Forecasting Areawide Hospital Utilization: A Comparison of Five Univariate Time Series Techniques

Time series analysis is one of the methods health services researchers, managers and planners have to examine and predict utilization over time. The focus of this study is univariate time series techniques, which model the change in a dependent variable over time, using time as the only independent variable. These techniques can be used with administrative healthcare databases, which typically contain reliable, time-specific utilization variables, but may lack adequate numbers of variables needed for behavioral or economic modeling. The inpatient discharge database of the Department of Veterans Affairs, the Patient Treatment File, was used to calculate monthly time series over a six-year period for the nation and across US Census Bureau regions for three hospital utilization indicators: Average length of stay, discharge rate, and multiple stay ratio, a measure of readmissions. The first purpose of this study was to determine the accuracy of forecasting these indicators 24 months into the future using five univariate time series techniques. In almost all cases, techniques were able to forecast the magnitude and direction of future utilization within a 10% mean monthly error. The second purpose of the study was to describe time series of the three hospital utilization indicators. This approach raised several questions concerning Department of Veterans Affairs hospital utilization.

Download Full-text

Veterans' Labor Force Participation: What Role Does the VA's Disability Compensation Program Play?

The American Economic Review ◽

10.1257/aer.p20151062 ◽

2015 ◽

Vol 105 (5) ◽

pp. 131-136 ◽

Cited By ~ 8

Author(s):

Courtney Coile ◽

Mark Duggan ◽

Audrey Guo

Keyword(s):

Labor Force ◽

Labor Force Participation ◽

Time Trends ◽

Veterans Affairs ◽

Department Of Veterans Affairs ◽

Economic Shocks ◽

Dc Program ◽

Compensation Program ◽

Over Time ◽

Disability Compensation

We explore time trends in the labor force participation of veterans and non-veterans and investigate whether they are consistent with a rising role for the Department of Veterans Affairs' Disability Compensation (DC) program, which pays benefits to veterans with service-connected disabilities and has grown rapidly since 2000. Using 35 years of March CPS data, we find that veterans' labor force participation declined over time in a way that coincides closely with DC growth and that veterans have become more sensitive to economic shocks. Our findings suggest that DC program growth has contributed to recent declines in veterans' labor force participation.

Download Full-text

Assessing Hospital Disaster Readiness Over Time at the US Department of Veterans Affairs

Prehospital and Disaster Medicine ◽

10.1017/s1049023x16001266 ◽

2016 ◽

Vol 32 (1) ◽

pp. 46-57 ◽

Cited By ~ 4

Author(s):

Claudia Der-Martirosian ◽

Tiffany A. Radcliff ◽

Alicia R. Gable ◽

Deborah Riopelle ◽

Farhad A. Hagigi ◽

...

Keyword(s):

Phase I ◽

Emergency Management ◽

Phase Ii ◽

Veterans Affairs ◽

Assessment Tools ◽

Department Of Veterans Affairs ◽

The Us ◽

Two Phases ◽

Hospital Readiness ◽

Over Time

AbstractIntroductionThere have been numerous initiatives by government and private organizations to help hospitals become better prepared for major disasters and public health emergencies. This study reports on efforts by the US Department of Veterans Affairs (VA), Veterans Health Administration, Office of Emergency Management’s (OEM) Comprehensive Emergency Management Program (CEMP) to assess the readiness of VA Medical Centers (VAMCs) across the nation.Hypothesis/ProblemThis study conducts descriptive analyses of preparedness assessments of VAMCs and examines change in hospital readiness over time.MethodsTo assess change, quantitative analyses of data from two phases of preparedness assessments (Phase I: 2008-2010; Phase II: 2011-2013) at 137 VAMCs were conducted using 61 unique capabilities assessed during the two phases. The initial five-point Likert-like scale used to rate each capability was collapsed into a dichotomous variable: “not-developed=0” versus “developed=1.” To describe changes in preparedness over time, four new categories were created from the Phase I and Phase II dichotomous variables: (1) rated developed in both phases; (2) rated not-developed in Phase I but rated developed in Phase II; (3) rated not-developed in both phases; and (4) rated developed in Phase I but rated not- developed in Phase II.ResultsFrom a total of 61 unique emergency preparedness capabilities, 33 items achieved the desired outcome – they were rated either “developed in both phases” or “became developed” in Phase II for at least 80% of VAMCs. For 14 items, 70%-80% of VAMCs achieved the desired outcome. The remaining 14 items were identified as “low-performing” capabilities, defined as less than 70% of VAMCs achieved the desired outcome.Conclusion:Measuring emergency management capabilities is a necessary first step to improving those capabilities. Furthermore, assessing hospital readiness over time and creating robust hospital readiness assessment tools can help hospitals make informed decisions regarding allocation of resources to ensure patient safety, provide timely access to high-quality patient care, and identify best practices in emergency management during and after disasters. Moreover, with some minor modifications, this comprehensive, all-hazards-based, hospital preparedness assessment tool could be adapted for use beyond the VA.Der-MartirosianC, RadcliffTA, GableAR, RiopelleD, HagigiFA, BrewsterP, DobalianA. Assessing hospital disaster readiness over time at the US Department of Veterans Affairs. Prehsop Disaster Med. 2017;32(1):46–57.

