Structural Time Series Model Research Articles

Long-term impact of COVID-19 pandemic on antibiotic use in primary care: lessons to optimize antimicrobial use

ABSTRACT Background Lockdowns due to the COVID-19 pandemic was associated with changes in the pattern of use of antimicrobials, but persistence of changes after lockdowns has not been described. The objective was to describe the number of patients with dispensed antibiotic treatment and consumption of antibiotics in outpatients from primary care in Catalonia 18 months after the end of the emergency period. Research design and methods Data for the COVID-19 pandemic period was obtained from March 2020 to December 2021. Four high transmission rate (Rt > 100) periods were established. To compare data, a simple Bayesian structural time series model was used. Results The observed number of patients with dispensed antibiotics decreased respect to the estimated, especially during the four high transmission rate periods: April–May 2020 (lockdown period) (−42.57% and −42.68%); December 2020-February 2021 (−41.65%, −49.97% and −43.64%); October 2021 (−16.23%), and December 2021 (−20.16%). Overall antibiotic consumption was reduced by 23.37% (p = 0.002). These differences were mainly observed in those ≤ 15 years. Conclusions We describe the reduction in the number of patients with dispensed antibiotics and antibiotic consumption after the COVID-19 lockdown persisted in a period of recovery of healthcare accessibility. This information may help to improve antimicrobial use at the primary care level.

Causal impact analysis of weir opening on cyanobacterial blooms and water quality in the Yeongsan River, Korea: A bayesian structural time-series analysis and median difference test

The construction of weirs in Korea's Four Major Rivers Project has led to an increase in cyanobacterial blooms, posing environmental challenges. To address this, the government began opening weirs in 2017. However, interpreting experimental results has proven to be complex due to the multifaceted nature of blooms. This study aimed to assess the impact of opening the Juksan Weir on cyanobacterial blooms and water quality in the Yeongsan River. Using a median difference test (MDT) and causal impact analysis (CIA) with Bayesian structural time-series (BSTS) models, changes in cyanobacterial cell density (Cyano) and chlorophyll a concentration (Chl-a) before (January 2013 to June 2017) and after (July 2017 to December 2021) the weir-opening event were analyzed. The MDT revealed no significant change in Cyano post-weir opening (p = 0.267), but Chl-a significantly increased by 48.1 % (p < 0.01). As a result of CIA, Cyano decreased, albeit statistically insignificantly (p = 0.454), while Chl-a increased by 59.0 % (p < 0.01). These findings contradict the expectation that Cyano decrease due to the increased flow velocity resulting from weir opening. The absence of changes in Cyano and the increase in Chl-a can be attributed to several factors, including the constrained and inadequate duration of full weir opening combined with conducive conditions for the proliferation of other algae such as diatoms and green algae. These findings suggest that the effectiveness of weir opening in controlling Cyano may have been compromised by factors influencing the overall aquatic ecosystem dynamics. Further analysis revealed that factors such as elevated water temperatures (≥ 30 °C) and reduced flow rates (< 37 m3/s) contributed to the flourishing of cyanobacteria, whose concentrations exceeded 10,000 cells/mL. In analyzing causal relationships in environmental management, especially when there are complex causal interactions, the application of MDT and CIA provides valuable insights.

A Comprehensive Framework for Measuring the Immediate Impact of TV Advertisements: TV-Impact.

Measuring the immediate impact of television advertisements (TV ads) on online traffic poses significant challenges in many aspects. Nonetheless, a comprehensive consideration is essential to fully grasp consumer reactions to TV ads. So far, the measurement of this effect has not been studied to a large extent. Existing studies have either determined how a specific focus group, i.e., toddlers, people of a certain age group, etc., react to ads via simple statistical tests using a case study approach or have examined the effects of advertising with simple regression models. This study introduces a comprehensive framework called TV-Impact. The framework uses a Bayesian structural time-series model called CausalImpact. There are additional novel approaches developed within the framework. One of the novelties of TV-Impact lies in its dynamic algorithm for selecting control variables which are supporting data sources and presumed to be unaffected by TV ads. In addition, we proposed the concept of Group Ads to combine overlapping ads into a single ad structure. Then, Random Forest Regressor, which is a commonly preferred supervised learning method, is used to decompose the impact into single ads. The TV-Impact framework was applied to the data of iLab, a venture company in Turkey, and manages its companies' advertising strategies. The findings reveal that the TV-Impact model positively influenced the companies' strategies for allocating their TV advertisement budgets and increased the amount of traffic driven to company websites, serving as an effective decision support system.

