Hierarchical Bayesian Analysis Research Articles

Identifying Spatiotemporal Clusters by Means of Agglomerative Hierarchical Clustering and Bayesian Regression Analysis with Spatiotemporally Varying Coefficients: Methodology and Application to Dengue Disease in Bandung, Indonesia

Dengue disease has serious health and socioeconomic consequences. Preventive actions are needed to avoid outbreaks. Bayesian spatiotemporal models with conditional autoregressive (CAR) and random walk (RW) priors are two common smoothing approaches used in disease mapping to develop early warning systems and action plans. However, this approach can lead to over‐smoothing, such that discontinuities in the risk surface are “averaged” out. Moreover, local variation in the relationship between disease risk and spatiotemporal risk factors can be concealed by the assumption of homogeneous regression effects. To overcome these problems, we propose a two‐stage spatiotemporal clustering approach. In the first stage, an agglomerative hierarchical clustering algorithm is applied to map out the complete set of cluster configurations. In the second stage, the optimal cluster configuration is selected using a spatiotemporal model with spatiotemporally varying coefficients. The methodology is applied to data for 30 districts in the city of Bandung, Indonesia, for the period January 2012 to December 2016. Weather factors are used as the risk factors. High‐risk and low‐risk clusters are identified that strongly vary across space and time, with sunshine duration, precipitation, evaporation, and humidity as the most important risk factors.

A long-term decline in the abundance of endangered leatherback turtles, Dermochelys coriacea, at a foraging ground in the California Current Ecosystem

Pacific leatherback turtles (Dermochelys coriacea) are critically endangered, and declines have been documented at multiple nesting sites throughout the Pacific. The western Pacific leatherback forages in temperate and tropical waters of the Indo-Pacific region, and about 38–57% of summer-nesting females from the largest remaining nesting population in Papua Barat (Indonesia) migrate to distant foraging grounds off the U.S. West Coast, including neritic waters off central California. In this study, we examined the trend in leatherback abundance off central California from 28 years of aerial survey data from coast-wide and adaptive fine-scale surveys. We used a Bayesian hierarchical analysis framework, including a process model of leatherback population density and an observation model relating leatherback observations to distance sampling methods. We also used time-depth data from biologgers deployed on 21 foraging leatherback turtles in the study area to account for detection biases associated with diving animals. Our results indicate that leatherback abundance has declined at an annual rate of −5.6% (95% credible interval −9.8% to −1.5%), without any marked changes in ocean conditions or prey availability. These results are similar to the nesting population trends of −5.9% and −6.1% per year estimated at Indonesian index beaches, which comprise 75% of western Pacific nesting activity. Combined, the declining trends underscore the need for coordinated international conservation efforts and long-term population monitoring to avoid extirpation of western Pacific leatherback turtles.

Anxiety Modulates Preference for Immediate Rewards Among Trait-Impulsive Individuals: A Hierarchical Bayesian Analysis

Trait impulsivity—defined by strong preference for immediate over delayed rewards and difficulties inhibiting prepotent behaviors—is observed in all externalizing disorders, including substance-use disorders. Many laboratory tasks have been developed to identify decision-making mechanisms and correlates of impulsive behavior, but convergence between task measures and self-reports of impulsivity are consistently low. Long-standing theories of personality and decision-making predict that neurally mediated individual differences in sensitivity to (a) reward cues and (b) punishment cues (frustrative nonreward) interact to affect behavior. Such interactions obscure one-to-one correspondences between single personality traits and task performance. We used hierarchical Bayesian analysis in three samples with differing levels of substance use ( N = 967) to identify interactive dependencies between trait impulsivity and state anxiety on impulsive decision-making. Our findings reveal how anxiety modulates impulsive decision-making and demonstrate benefits of hierarchical Bayesian analysis over traditional approaches for testing theories of psychopathology spanning levels of analysis.