Download Full-text

Patients with schizophrenia and breast cancer: Adjuvant radiation treatment.

Journal of Clinical Oncology ◽

10.1200/jco.2013.31.15_suppl.e17506 ◽

2013 ◽

Vol 31 (15_suppl) ◽

pp. e17506-e17506

Author(s):

Lauren Davies ◽

Kashif Abdullah ◽

Radhakrishna Janardhan ◽

Mark Hwang ◽

Michael Farasatpour ◽

...

Keyword(s):

Breast Carcinoma ◽

Radiation Treatment ◽

Care Delivery ◽

Veterans Affairs ◽

Tnm Stage ◽

Patient Treatment ◽

Adjuvant Radiation ◽

Department Of Veterans Affairs ◽

Medical Centers ◽

History Of

e17506 Background: Schizophrenia affects about 1% of subjects in all populations studied so far. It impairs medical care delivery. We sought to evaluate how patients with schizophrenia who are later diagnosed with breast carcinoma fare when adjuvant radiation therapy (ART) is indicated. Methods: We searched the Patient Treatment File (PTF) of the Department of Veterans Affairs (DVA) to identify subjects with schizophrenia who later developed breast carcinoma and were treated in DVA Medical Centers (DVAMCs) from 1999 - 2005. Chart-based clinical data from the DVAMCs where the subjects had been treated supplemented PTF data. Results: 42 patients had preexisting schizophrenia, later developed breast carcinoma, and were candidates for ART. There were 31 women (74%) and 11 men (26%). 27 of the 42 study subjects had records specifying TNM stage; 18 of the 27 (67%) had TNM stages III-IV. 31 subjects had records about compliance with indicated medical therapies; 24 (77%) had previously been non-compliant. 39 subjects had records regarding therapy delay; 20 (51%) had previously delayed medically indicated therapy. Of the 42 subjects who were candidates for ART based on TNM stage, we found data about the decision to recommend ART in 37; only 23 (26%) were offered ART and 3 of those 23 (26%) refused it. Of the 6 subjects who refused ART, 4 had been non-compliant with previous medically indicated care, 3 had delayed initial cancer treatment, 4 had documented suicidal ideation, and 2 had documented homicidal ideation before being offered ART. Conclusions: Subjects with schizophrenia and breast carcinoma often do not understand their illnesses well. They often do not comply with recommended standard therapies such as ART. Treatments that rely on ART are likely to be met with non-compliance. A history of non-compliance appears to be a predictor of non-compliance with ART. Our results should be of interest to caregivers because we describe ways to avoid suboptimal ART in patients with breast carcinoma. Breast-conserving multi-modality treatment with ART is frequently not appropriate; radical surgery is often preferable.

Download Full-text

0462 Identifying and Characterizing Insomnia Symptom Groups Across the Deployment Cycle in Current Army Soldiers

SLEEP ◽

10.1093/sleep/zsaa056.459 ◽

2020 ◽

Vol 43 (Supplement_1) ◽

pp. A177-A177

Author(s):

K E Miller ◽

E M Boland ◽

E A Klingaman ◽

P R Gehrman

Keyword(s):