Bayesian Structural Time Series Models for Predicting the $${\textrm{CO}}_2$$ Emissions in Afghanistan

Bayesian Structural Time Series Models for Predicting the $${\textrm{CO}}_2$$ Emissions in Afghanistan

Assessing the impact of COVID-19 interventions on the hand, foot and mouth disease in Guangdong Province, China: a Bayesian modeling study.

The non-pharmaceutical interventions (NPIs) against COVID-19 may have affected the transmission of hand, foot and mouth disease (HFMD). We aimed to assess the impact of the NPIs on HFMD in the high epidemic area of HFMD, Guangdong Province. The data of HFMD cases, etiological information, and meteorological factors in Guangdong from January 1, 2012, to December 31, 2021, were collected. Using a Bayesian structural time series (BSTS) model integrated counterfactual framework, we assessed the effect of NPIs on HFMD by different intervention periods, populations (gender, age, occupation), and cities. We further explored the correlation between the reduction of HFMD and socioeconomic factors in 21 cities. A total of 351,217 HFMD cases were reported and 455,327 cases were averted in Guangdong Province during 2020-2021 with a reduction of 84.94% (95%CI: 81.63-87.22%) in 2020 and 29.49% (95%CI: 15.26-39.54%) in 2021. The impact of NPIs on HFMD differed by age and gender. The effects of NPIs were more remarkable for children aged 0-2 years and scattered children. We found that the relative reductions in 21 cities were related to the composition ratio of children and COVID-19 incidence. The reduction of HFMD incidence was significantly associated with COVID-19 NPIs, and school closure was an effective intervention to prevent HFMD outbreaks. Our findings will contribute to the development of HFMD prevention and control measures.

Evaluating the effects of the 2017 National Health Insurance coverage expansion on amenable mortality and its disparities between areas in South Korea using Bayesian structural time-series models

To improve the low coverage rate of the National Health Insurance (NHI), South Korea implemented the NHI coverage expansion plan in 2017 to cover medically essential non-covered services and reduce copayment rates. This study aimed to estimate the effects of the 2017 NHI coverage expansion on amenable mortality and its disparities between areas in South Korea under a controlled interrupted time-series design using Bayesian structural time-series models. Age-standardized amenable mortality rates and rate differences (RDs) and rate ratios (RRs) between areas for amenable mortality were calculated monthly between July 2012 and December 2021 and used as the response series. The non-equivalent control series were monthly non-avoidable mortality rates and their regional disparities. After the coverage expansion, amenable mortality rates decreased for both males (−8.8%, 95% credible interval [CrI] −13.4% to −3.9%) and females (−8.3%, 95% CrI −13.4% to −2.4%), with the largest decline in the non-Seoul-Capital metropolitan area (−11.6%, 95% CrI −16.5% to −6.3%) rather than the Seoul Capital Area (−7.5%, 95% CrI −11.9% to −2.5%) and a non-significant reduction in the non-Seoul-Capital non-metropolitan area in females. RDs and RRs between areas for amenable mortality decreased non-significantly (−16.2%, 95% CrI −31.3% to 2.6% for RD and −1.2%, 95% CrI −3.7% to 1.5% for RR), except for a significant decrease in RD in males (−21.8%, 95% CrI −38.0% to −1.5%), and decreased less in females than in males. The coverage expansion was generally effective in reducing amenable mortality rates by area, but had limited effects in closing amenable mortality disparities between areas, favoring males and the non-Seoul-Capital metropolitan area. These results implied that additional measures are necessary to improve access to quality health care for females and underserved areas to enhance the effectiveness of the coverage expansion.