Discriminating between Neutron Stars and Black Holes with Imperfect Knowledge of the Maximum Neutron Star Mass

Abstract Although gravitational-wave signals from exceptional low-mass compact binary coalescences, like GW170817, may carry matter signatures that differentiate the source from a binary black hole system, only one out of every eight events detected by the current Advanced LIGO and Virgo observatories are likely to have signal-to-noise ratios large enough to measure matter effects, even if they are present. Nonetheless, the systems’ component masses will generally be constrained precisely. Constructing an explicit mixture model for the total rate density of merging compact objects, we develop a hierarchical Bayesian analysis to classify gravitational-wave sources according to the posterior odds that their component masses are drawn from different subpopulations. Accounting for current uncertainty in the maximum neutron star mass and adopting a power-law mass distribution with or without a mass gap and either random or mass-ratio-dependent pairing, we examine two recent events from the LIGO–Virgo Collaboration’s third observing run, GW190425 and GW190814. For population models with no overlap between the neutron star and black hole mass distributions, we typically find that there is a chance, depending on the exact population model, that GW190425 was a binary neutron star merger rather than a neutron-star–black-hole merger. On the other hand, we find that there is a chance that GW190814 involved a slowly spinning neutron star, regardless of our assumed population model.

Immediate <i>Versus</i> Delayed Stenting in Patients with Acute ST-Elevation Myocardial Infarction: A Randomized Open-Label Bayesian Trial

Background: Stents immediately implanted in thrombus-laden arteries cause embolization, which paradoxically impairs myocardial reperfusion and function. Methods: We conducted a Bayesian open-label randomized trial across 15 centres in France and Canada. Patients with acute ST-elevation myocardial infarction (STEMI) and reperfused by thrombectomy or small-size balloon angioplasty were randomly assigned to immediate stenting or to a bridging anti-thrombin therapy and delayed stenting between 18h to 48h after reperfusion. Using a hierarchical Bayesian analysis, we borrowed data from exchangeable twin trials to form a posterior probability of the risk ratio (RR) for the combined occurrence of cardiovascular death, heart failure, non-fatal myocardial infarction (MI), or target-vessel revascularisation (TVR) over 9 months in the intent-to-treat populations. Findings: Between April 2014 and September 2017, 305 participants were randomly assigned to delayed (n = 152) or immediate stenting (n = 153). In parallel, four similar RCTs were reported that satisfied exchangeability requirements, leading to 1,873 participants, including 930 assigned to delayed and 943 assigned to immediate stenting. The primary endpoint occurred in 167 participants (18.1%) assigned to delayed stenting compared to 194 participants (20.6%) assigned to immediate stenting (RR, 0.91; 95% credible interval [95%CrI], 0.76 to 1.09; probability of superiority, 89%). Delayed stenting led to a meaningful increase in unplanned target revascularization acutely (RR, 1.54; 95% CrI, 0.98 to 2.41, probability of inferiority, 99%). At long term, delayed stenting reduced the combined occurrence of CV death or heart failure (RR, 0.83; 95% CrI, 0.67 to 1.02, probability of superiority, 99%). Interpretation: In patients with STEMI, delayed stenting did not reduce the occurrence of cardiovascular death, non-fatal MI, heart failure and unplanned TVR compared to the immediate stenting. There is a high probability that delayed stenting substantially reduces cardiovascular death and heart failure in the longer-term. Trial Registration: ClinicalTrials.gov identifier: NCT01542385. Funding Statement: The trial was supported by an operating grant from the peer-reviewed, publically funded Canadian Institute of Health Research (CIHR), through the Industry-Partnered Collaborative Research program (grant #298937), in collaboration with Boston Scientific and AstraZeneca (unrestricted grants support). The industry partners played no role in the conduct and oversight of the study. Declaration of Interests: The investigators disclose the following relationships: E. Marc Jolicoeur: Advisory Board: Institut National d’Excellence en Sante et Services Sociaux du Quebec, Servier, Imbria Pharmaceutical, Carre technology; Board of Directors, Societe de Sciences Vasculaires du Quebec; Data Safety Monitoring Board: XyloCor; Research grants; Boston Scientific, AstraZeneca; The University of Glasgow (employer, C Berry) holds research and/or consultancy agreements with AstraZeneca, Abbott Vascular, Boehringer Ingelheim, GSK, HeartFlow, Novartis and Siemens Healthcare; Francois Roubille: Advisory Board: Abbott, Air liquide, Research grants; Boston Scientific, AstraZeneca. Consultancy agreements with AstraZeneca, Abbott Vascular, Air liquide, Amgen, Boehringer Ingelheim, GSK, MSD, Novartis, Sanofi, Novartis, Novonordisk, Vifor; Geraud Souteyrand: Consultancy agreements; Medtronic, Abbott, and Terumo; Guillaume Cayla: Consultancy agreements; Astra Zeneca, Amgen, Abbot Vascular, BMS, Boston, Biotronik, Medtronic, Pfizer, and Sanofi; Nicolas Delarche: Advisory board: Novartis; Samer Mansour: Consultant/Advisory board/Speaker’s Bureau: Abbott vascular, Soundbite, Boehringer Ingelheim, AstraZeneca, Bayer, BMS-Pfizer Alliance, Medtronic, Amgen, Sanofi, Servier, Gilead and Novartis; Erick Schampaert: Consultant/Advisory board/Speaker’s Bureau: Abbott, AstraZeneca, Bayer, BMS-Pfizer Alliance, Medtronic, Philips-Volcano, Servier; Loic Belle: Research Grant: AstraZeneca, Medtronic, Biotronic, Boston Scientific, Abbott Vascular, Terumo Ethics Approval Statement: The protocol complied with the Declaration of Helsinki and was approved by IRB at each participating centers. The trial was coordinated by the Montreal Health Innovation Coordinating Centre under the direction of the executive committee, which was responsible for trial conduct, the integrity of the data analysis, and the reporting of results.