Research And Development ◽

Military Personnel ◽

Veterans Affairs ◽

Department Of Veterans Affairs ◽

Insomnia Symptom ◽

Symptom Status ◽

Army Soldiers ◽

Insomnia Symptoms ◽

Over Time ◽

New Onset

Abstract Introduction Most research conducted on insomnia and its development in military personnel focuses on cross-sectional data, precluding examination of the course of sleep changes over time. The present study characterized Army Soldiers based on insomnia symptom status trajectory from pre to post-deployment and explored baseline factors predictive of these trajectories in a sample of 7,245 soldiers across 3 Brigade Combat Teams. Methods Data were analyzed from the Army Study to Assess Risk and Resilience in Service members (STARRS)-All-Army Study (AAS) Pre Post Deployment Study, using surveys that captured 1-2 months pre-deployment, during deployment, and 6-months post-deployment. Insomnia symptom status was defined at each timepoint as insomnia symptoms that interfered with one or more domains of functioning at least some of the time in the past month. Theoretically-derived variables linked to sleep disturbance were selected as predictors of insomnia symptom trajectory and evaluated using a general linear selection model. Results Four trajectories characterized the majority of the sample: ‘good sleepers’ (no insomnia symptoms across time; 44.4%), ‘non-remitting new onset insomnia’ (no pre-deployment insomnia, developed insomnia symptoms during deployment that remained at 6 months; 22.8%), ‘deployment-only insomnia symptoms’ (no pre-deployment insomnia, developed insomnia during deployment but recovered by follow-up; 12.8%), and ‘chronic insomnia’ (insomnia both pre- and post-deployment; 7.4%). Several pre-deployment factors predicted insomnia trajectory, the strongest of which were past six-month attention deficit disorder symptoms, number of lifetime exposures to potentially traumatic events, and past month depression symptoms. Conclusion Insomnia is one of the most common reasons that military personnel seek behavioral health treatment and is associated with poorer military readiness. Better characterization and identification of insomnia symptoms over time can improve intervention during post-deployment transitions, particularly for those with new onset insomnia that does not remit. Support Cooperative agreement U01MH087981 (Department of the Army; U.S. Department of Health and Human Services, National Institutes of Health, National Institute of Mental Health); U.S. Department of Veterans Affairs, Clinical Science Research and Development-IK2CX001874-PI:Katherine E. Miller, IK2CX001501-PI:Elaine M. Boland; Rehabilitation Research and Development-1IK2RX001836-PI:Elizabeth A. Klingaman. The views expressed here are those of the authors and do not represent the Department of Veterans Affairs.

Download Full-text

Semi-Supervised Time Series Anomaly Detection Based on Statistics and Deep Learning

Applied Sciences ◽

10.3390/app11156698 ◽

2021 ◽

Vol 11 (15) ◽

pp. 6698

Author(s):

Jehn-Ruey Jiang ◽

Jian-Bin Kao ◽

Yu-Lin Li

Keyword(s):

Time Series ◽

Deep Learning ◽

Anomaly Detection ◽

Stationary Time Series ◽

Test Time ◽

Pearson Product Moment Correlation ◽

Novel Technologies ◽

Univariate Time Series ◽

Comparison Results ◽

Over Time

Thanks to the advance of novel technologies, such as sensors and Internet of Things (IoT) technologies, big amounts of data are continuously gathered over time, resulting in a variety of time series. A semi-supervised anomaly detection framework, called Tri-CAD, for univariate time series is proposed in this paper. Based on the Pearson product-moment correlation coefficient and Dickey–Fuller test, time series are first categorized into three classes: (i) periodic, (ii) stationary, and (iii) non-periodic and non-stationary time series. Afterwards, different mechanisms using statistics, wavelet transform, and deep learning autoencoder concepts are applied to different classes of time series for detecting anomalies. The performance of the proposed Tri-CAD framework is evaluated by experiments using three Numenta anomaly benchmark (NAB) datasets. The performance of Tri-CAD is compared with those of related methods, such as STL, SARIMA, LSTM, LSTM with STL, and ADSaS. The comparison results show that Tri-CAD outperforms the others in terms of the precision, recall, and F1-score.

Download Full-text

Insomnia and Insomnia-Related Care in the Department of Veterans Affairs: An Electronic Health Record Analysis

International Journal of Environmental Research and Public Health ◽

10.3390/ijerph18168573 ◽

2021 ◽

Vol 18 (16) ◽

pp. 8573

Author(s):

Adam D. Bramoweth ◽

Caitlan A. Tighe ◽

Gregory S. Berlin

Keyword(s):

Behavioral Therapy ◽

Veterans Affairs ◽

Veteran Health ◽

Department Of Veterans Affairs ◽

Regional Network ◽

Health Records ◽

Trends Over Time ◽

Electronic Health ◽

Over Time ◽

Roll Out

The objective was to examine insomnia and insomnia-related care within a regional network of Department of Veterans Affairs (VA) facilities since the VA roll-out of cognitive behavioral therapy for insomnia (CBT-I) in 2011. A retrospective analysis of VA electronic health records (EHR) data from 2011 to 2019 was conducted. The annual and overall prevalence of four insomnia indicators was measured: diagnoses, medications, consultations for assessment/treatment, and participation in CBT-I. Also examined were sociodemographic and clinical differences among veterans with and without an insomnia indicator, as well as differences among the four individual insomnia indicators. The sample included 439,887 veterans, with 17% identified by one of the four indicators; medications was most common (15%), followed by diagnoses (6%), consults (1.5%), and CBT-I (0.6%). Trends over time included increasing yearly rates for diagnoses, consults, and CBT-I, and decreasing rates for medications. Significant differences were identified between the sociodemographic and clinical variables across indicators. An evaluation of a large sample of veterans identified that prescription sleep medications remain the best way to identify veterans with insomnia. Furthermore, insomnia continues to be under-diagnosed, per VA EHR data, which may have implications for treatment consistent with clinical practice guidelines and may negatively impact veteran health.

Download Full-text