Bayesian Structural Time Series and Geographically Weighted Logistic Regression Modelling Impacts of COVID-19 Lockdowns on the Spatiotemporal Patterns of London’s Crimes

Given the paramount impacts of COVID-19 on people’s lives in the capital of the UK, London, it was foreseeable that the city’s crime patterns would have undergone significant transformations, especially during lockdown periods. This study aims to testify the crime patterns’ changes in London, using data from March 2020 to March 2021 to explore the driving forces for such changes, and hence propose data-driven insights for policy makers and practitioners on London’s crime deduction and prevention potentiality in post-pandemic era. (1) Upon exploratory data analyses on the overall crime change patterns, an innovative BSTS model has been proposed by integrating restriction-level time series into the Bayesian structural time series (BSTS) model. This novel method allows the research to evaluate the varied effects of London’s three lockdown periods on local crimes among the regions of London. (2) Based on the predictive results from the BSTS modelling, three regression models were deployed to identify the driving forces for respective types of crime experiencing significant increases during lockdown periods. (3) The findings solidified research hypotheses on the distinct factors influencing London’s specific types of crime by period and by region. In light of the received evidence, insights on a modified policing allocation model and supporting the unemployed group was proposed in the aim of effectively mitigating the surges of crimes in London.

Integrating Bayesian Approaches and Expert Knowledge for Forecasting Continuous Glucose Monitoring Values in Type 2 Diabetes Mellitus.

Precise and timely forecasting of blood glucose levels is essential for effective diabetes management. While extensive research has been conducted on Type 1 diabetes mellitus, Type 2 diabetes mellitus (T2DM) presents unique challenges due to its heterogeneity, underscoring the need for specialized blood glucose forecasting systems. This study introduces a novel blood glucose forecasting system, applied to a dataset of 100 patients from the ShanghaiT2DM study. Our study uniquely integrates knowledge-driven and data-driven approaches, leveraging expert knowledge to validate and interpret the relationships among diabetes-related variables and deploying the data-driven approach to provide accurate forecast blood glucose levels. The Bayesian network approach facilitates the analysis of dependencies among various diabetes-related variables, thus enabling the inference of continuous glucose monitoring (CGM) trajectories in similar individuals with T2DM. By incorporating past CGM data including inference CGM trajectories, dietary records, and individual-specific information, the Bayesian structural time series (BSTS) model effectively forecasts glucose levels across time intervals ranging from 15 to 60 minutes. Forecast results show a mean absolute error of 6.41 ±0.60 mg/dL, a root mean square error of 8.29 ±0.95 mg/dL, and a mean absolute percentage error of 5.28 ±0.33%, for a 15-minute prediction horizon. This study makes the first application of the ShanghaiT2DM dataset for glucose level forecasting, considering the influences of diabetes-related variables. Its findings establish a foundational framework for developing personalized diabetes management strategies, potentially enhancing diabetes care through more accurate and timely interventions.

Impact of organised colorectal cancer screening on age-specific population incidences: evidence from a quasi-experimental study in Sweden

Colorectal cancer (CRC) incurs a significant disease burden globally. Organised CRC screening programmes have been widely implemented for early detection and prevention. To understand the public health impact of these programmes, quantitative evidence of changes in overall and age-specific population incidences is fundamental. We aimed to provide such evidence by exploiting a time lag in the implementation of organised screening in Sweden: two out of 21 regions (these two regions comprise nearly 20% of the total Swedish population) have offered organised screening since 2008; the other regions have offered CRC screening since 2021. Using registry data on diagnosed CRC cases and socio-demographics for all regions in Sweden over the period 1970–2019, Bayesian structural time series modelling and difference-in-differences were applied to analyse the impact of screening on age-specific population incidences over time (CRC cases per 100.000 persons/year). After inviting birth-year cohorts aged 60–69 years for stool-based testing, the incidence rate in the 70–74-year age group decreased significantly over time, with an average reduction of − 44·40 (95% CI − 58·15 to − 31·31) from 2011 to 2019 in the intervention regions. In the overall population aged 60–74 years, there was a net incidence decrease of − 7·99 (95% CI − 13·85 to − 2·39) since the initiation of organised screening in the intervention regions (2008–2019). Organised CRC screening for 60–69-year-olds generated a change in age-specific incidence patterns with a long-lasting incidence decrease in the 70–74-year-old population, implying reductions in the excess mortality and burden of the disease.

