Bayesian Time Series Model Research Articles

Influence of population mobility on electricity consumption in seven U.S. cities during the COVID-19 pandemic

We examine the impact of the COVID-19 pandemic on electricity consumption in seven U.S. cities. A high-level analysis reveals that reductions in electricity consumption were mostly short-term, mainly when lockdowns were first introduced. Bayesian structural time series modeling was used to decompose electricity consumption into multiple tailored components to better understand the pandemic's impact. We find that models incorporating population mobility achieved high accuracy rates using pre-pandemic data and even better rates using post-pandemic data. Electricity usage dropped during the first six weeks of the pandemic in all but one of the cities studied.

A Bayesian non-stationary heteroskedastic time series model for multivariate critical care data.

We propose a multivariate GARCH model for non-stationary health time series by modifying the observation-level variance of the standard state space model. The proposed model provides an intuitive and novel way of dealing with heteroskedastic data using the conditional nature of state-space models. We follow the Bayesian paradigm to perform the inference procedure. In particular, we use Markov chain Monte Carlo methods to obtain samples from the resultant posterior distribution. We use the forward filtering backward sampling algorithm to efficiently obtain samples from the posterior distribution of the latent state. The proposed model also handles missing data in a fully Bayesian fashion. We validate our model on synthetic data and analyze a data set obtained from an intensive care unit in a Montreal hospital and the MIMIC dataset. We further show that our proposed models offer better performance, in terms of WAIC than standard state space models. The proposed model provides a new way to model multivariate heteroskedastic non-stationary time series data. Model comparison can then be easily performed using the WAIC.

Multivariate Bayesian Time-Series Model with Multi-temporal Convolution Network for Forecasting Stock Market During COVID-19 Pandemic

The paper proposes a hybrid algorithm for forecasting multiple correlated time-series data, which consists of two main steps. First, it employs a multivariate Bayesian structural time series (MBSTS) approach as a base step. This method allows for the incorporation of potentially high-dimensional regression components, and it utilizes spike and slab priors to identify a parsimonious model. Second, the algorithm includes a post-model fitting diagnostic step where the residuals from the MBSTS step are processed through a multi-input/output temporal convolutional network (M-TCN) with multiple time scale feature learning. This step serves as an alternative to traditional subjective residual-based diagnostic procedures in time-series analysis, with the aim of improving forecasting accuracy. The key advantage of the M-TCN is its ability to capture sequential information efficiently. The M-TCN expands the field of convolution kernel without increasing the number of parameters, thus enhancing the capacity of model to capture complex sequential patterns. The paper presents two applications showcasing the effectiveness of the proposed hybrid algorithm. First, it utilizes pre-lockdown data from eleven Nifty stock sectoral indices to predict stock price movements, including the initial post-lockdown upturn. In the second application, it focuses on stock market data from pharmaceutical companies involved in manufacturing COVID-19 vaccines. In both cases, sentiment data sourced from newspapers and social media serve as the regression component. Through rigorous analysis, the paper demonstrates that the hybrid model outperforms various benchmark models, including LSTM, Bidirectional Encoder Representations from Transformers (BERT)-based LSTM, Deep Transformer Model, and GRU, among others, in terms of forecasting accuracy. This underscores the utility of the hybrid algorithm, particularly in predicting stock market trends during the COVID-19 pandemic period and its associated market dynamics.

Relationships of State Alcohol Policy Environments With Homicides and Suicides

IntroductionAlcohol use is involved in a large proportion of homicides and suicides each year in the United States, but there is limited evidence on how policies targeting alcohol influence violence in the U.S. context. Extant studies generally focus on individual policies in isolation of each other. This study examines of the effects of changes in states’ alcohol policy restrictions on overall homicide and suicide rates and firearm-related homicide and suicide rates using a holistic measure of states’ alcohol policy environments. MethodsUsing a composite measure of state-level alcohol policies (Alcohol Policy Scale) and data from the National Vital Statistics System from 2002 to 2018, this study applied a Bayesian time series model to estimate the effects of alcohol policy changes on overall and firearm-involved homicide and suicide rates. The analysis was performed in 2023 and 2024. ResultsA one standard deviation change in the Alcohol Policy Scale was associated with a 6 percent decline in homicide rates both overall (IRR=0.94; 95-percent credibility interval = [0.89, 1.00]) and for firearm homicides specifically (IRR=0.94, 95-percent CI=[0.88, 1.01]). There was no clear association of alcohol policy with suicides. The model predicts that a nationwide increase in alcohol restrictions equivalent to a shift from the 25th to 75th percentile of the scale's distribution would result in almost 1200 fewer homicides annually. ConclusionsIncreases in the restrictiveness of state-level alcohol policies are associated with reductions in homicides. More restrictive alcohol policy environments may offer an opportunity to reduce homicides.