NSAIDs for Prophylaxis for Heterotopic Ossification After Total Hip Arthroplasty: A Bayesian Network Meta-analysis

Non-steroidal anti-inflammatory drugs (NSAID) have been recommended to prevent of heterotopic ossification (HO) after total hip arthroplasty (THA), but debates are still ongoing. The present Bayesian network meta-analysis of randomized clinical trials (RCTs) compared all available pathways of NSAID treatment as prophylaxis for HO after THA. The present Bayesian network meta-analysis was conducted according to The PRISMA Extension Statement for Reporting of Systematic Reviews Incorporating Network Meta-analyses of Health Care Interventions guidelines. All randomized clinical trials comparing two or more interventions to prevent HO after THA were considered for analysis. HO was classified according to Brooker. The quality of the methodological assessment was performed through the risk of bias summary tool of the Review Manager Software 5.3 (The Cochrane Collaboration, Copenhagen). The network meta-analysis was performed through a STATA routine for Bayesian hierarchical random-effects model analysis, with log odd ratio (LOR) effect measure. Data from 26 studies (6396 THAs; 58% females) were collected. The mean follow-up was 10.50 ± 5.7 months. ANOVA showed good comparability among mean age and gender (P > 0.5). Celecoxib demonstrated the highest rate of Brooker class 0 (LOR 6.96), followed by diclofenac (LOR 6.94). Naproxen demonstrated the lowest rate of Brooker I HO (LOR 2.82), followed by celecoxib (LOR 3.52). Celecoxib demonstrated lowest rate of Brooker class II HO (LOR 1.66), class III (LOR), and class IV (LOR 0.25). The equation for global linearity detected no statistically significant inconsistency (P > 0.5) in all the comparisons. The present Bayesian network meta-analysis encourages the use of celecoxib as a prophylaxis for HO. These conclusions must be interpreted in light of the limitations of the present study. Future investigations are required to establish more definitely the role of celecoxib.Level of Evidence: I, Bayesian network analysis of RCTs.