Evaluation of the impact of the implementation of the new primary health care information system on records of patient care and home visits in Brazil

The replacement of the Primary Care Information System (SIAB, 1998-2015), as of January 2016, by the new Health Information System for Primary Care (SISAB) determined new forms of collecting, processing, and using information, with a possible impact on the records of activities carried out in primary health care in Brazil. This study aimed to evaluate the implementation impact of the new information system on records of physicians' and nurses' patient care and home visits of community health workers (CHW) in Brazil from 2007 to 2019. To this end, a Bayesian structural time-series model approach was used, based on a diffuse state-space regression. From 2016 to 2019, 463.47 million physician care, 210.61 million nursing care, and 1.28 billion CHW visits were recorded. Following the trend recorded before the implementation, 598.86 million, 430.46 million, and 1.5 billion physician and nursing appointments and CHW visits would be expected, respectively. In relative terms, there was a decrease of 25% in physician care, 51% in nursing care, and 15% in CHW visits when compared to the value expected by the Bayesian method. The negative impact on the records of patient care and home visits identified in this study, whether due to difficulties in adapting to the new system or a reduction in improper records, must be investigated so that the challenge of improving the primary care information system can be understood and overcome in a planned way.

Window dressing in the banking sector of an emerging economy: evidence from aggregate data

PurposeThis study aims to investigate window dressing as practiced by commercial banks in Kuwait, using monthly aggregate balance sheet data covering the period January 1993 to December 2017.Design/methodology/approachThis study applies the structural time series model to decompose an observed time series into unobserved components based on monthly data covering January 1993 to December 2017 on the consolidated balance sheet of commercial banks in Kuwait.FindingsThe empirical results indicate that Kuwaiti commercial banks indulge in upward window dressing to boost size and liquidity. This kind of behaviour is indicated by a statistically significant rise in assets under the control of banks in December, followed by a statistically significant decline in January. The operation is funded by borrowing, leading to a December rise and a January fall in foreign and other liabilities, which are also under the control of commercial banks.Originality/valueThis study uses a novel methodology to detect window dressing based on the seasonal behaviour of balance sheet items. This study suggests a unified framework for the motives, targets, types and consequences of window dressing and how they are related.

Effectiveness of public policies related to traffic emissions in improving air quality in Brazil: A causal inference study using Bayesian structural time-series models

Air pollution is a major environmental health risk, particularly in low- and middle-income countries. In Brazil, the government has implemented progressive stages of traffic emission controls since 1986 through a program called “Proconve”. In this study, we evaluated the effectiveness of two of these stages, Proconve stages L-5 and P-7, in reducing air pollution in the 10 most populated cities in Brazil. We used Bayesian structural time-series models to estimate the causal effect of the vehicle emission control measures on the air pollution time series for each city and air pollutant (PM2.5, O3, NO2, and CO), accounting for potential confounding factors such as temperature and humidity. While previous studies have shown positive effects of control measures on air quality, our study suggests that the effects may not always be statistically significant, which has implications for the design and interpretation of future studies. Our study provides valuable information for policymakers in Brazil and adds to the international literature on the effectiveness of air pollution control measures. Despite the lack of statistically significant results, our study highlights the importance of considering statistical power and unmeasured confounding factors when evaluating the effectiveness of air pollution control measures.

Bayesian Structural Time Series Model for Forecasting the Composite Stock Price Index in Indonesia

One of the models that can be used to predict time series data is the Bayesian Structural Time Series (BSTS) model. The BSTS model is a more modern model and can handle data movement better. In the BSTS model, the Markov Chain Monte Carlo (MCMC) sampling algorithm is used to simulate the posterior distribution, which smoothes the forecasting results over a large number of potential models using Bayesian averaging models. The purpose of this study was to obtain the best BSTS model for Composite Stock Price Index (CSPI) data in Indonesia based on the state component and the number of MCMC iterations, and obtain forecasting results for CSPI value in Indonesia for the next 24 months, namely the period July 2023 to June 2024. The results obtained are based on a comparison of the R-square values in the model, the BSTS model with local linear trend and seasonal state components, and the number of MCMC iterations n = 5 00 is the best BSTS model that can be used for forecasting the CSPI value in Indonesia with an R-square value of 99.96%. The results of forecasting the CSPI value in Indonesia for the period July 2023 to June 2024 range from 6589 to 6760, with the lowest forecasting value in October 2023 and the highest in March 2023.