Multilayer factors associated with excess all-cause mortality during the omicron and non-omicron waves of the COVID-19 pandemic: time series analysis in 29 countries.

The COVID-19 pandemic has resulted in significant excess mortality globally. However, the differences in excess mortality between the Omicron and non-Omicron waves, as well as the contribution of local epidemiological characteristics, population immunity, and social factors to excess mortality, remain poorly understood. This study aims to solve the above problems. Weekly all-cause death data and covariates from 29 countries for the period 2015-2022 were collected and used. The Bayesian Structured Time Series Model predicted expected weekly deaths, stratified by gender and age groups for the period 2020-2022. The quantile-based g-computation approach accounted for the effects of factors on the excess all-cause mortality rate. Sensitivity analyses were conducted using alternative Omicron proportion thresholds. From the first week of 2021 to the 30th week of 2022, the estimated cumulative number of excess deaths due to COVID-19 globally was nearly 1.39million. The estimated weekly excess all-cause mortality rate in the 29 countries was approximately 2.17 per 100,000 (95% CI: 1.47 to 2.86). Weekly all-cause excess mortality rates were significantly higher in both male and female groups and all age groups during the non-Omicron wave, except for those younger than 15 years (P < 0.001). Sensitivity analysis confirmed the stability of the results. Positive associations with all-cause excess mortality were found for the constituent ratio of non-Omicron in all variants, new cases per million, positive rate, cardiovascular death rate, people fully vaccinated per hundred, extreme poverty, hospital patients per million humans, people vaccinated per hundred, and stringency index. Conversely, other factors demonstrated negative associations with all-cause excess mortality from the first week of 2021 to the 30th week of 2022. Our findings indicate that the COVID-19 Omicron wave was associated with lower excess mortality compared to the non-Omicron wave. This study's analysis of the factors influencing excess deaths suggests that effective strategies to mitigate all-cause mortality include improving economic conditions, promoting widespread vaccination, and enhancing overall population health. Implementing these measures could significantly reduce the burden of COVID-19, facilitate coexistence with the virus, and potentially contribute to its elimination.

Abstract TP93: A Bayesian Hierarchical Time Series Model to Forecast Patient Recovery Outcomes After an Acute Stroke When Activity Monitor Data is Incomplete

Background: Activity monitor data is a useful tool in clinical trials to monitor patient recovery after stroke. These data offer important information on activity, sedentary and sleep behaviors throughout recovery. However, not all participants wear activity monitors, or their activity data may be incomplete. The purpose of this study was to develop a model-based approach to obtain individual patient-level longitudinal forecast estimates of activity and sleep behaviors for both those with and without complete activity data. Methods: Participants wore an ActivPal (steps/day) and Actigraph (total sleep/day) monitors for 3-7 days at 10-60-90 days post stroke. We developed a Bayesian hierarchical time series model to forecast daily steps and sleep time by participant. We included a logistic growth curve in which we use predictors comprising age, sex, race, marital status, smoking, BMI, cognition, gait speed, depression, self-reported sleep, and patients’ previous time trends to forecast patient specific future time trends during the first 90 days following stroke. Results: Data were collected on a total sample of 114 participants. Out of the 114 participants, 73 participants had recorded sleep data, and 76 participants had step data reported, which are not mutually exclusive groups. Out of 114 total participants, 41 participants did not have any sleep or step data recorded. We obtained patient-specific time trends for both patients with and without monitor data. To assess model predictive performance, we compared predicted estimates to observed data using a validation study and found that on average, there was a difference of 0.52 daily steps (counts). Additionally, on average, there was a difference of 0.46 between observed and estimated daily sleep count (minutes). Discussion: The contribution of our study is a model-based approach to accurately predict patient-specific longitudinal step and sleep behaviors based on predictor variables and previous time trends for patients in the subacute stroke recovery phase. Our approach may be extended to other applications using activity monitor data. A limitation of this study is the increased uncertainty surrounding forecast estimates for periods/days without data, which will be addressed in future work.