Evolution of Angiosperm Pollen: 8. Lamiids

Lamiids, a clade composed of approximately 15% of all flowering plants, contains more than 50,000 species dispersed across 49 families and eight orders (APG IV, 2016). This paper is the eighth in a series that analyzes pollen characters across angiosperms. We reconstructed a maximum likelihood tree based on the most recent phylogenetic studies for the Lamiids, comprising 150 terminal genera (including six outgroups) and covering all eight orders and 49 families within the clade. To illustrate pollen diversity across the Lamiids, pollen grains from 22 species (22 genera in 14 families) were imaged under light, scanning, and transmission electron microscopy. Eighteen pollen characters that were documented from previous publications, websites, and our new observations were coded and optimized onto the reconstructed phylogenetic tree using Fitch parsimony, maximum likelihood, and hierarchical Bayesian analysis. Pollen morphology of the Lamiids is highly diverse, particularly in shape class, pollen size, aperture number, endoaperture shape, supratectal element shape, and tectum sculpture. In addition, some genera show relatively high infrageneric pollen variation within the Lamiids: i.e., Coffea L., Jacquemontia Choisy, Justicia L., Pedicularis L., Psychotria L. nom. cons., Sesamum L., Stachytarpheta Vahl, and Veronica L. The plesiomorphic states for 16 pollen characters were inferred unambiguously, and 10 of them displayed consistent plesiomorphic states under all optimization methods. Seventy-one lineages at or above the family level are characterized by pollen character state transitions. We identified diagnostic character states for monophyletic clades and explored palynological evidence to shed light on unresolved relationships. For example, palynological evidence supports the monophyly of Garryales and Metteniusaceae, and sister relationships between Icacinaceae and Oncothecaceae, as well as between Vahliales and Solanales. The evolutionary patterns of pollen morphology found in this study reconfirm several previously postulated evolutionary trends, which include an increase in aperture number, a transition from equatorially arranged apertures to globally distributed ones, and an increase in exine ornamentation complexity. Furthermore, there is a significant correlation between pollen characters and a number of ecological factors, e.g., pollen size and pollination type, pollen ornamentation and pollination type, and shape class and plant growth form. Our results provide insight into the ecological, environmental, and evolutionary mechanisms driving pollen character state changes in the Lamiids.

Choice data generation using usage scenarios and discounted cash flow analysis

Discrete choice analysis is a popular method of estimating heterogeneous customer preferences. Although model accuracy can be increased by including more choice data, this option is untenable when the obtaining of choice data from target customers is costly and time-consuming.. We thus propose a method for choice data generation for commercial products whose expected money value is a key factor in consumer choice (e.g., commercial vehicles and financial product). Using an individual usage scenario, we generate a discounted cash flow (DCF) model instead of a utility model to estimate the discount rates, rather than partworths, of individual consumers. The DCF model helps us generate synthetic choice data from choice sets consisting of various combinations of attribute levels. Using these data, we employ a hierarchical Bayesian (HB) discrete choice analysis. We conclude the study with a case study on the preference estimation of a hybrid courier truck conversion. The results reveal that the DCF-based HB estimation outperforms the traditional HB estimation.

θ-γ Cross-Frequency Transcranial Alternating Current Stimulation over the Trough Impairs Cognitive Control.

Cognitive control is a mental process, which underlies adaptive goal-directed decisions. Previous studies have linked cognitive control to electrophysiological fluctuations in the θ band and θ-γ cross-frequency coupling (CFC) arising from the cingulate and frontal cortices. However, to date, the behavioral consequences of different forms of θ-γ CFC remain elusive. Here, we studied the behavioral effects of the θ-γ CFC via transcranial alternating current stimulation (tACS) designed to stimulate the frontal and cingulate cortices in humans. Using a double-blind, randomized, repeated measures study design, 24 healthy participants were subjected to three active and one control CFC-tACS conditions. In the active conditions, 80-Hz γ tACS was coupled to 4-Hz θ tACS. Specifically, in two of the active conditions, short γ bursts were coupled to the delivered θ cycle to coincide with either its peaks or troughs. In the third active condition, the phase of a θ cycle modulated the amplitude of the γ oscillation. In the fourth, control protocol, 80-Hz tACS was continuously superimposed over the 4-Hz tACS, therefore lacking any phase specificity in the CFC. During the 20 min of stimulation, the participants performed a Go/NoGo monetary reward-based and punishment-based instrumental learning task. A Bayesian hierarchical logistic regression analysis revealed that relative to the control, the peak-coupled tACS had no effects on the behavioral performance, whereas the trough-coupled tACS and, to a lesser extent, amplitude-modulated tACS reduced performance in conflicting trials. Our results suggest that cognitive control depends on the phase specificity of the θ-γ CFC.

