frequency range of interest. The particular example is concerned with a spring-mass system, which is attached to the shell. Three cases of forced harmonic motion are considered. The radiated pressure is computed as a function of the frequency. The radiated pressure for the shell/substructure system is compared with that for a shell without an oscillator subjected to the same input force.

Modeling the Impact of School-Based Universal Depression Screening on Additional Service Capacity Needs: A System Dynamics Approach

Although it is widely known that the occurrence of depression increases over the course of adolescence, symptoms of mood disorders frequently go undetected. While schools are viable settings for conducting universal screening to systematically identify students in need of services for common health conditions, particularly those that adversely affect school performance, few school districts routinely screen their students for depression. Among the most commonly referenced barriers are concerns that the number of students identified may exceed schools’ service delivery capacities, but few studies have evaluated this concern systematically. System dynamics (SD) modeling may prove a useful approach for answering questions of this sort. The goal of the current paper is therefore to demonstrate how SD modeling can be applied to inform implementation decisions in communities. In our demonstration, we used SD modeling to estimate the additional service demand generated by universal depression screening in a typical high school. We then simulated the effects of implementing “compensatory approaches” designed to address anticipated increases in service need through (1) the allocation of additional staff time and (2) improvements in the effectiveness of mental health interventions. Results support the ability of screening to facilitate more rapid entry into services and suggest that improving the effectiveness of mental health services for students with depression via the implementation of an evidence-based treatment protocol may have a limited impact on overall recovery rates and service availability. In our example, the SD approach proved useful in informing systems’ decision-making about the adoption of a new school mental health service.

Incorporating Systems Science Principles into the Development of Obesity Prevention Interventions: Principles, Benefits, and Challenges

Systems modeling represents an innovative approach for addressing the obesity epidemic at the community level. We developed an agent-based model of the Baltimore City food environment that permits us to assess the relative impact of different programs and policies, alone and in combination, and potential unexpected consequences. Based on this experience, and a review of literature, we have identified a set of principles, potential benefits, and challenges. Some of the key principles include the importance of early and multilevel engagement with the community prior to initiating model development and continued engagement and testing with community stakeholders. Important benefits include improving community stakeholder understanding of the system, testing of interventions before implementation, and identification of unexpected consequences. Challenges in these models include deciding on the most important, yet parsimonious factors to consider, how to model food source and food selection behavior in a realistic yet transferable manner, and identifying the appropriate outcomes and limitations of the model.

Examining social norm impacts on obesity and eating behaviors among US school children based on agent-based model

BackgroundAlthough the importance of social norms in affecting health behaviors is widely recognized, the current understanding of the social norm effects on obesity is limited due to data and methodology limitations. This study aims to use nontraditional innovative systems methods to examine: a) the effects of social norms on school children’s BMI growth and fruit and vegetable (FV) consumption, and b) the effects of misperceptions of social norms on US children’s BMI growth.MethodsWe built an agent-based model (ABM) in a utility maximization framework and parameterized the model based on empirical longitudinal data collected in a US nationally representative study, the Early Childhood Longitudinal Study – Kindergarten Cohort (ECLS-K), to test potential mechanisms of social norm affecting children’s BMI growth and FV consumption.ResultsIntraclass correlation coefficients (ICC) for BMI were 0.064-0.065, suggesting that children’s BMI were similar within each school. The correlation between observed and ABM-predicted BMI was 0.87, indicating the validity of our ABM. Our simulations suggested the follow-the-average social norm acts as an endogenous stabilizer, which automatically adjusts positive and negative deviance of an individual’s BMI from the group mean of a social network. One unit of BMI below the social average may lead to 0.025 unit increase in BMI per year for each child; asymmetrically, one unit of BMI above the social average, may only cause 0.015 unit of BMI reduction. Gender difference was apparent. Social norms have less impact on weight reduction among girls, and a greater impact promoting weight increase among boys. Our simulation also showed misperception of the social norm would push up the mean BMI and cause the distribution to be more skewed to the left. Our simulation results did not provide strong support for the role of social norms on FV consumption.ConclusionsSocial norm influences US children’s BMI growth. High obesity prevalence will lead to a continuous increase in children’s BMI due to increased socially acceptable mean BMI. Interventions promoting healthy body image and desirable socially acceptable BMI should be implemented to control childhood obesity epidemic.

Resonance characteristics of connected subsystems: Theory and simple configurations

Abstract The dynamic properties of connected continuous subsystems are examined with use of analytical expressions for the modal properties. The analysis begins with a Lagrange multiplier formulation to develop a characteristic equation in terms of subsystem mobilities and impedances. The complexity of the problem is examined in terms of the order of the polynomial expressions in the characteristic equation. To obtain insight into the systems complicated dynamic characteristics, the complexity of the problem is reduced. It is shown that the reduction of complexity can be obtained only with a reduction of accuracy, but by retaining the dominant effects of the dynamics problem, the loss of accuracy is not excessive. The reduced problem is examined further to develop simple, yet powerful expressions for the modal properties which provide insight into the resonance characteristics of the connected subsystem problem. The results are useful as a complement to existing computational techniques for understanding and interpreting dynamic analysis results. In this paper, the theory and simplest configurations are examined, and in the companion paper, more general configurations are studied.