The impact of EV71 vaccination program on hand, foot and mouth disease in Zhejiang Province, China: A negative control study

ObjectiveTo estimate the potential causal impact of Enterovirus A71 (EV71) vaccination program on the reduction of EV71-infected hand, foot, and mouth disease (HFMD) in Zhejiang Province. MethodsWe utilized the longitudinal surveillance dataset of HFMD and EV71 vaccination in Zhejiang Province during 2010–2019. We estimated vaccine efficacy using a Bayesian structured time series (BSTS) model, and employed a negative control outcome (NCO) model to detect unmeasured confounding and reveal potential causal association. ResultsWe estimated that 20,132 EV71 cases (95% CI: 16,733, 23,532) were prevented by vaccination program during 2017–2019, corresponding to a reduction of 29% (95% CI: 24%, 34%). The effectiveness of vaccination increased annually, with reductions of 11% (95% CI: 6%, 16%) in 2017 and 66% (95% CI: 61%, 71%) in 2019. Children under 5 years old obtained greater benefits compared to those over 5 years. Cities with higher vaccination coverage experienced a sharper EV71 reduction compared to those with lower coverage. The NCO model detected no confounding factors in the association between vaccination and EV71 cases reduction. ConclusionsThis study suggested a potential causal effect of the EV71 vaccination, highlighting the importance of achieving higher vaccine coverage to control the HFMD.

Bayesian Time Series Modeling of Necrotizing Fasciitis Count in Mahasarakham and Roi-Et Hospitals

Bayesian Time Series Modeling of Necrotizing Fasciitis Count in Mahasarakham and Roi-Et Hospitals

The “war on cops,” retaliatory violence, and the murder of George Floyd*

AbstractThe police murder of George Floyd sparked nationwide protests in the summer of 2020 and revived claims that public outcry over such high‐profile police killings perpetuated a violent “war on cops.” Using data collected by the Gun Violence Archive (GVA) on firearm assaults of U.S. police officers, we use Bayesian structural time series (BSTS) modeling to empirically assess if and how patterns of firearm assault on police officers in the United States were influenced by the police murder of George Floyd. Our analysis finds that the murder of George Floyd was associated with a 3‐week spike in firearm assaults on police, after which the trend in firearms assaults dropped to levels only slightly above that which were predicted by pre‐Floyd data. We discuss potential explanations for these findings and consider their relevance to the contemporary discussion of a “war on cops,” violence, and officer safety.

WITHDRAWN: The Impact of Industrial Structure Evolution on Fiscal Revenue in Shenzhen China

Adjustment, optimization and upgrading of industrial structure is one of the important indicators that affect local fiscal revenue. To explore the relationship between industrial structure and local fiscal revenue is the basis for industrial policy formulation and fiscal expenditure arrangement and an important issue in the field of structural economics. This paper aims at exploring the inherent regularity public fiscal revenue and the upgrading of industrial structure in the process of urban economic development. We utilize data from 1996 to 2020 in the “Shenzhen Statistical Yearbook” and employ Bayesian structure time series (BSTS) model to analyze the effect of industrial structure evolution on fiscal revenue in Shenzhen. The tax revenue data is used to represent Shenzhen's fiscal revenue and the ratio of the tertiary to secondary industry added values is used to represent industrial structure evolution index. Our findings demonstrate that the impact of Shenzhen's industrial structure evolution on fiscal revenue shows a steady and positive trend from 2000 to 2004. Then it significantly increases to a peak in 2006 and goes downward. After 2017, the evolution of Shenzhen's industrial structure has negative impact on the overall fiscal revenue. Based on the results, this paper puts forward relevant policy implications for other cities in China to promote the transformation and upgrading of industrial structures.

Real-time estimation and forecasting of COVID-19 cases and hospitalizations in Wisconsin HERC regions for public health decision making processes