Quantitative easing effectiveness: Evidence from Euro private assets

AbstractProponents of quantitative easing (QE) unconventional policy have rather overstated some evidence that structural time series models do not predict long‐term asset prices and yields as well as naive random walk forecasts, implying that predictions of price reversals cannot be profitable and, therefore, that QE effects are not transitory. Indeed, in this work we present evidence that naive models do not outperform structural vector autoregressive and Markov switching models in out‐of‐sample forecasting of corporate bond yields purchased by the European Central Bank, when the information set includes base money growth. It turns out that structural time series models provide additional information regarding the likelihood of price reversals, thus motivating investors to offset the effects of QE interventions if they perceive unconventional monetary policy regimes as temporary.

Structural Time Series Model using Hamiltonian Monte Carlo for Rice Price

Although forecasts of future events are simply uncertain, predicting is one of the most important aspects of future planning. Accurate rice price predictions tend to be helpful for wholesalers, producers, and farmers to develop plans and strategies to reduce the risks that can be faced. Structural time series models are the most plausible alternative for long-term forecasting. This paper proposes an alternate method for modeling average rice prices using structural time series along with Bayesian parameter inference via Hamiltonian Monte Carlo (HMC). The model has been built using the monthly average wholesale rice price from January 2010 to December 2019. For working out both structural time series and HMC, the TensorFlow Probability Library was used. Linear trend, seasonal, and autoregressive components were combined as an additive model to the structural time model. The proposed Hamiltonian parameter produces an optimal acceptance rate. Their trace plot was used to diagnose the convergence of their chain. One of the predictive accuracy of models was assessed using the mean absolute percent error (MAPE). Through both single and multiple chain iterations, the prediction accuracy of a year-ahead is highly accurate, with MAPE less than 2%. Long-term iteration draws during Hamiltonian Monte Carlo should be considered when attempting to achieve more convergence.

Modeling and forecasting industrial electricity demand for Saudi Arabia: Uncovering regional characteristics

This study aims to investigate and forecast Saudi Arabia's industrial electricity consumption at the regional level. By applying structural time series modeling to annual data from 1990 to 2019, we find the electricity demand's long-run income and price elasticities to vary across regions. Also, the underlying energy demand trend analysis indicates efficiency improvements in industrial electricity consumption patterns in all regions. In the low-case scenario, we project the 2030 demand to be 11.6 terawatt-hours (TWh), 63.5 TWh, 0.7 TWh, and 6.8 TWh for the central, eastern, southern, and western operation areas, respectively, with the total demand forecast to be 82.5 TWh. The study's findings can help policymakers understand the effect of future energy policies and economic growth trajectories.

Measuring the causal effect of Panama Canal expansion on Latin America and the Caribbean's economic growth: a Bayesian structural time series approach

PurposeThis study provides empirical evidence on the impact of the Panama Canal expansion (PCE) on the economies of Latin American and Caribbean (LAC) countries, particularly in light of the emergence of larger container ships such as neo-Panamax and post-Panamax vessels.Design/methodology/approachThis study uses the Bayesian structural time Series (BSTS) model to evaluate the economic effects of the PCE on 21 countries within the LAC region. It utilized the World Bank's gross domestic product (GDP) figures between 2000 and 2019 as the primary variable, alongside the human development index (HDI) (X1), container throughput (TEU) (X2) and unemployment rates (UNEMPL) (X3) covariates. This allowed a precise and robust approach to analyzing time series data while accounting for uncertainties and allowing the inclusion of various components and external factors.FindingsThe findings revealed that the PCE has a positive and statistically significant impact on most countries within the Caribbean Transshipment Triangle, ranging from 9.2% in Belize to 46% in Cuba. This suggests that the causal effect of the PCE on regional economies was not confined to any specific type of economy or geographical location within the LAC region. Where the growth rates were statistically insignificant, primarily in some Latin American countries, it coincided with countries that are primarily driven by exports and service industries, where bulk and oil tanker vessels are likely to be the main carriers for exports rather than container vessels.Originality/valueThe practical implications of this research are crucial for various stakeholders in the maritime industry and economic planning. The factors influencing economic growth resulting from investing in maritime activities are vital for decision-makers to create policies that lead to positive outcomes and sustainable development in regions and countries with flourishing maritime industries. The methodology and findings have significant implications for governments, managers, professionals, policy-makers and investors.