BackgroundThe spread of the COVID-19 (SARS-CoV-2) and the surging number of cases across the United States have resulted in full hospitals and exhausted health care workers. Limited availability and questionable reliability of the data make outbreak prediction and resource planning difficult. Any estimates or forecasts are subject to high uncertainty and low accuracy to measure such components. The aim of this study is to apply, automate, and assess a Bayesian time series model for the real-time estimation and forecasting of COVID-19 cases and number of hospitalizations in Wisconsin healthcare emergency readiness coalition (HERC) regions.MethodsThis study makes use of the publicly available Wisconsin COVID-19 historical data by county. Cases and effective time-varying reproduction number R_t by the HERC region over time are estimated using Bayesian latent variable models. Hospitalizations are estimated by the HERC region over time using a Bayesian regression model. Cases, effective Rt, and hospitalizations are forecasted over a 1-day, 3-day, and 7-day time horizon using the last 28 days of data, and the 20%, 50%, and 90% Bayesian credible intervals of the forecasts are calculated. The frequentist coverage probability is compared to the Bayesian credible level to evaluate performance.ResultsFor cases and effective R_t, all three time horizons outperform the three credible levels of the forecast. For hospitalizations, all three time horizons outperform the 20% and 50% credible intervals of the forecast. On the contrary, the 1-day and 3-day periods underperform the 90% credible intervals. Questions about uncertainty quantification should be re-calculated using the frequentist coverage probability of the Bayesian credible interval based on observed data for all three metrics.ConclusionsWe present an approach to automate the real-time estimation and forecasting of cases and hospitalizations and corresponding uncertainty using publicly available data. The models were able to infer short-term trends consistent with reported values at the HERC region level. Additionally, the models were able to accurately forecast and estimate the uncertainty of the measurements. This study can help identify the most affected regions and major outbreaks in the near future. The workflow can be adapted to other geographic regions, states, and even countries where decision-making processes are supported in real-time by the proposed modeling system.

The relation between wheat, soybean, and hemp acreage: a Bayesian time series analysis

The 2018 United States Farm Bill has opened the possibility for farmers to increase their profits through hemp cultivation. The literature suggests hemp has the potential to replace soybeans in soybean–wheat double-cropping because hemp shares key attributes of soybeans as a rotation crop (profitability, potential as an energy crop, and maintenance of soil fertility). Nonetheless, due to a short history of hemp cultivation in the USA, it is difficult to predict a time series relationship between hemp, soybean, and wheat through conventional approaches. In this article, we use Bayesian time series models and data from Statistics Canada and the Alberta Agricultural and Rural Development Department to examine a time series relationship between hemp, wheat, and soybean acreage and therefore predict farmers’ decision when hemp is a legal alternative agricultural commodity. Our results show evidence of complementary and substitution relationships for hemp–wheat and hemp–soybean, respectively. In addition, the results indicate a potential of hemp monoculture as a positive response to self-positive shock on hemp acreage that lasts for years.

A Bayesian time series model for reconstructing hydroclimate from multiple proxies

AbstractWe propose a Bayesian model which produces probabilistic reconstructions of hydroclimatic variability in Queensland Australia. The model provides a standardized approach to hydroclimate reconstruction using multiple palaeoclimate proxy records derived from natural archives such as speleothems, ice cores and tree rings. The method combines time‐series modeling with inverse prediction to quantify the relationships between a given hydroclimate index and relevant proxies over an instrumental period and subsequently reconstruct the hydroclimate back through time. We present case studies for Brisbane and Fitzroy catchments focusing on two hydroclimate indices, the Rainfall Index (RFI) and the Standardized Precipitation‐Evapotranspiration Index (SPEI). The probabilistic nature of the reconstructions allows us to estimate the probability that a hydroclimate index in any reconstruction year was lower (higher) than the minimum (maximum) value observed over the instrumental period. In Brisbane, the RFI is unlikely (probabilities &lt; 5%) to have exhibited extremes beyond the minimum/maximum values observed between 1889 and 2019. However, in Fitzroy there are several years during the reconstruction period where the RFI is likely (&gt;50% probability) to have exhibited behavior beyond the minimum/maximum of what has been observed, during the instrumental period. For SPEI, the probability of observing such extremes prior to the beginning of the instrumental period in 1889 doesn't exceed 30% in any reconstruction year in Brisbane, but exceeds 50% in multiple years in Fitzroy.

Evaluating federal policies using Bayesian time series models: estimating the causal impact of the hospital readmissions reduction program

Researchers are often faced with evaluating the effect of a policy or program that was simultaneously initiated across an entire population of units at a single point in time, and its effects over the targeted population can manifest at any time period afterwards. In the presence of data measured over time, Bayesian time series models have been used to impute what would have happened after the policy was initiated, had the policy not taken place, in order to estimate causal effects. However, the considerations regarding the definition of the target estimands, the underlying assumptions, the plausibility of such assumptions, and the choice of an appropriate model have not been thoroughly investigated. In this paper, we establish useful estimands for the evaluation of large-scale policies. We discuss that imputation of missing potential outcomes relies on an assumption which, even though untestable, can be partially evaluated using observed data. We illustrate an approach to evaluate this key causal assumption and facilitate model elicitation based on data from the time interval before policy initiation and using classic statistical techniques. As an illustration, we study the Hospital Readmissions Reduction Program (HRRP), a US federal intervention aiming to improve health outcomes for patients with pneumonia, acute myocardial infraction, or congestive failure admitted to a hospital. We evaluate the effect of the HRRP on population mortality among the elderly across the US and in four geographic subregions, and at different time windows. We find that the HRRP increased mortality from pneumonia and acute myocardial infraction across at least one geographical region and time horizon, and is likely to have had a detrimental effect on public health.

Levels and trends estimate of sex ratio at birth for seven provinces of Pakistan from 1980 to 2020 with scenario-based probabilistic projections of missing female birth to 2050: A Bayesian modeling approach

Most evidence on son preference in Pakistan is reflected in the higher child mortality among females than males. The sex discrimination before birth is rarely reported in Pakistan. This is the first study to quantify prenatal sex discrimination in Pakistan on a subnational level. We provide annual estimates of the sex ratio at birth (SRB) from 1980 to 2020 and scenario-based projections of the number of missing female births up to 2050 by Pakistan province. The results are based on a comprehensive database consisting of 832,091 birth records from all available surveys and censuses. We adopted a Bayesian hierarchical time series model to synthesize different data sources. We identified Balochistan with an existing imbalanced SRB since 1980. For the rest provinces without past or ongoing SRB inflation, we projected the largest female birth deficit to occur in Punjab in 2033 under the scenario that the SRB transition process starts in 2021. We demonstrated important disparities in the occurrence and quantification of missing female births up to 2050.

Disparities in the Effects of Acuity Circle-based Liver Allocation on Waitlist and Transplant Practice Between Centers.

Background.Liver allocation in the United States was updated on February 4, 2020, by introducing the acuity circle (AC)–based model. This study evaluated the early effects of the AC-based allocation on waitlist outcomes.Methods.Adult liver transplant (LT) candidates listed between January 1, 2019, and September 30, 2021, were assessed. Two periods were defined according to listing date (pre- and post-AC), and 90-d waitlist outcomes were compared. Median transplant Model for End-stage Liver Disease (MELD) score of each transplant center was calculated, with centers categorized as low- (<25 percentile), mid- (25–75 percentile), and high-MELD (>75 percentile) centers.Results.A total of 12 421 and 17 078 LT candidates in the pre- and post-AC eras were identified. Overall, the post-AC era was associated with higher cause-specific 90-d hazards of transplant (csHR, 1.32; 95% confidence interval [CI], 1.27-1.38; P < 0.001) and waitlist mortality (cause-specific hazard ratio [csHR], 1.20; 95% CI, 1.09-1.32; P < 0.001). The latter effect was primarily driven by high-MELD centers. Low-MELD centers had a higher proportion of donations after circulatory death (DCDs) used. Compared with low-MELD centers, mid-MELD and high-MELD centers had significantly lower cause-specific hazards of DCD-LT in both eras (mid-MELD: csHR, 0.47; 95% CI, 0.38-0.59 in pre-AC and csHR, 0.56; 95% CI, 0.46-0.67 in post-AC and high-MELD: csHR, 0.11; 95% CI, 0.07-0.17 in pre-AC and csHR, 0.14; 95% CI, 0.10-0.20 in post-AC; all P < 0.001). Using a structural Bayesian time-series model, the AC policy was associated with an increase in the actual monthly DCD-LTs in low-, mid-, and high-MELD centers (actual/predicted: low-MELD: 19/16; mid-MELD: 21/14; high-MELD: 4/3), whereas the increase in monthly donation after brain death–LTs were only present in mid- and high-MELD centers.Conclusions.Although AC-based allocation may improve waitlist outcomes, regional variation exists in the drivers of such outcomes between centers.