A systematic survey of the online trade in elephant ivory in Singapore before and after a domestic trade ban

AbstractThe illegal online trade in wildlife has increased significantly over the last 2 decades. Applying a systematic survey approach, we examined the online trade in elephant ivory before and after the implementation of a domestic trade ban in Singapore, to understand the forms of elephant ivory offered for sale, the e-commerce sites used and the effect of the domestic trade ban. Using elephant ivory and lookalikes in form, shape and colour as proxies for elephant ivory, we found that most of the online listings consisted of bird cages and related accessories such as bird feeding cups and ornaments used to adorn them. After the domestic trade ban in elephant ivory was implemented on 1 September 2021, a 76% drop in total listings was observed. Causal impact analysis indicated a strong causal effect of the domestic trade ban, resulting in an 83% drop in the mean number of new listings found weekly relative to the predicted counterfactual market response based on a Bayesian structural time-series model in the absence of the ban. E-commerce sites play an important role in combatting illegal wildlife trade online by removing flagged listings and sending notifications to raise awareness of the ban. We found textual descriptions of products to be more effective for preliminary differentiation of legal and potentially illegal listings compared to the detection of Schreger lines in images, although this remains a significant challenge for law enforcement. Our findings can inform future efforts to develop automated detection methods for elephant ivory in online markets.

가정폭력 범죄에 대한 체포우선주의 정책이 여성의 인명 피해 감소에 미친 영향

This study examines the relationship between arrest prioritization policies and arrests of domestic violence perpetrators and the number of female victims killed using a time series analysis, as well as the relationship between judicial action against domestic violence perpetrators and the number of female victims killed by male partners. Data on the number of female victims killed by their male partners came from the Korea Women's Helpline, and data on the number of female homicide victims came from the National Police Agency, including arrests of domestic violence perpetrators. The intervention analysis in time series with ARIMA model and Bayesian Structural Time-Series model(BSTS) showed that the number of arrests of domestic violence perpetrators increased significantly after implementing the arrest prioritization policy in 2013, but the number of female homicide victims decreased without a significant effect. The number of female victims killed by their male partners was inversely related to domestic violence-related measures, with the number of female victims decreasing as the number of arrests increased. An essential aspect of this study is the use of ARIMA intervention time series and BSTS in assessing the effectiveness of national crime prevention measures, as well as testing the impact of arrest prioritization on domestic violence offenses.

Effects of tougher school zone laws on road traffic safety in school zones for children in South Korea

IntroductionNew school zone laws in Korea were enforced in March 2020 to increase penalties for traffic violations and preferentially install road safety equipment including automated traffic enforcement devices in school zones. This study aimed to evaluate the effectiveness of tougher school zone laws on school-zone-related road traffic outcome (RTO) rates for children aged 0–12 years in Korea. MethodsMonthly child RTO rates in school zones were calculated per million vehicles and per million children using data on police-reported road traffic crashes, registered motor vehicles, and registered resident population between 2015 and 2021: the intervention series involved school-zone-related rates and the non-equivalent control series involved total and non-school-zone-related rates. The effectiveness of the laws on RTO rates in school zones was assessed using the Bayesian structural time-series models under the controlled interrupted time-series design. ResultsThe results found that considering the trends in the overall and non-school-zone-related RTO rates, enforcement of the laws did not significantly reduce any school-zone-related RTO rates for children aged 0–12 years, both per million vehicles and per million children: equivalent fatality, road traffic injury (RTI), severe RTI, minor RTI, and road traffic crash rates. ConclusionsThe ineffectiveness of Korea's new school zone laws on child RTO rates in school zones implies that other effective measures need to be accompanied to improve children's road traffic safety in school zones, rather than just focusing on raising the severity of punishment. Further research is needed to ensure the effectiveness of school zone laws and measures to mitigate child road traffic casualties in school zones.