Using Bayesian time series models to estimate bycatch of an endangered albatross

Developing unbiased estimates of incidental bycatch poses a challenge for species where fishing-induced mortality is a rare occurrence. Expanding rare mortality events using ratio estimators or bycatch of proxy species can result in highly variable estimates based on untested and often untestable assumptions. We estimated short-tailed albatross bycatch in a U.S. West Coast groundfish fishery using Bayesian time series modeling. The best model used a constant bycatch rate and inferred annual expected bycatch and variability using a Poisson distribution, given specified levels of observed effort. Fleet-wide bycatch estimates varied annually and peaked at 1.35 birds in 2011 (the year of the only observed mortality). The probability of exceeding the limit of five estimated takes in a 2-year period was very low throughout the time series, and estimated takes in the unobserved portion of the fleet are more likely with lower observer coverage and higher fishing effort. The Bayesian model-based approach avoids assumptions inherent in ratio estimators and proxy methods; it incorporates uncertainty, reduces volatility, and enables comparisons of bycatch estimates to management thresholds. This analytical approach offers natural resource managers a framework for estimating bycatch in data-limited contexts, which can result in better guidance for management actions and mitigation strategies.

Neighborhood, built environment and resilience in transportation during the COVID-19 pandemic.

COVID-19 has swept the world, and the unprecedented decline in transit ridership has been noticed. However, little attention has been paid to the resilience of the transportation system, particularly in medium-sized cities. Drawing upon a light rail ridership dataset in Salt Lake County from 2017 to 2021, we develop a novel method to measure the vulnerability and resilience of transit ridership using a Bayesian structure time series model. The results show that government policies have a more significant impact than the number of COVID-19 cases on transit ridership. Regarding the built environment, a highly compact urban design might reduce the building coverage ratio and makes transit stations more vulnerable and less resilient. Furthermore, the high rate of minorities is the primary reason for the drops in transit ridership. The findings are valuable for understanding the vulnerability and resilience of transit ridership to pandemics for better coping strategies in the future.

The impact of enterovirus A71 vaccination program on hand, foot, and mouth disease in Guangdong, China: A longitudinal surveillance study

Enterovirus A71 (EV71) vaccination program was introduced in 2016 in China. Based on a longitudinal surveillance dataset from 2012 to 2019 in Guangdong, China, we estimated the impact of the EV71 vaccination program on hand, foot, and mouth disease (HFMD) incidence, by using a counterfactual prediction made from synthetic control approach integrated with a Bayesian time-series model. We observed a relative reduction of 41.4% for EV71-associated HFMD cases during the post-vaccination period of 2017–2019, corresponding to 26,226 cases averted. The reduction of EV71-associated HFMD cases raised with the elevation of EV71 vaccine coverage by year. We found an indirect effect for the children aged 6–14 years who were less likely to be vaccinated. Whereas, the EV71 vaccine may not protect against non-EV71-associated HFMD. This study provides a template for ongoing public health surveillance of EV71 vaccine effectiveness with a counterfactual study design. Our results show strong evidence of the EV71 vaccination program working on reducing EV71-associated HFMD in real-world settings. The finding will benefit policy-making of EV71 vaccination and the prevention of HFMD.

Military Health System Opioid, Tramadol, and Gabapentinoid Prescription Volumes Before and After a Defense Health Agency Policy Release

Clinical practice guidelines (CPGs) and health system policies to mitigate inappropriate opioid prescribing practices may have an extended impact on low-dose opioid (e.g., tramadol) and non-opioid (e.g., gabapentinoid) pain medication prescribing practices. To evaluate changes in opioid, tramadol, and gabapentinoid prescribing rates from January 2016 to February 2020 within the Military Health System, including the degree to which prescribing rates changed after release of a US Defense Health Agency Procedural Instruction. In this observational health services research study, opioid, tramadol, and gabapentin prescription dispense events of US Military Health System beneficiaries enrolled in care at military treatment facilities prior to US Defense Health Agency Procedural Instruction release (January 2016-May 2018) were used to forecast values from the post-intervention period (June 2018-February 2020). The median opioid and tramadol prescribing rates decreased from January 2016 to February 2020, aside from tramadol prescribing in Surgery Clinics, which increased. Gabapentinoid prescribing rate changes were mixed. In Bayesian time series models, the forecasted proportion of patients receiving each of the three medications, regardless of age group or clinic type, did not significantly vary from the actual prescribing rates in the post-intervention period. Overall, CPGs and policies targeting opioid prescribing practices may have provided the maximal impetus for providers to re-evaluate their prescribing practices, as the policy did not appear to change the slope in prescribing rates. However, it is unclear whether the policies mitigated the likelihood of plateaus in prescribing rates. Further work is needed to assess the degree to which providers simultaneously altered other non-opioid pain medication prescribing practices, self-management recommendations, and non-pharmacological therapy referrals.