Imprecise action selection in substance use disorder: Evidence for active learning impairments when solving the explore-exploit dilemma

BackgroundSubstance use disorders (SUDs) are a major public health risk. However, mechanisms accounting for continued patterns of poor choices in the face of negative life consequences remain poorly understood. MethodsWe use a computational (active inference) modeling approach, combined with multiple regression and hierarchical Bayesian group analyses, to examine how treatment-seeking individuals with one or more SUDs (alcohol, cannabis, sedatives, stimulants, hallucinogens, and/or opioids; N = 147) and healthy controls (HCs; N = 54) make choices to resolve uncertainty within a gambling task. A subset of SUDs (N = 49) and HCs (N = 51) propensity-matched on age, sex, and verbal IQ were also compared to replicate larger group findings. ResultsResults indicate that: (a) SUDs show poorer task performance than HCs (p = 0.03, Cohen’s d = 0.33), with model estimates revealing less precise action selection mechanisms (p = 0.004, d = 0.43), a lower learning rate from losses (p = 0.02, d = 0.36), and a greater learning rate from gains (p = 0.04, d = 0.31); and (b) groups do not differ significantly in goal-directed information seeking. ConclusionsFindings suggest a pattern of inconsistent behavior in response to positive outcomes in SUDs combined with a tendency to attribute negative outcomes to chance. Specifically, individuals with SUDs fail to settle on a behavior strategy despite sufficient evidence of its success. These learning impairments could help account for difficulties in adjusting behavior and maintaining optimal decision-making during and after treatment.

HBA-DEALS: accurate and simultaneous identification of differential expression and splicing using hierarchical Bayesian analysis

We present Hierarchical Bayesian Analysis of Differential Expression and ALternative Splicing (HBA-DEALS), which simultaneously characterizes differential expression and splicing in cohorts. HBA-DEALS attains state of the art or better performance for both expression and splicing and allows genes to be characterized as having differential gene expression, differential alternative splicing, both, or neither. HBA-DEALS analysis of GTEx data demonstrated sets of genes that show predominant DGE or DAST across multiple tissue types. These sets have pervasive differences with respect to gene structure, function, membership in protein complexes, and promoter architecture.

Strong genetic structure among populations of the tick Ixodes ricinus across its range

Ixodes ricinus is the most common and widely distributed tick species in Europe, responsible for several zoonotic diseases, including Lyme borreliosis. Population genetics of disease vectors is a useful tool for understanding the spread of pathogens and infection risks. Despite the threat to the public health due to the climate-driven distribution changes of I. ricinus, the genetic structure of tick populations, though essential for understanding epidemiology, remains unclear. Previous studies have demonstrated weak to no apparent spatial pattern of genetic differentiation between European populations. Here, we analysed the population genetic structure of 497 individuals from 28 tick populations sampled from 20 countries across Europe, the Middle-East, and northern Africa. We analysed 125 SNPs loci after quality control. We ran Bayesian and multivariate hierarchical clustering analyses to identify and describe clusters of genetically related individuals. Both clustering methods support the identification of three spatially-structured clusters. Individuals from the south and north-western parts of Eurasia form a separated cluster from northern European populations, while central European populations are a mix between the two groups. Our findings have important implications for understanding the dispersal processes that shape the spread of zoonotic diseases under anthropogenic global changes.

Dataset for estimating occurrence probability of causations for plugged, abandoned and decommissioned oil and gas wells.

This article contains the dataset on the failure frequencies of the barrier and mechanical plugs in place within the hydrocarbon-containing wellbore during plugging and abandonment operation. The interpretation and application of this data can be found in the research article (“https://doi.org/10.1016/j.psep.2019.09.015” Babaleye et al., 2019). These datasets were collected through a comprehensive hazard identification technique workshop involving 10 engineers and academics with considerable years of field experience. The data were collected based on how likely it is for each causation to occur and these likelihoods are ranked from 1 to 10. The process is experience-driven and is complemented by a 1–10 rating of the duration of leak of hydrocarbon before remediation, should the leak reach the mudline. The ranked data was a representative of raw failure data (failure rate or mean time to failure (MTTF)) for each causation and are coded in MATLAB using gamma distribution based on hierarchical Bayesian analysis. The dataset offers unique opportunity for reuse due to its accessibility and discreteness.

Exploration of Hazardous Material Truck Crashes on Wyoming’s Interstate Roads using a Novel Hamiltonian Monte Carlo Markov Chain Bayesian Inference

Crash severity of a hazardous material (HAZMAT) transporting truck increases manyfold compared with normal truck crash because of the possible exposure to dangerous substances. Crashes which involve a HAZMAT truck might result in a catastrophic incident causing horrendous damage to individuals involved in the crash. In-transit HAZMAT crashes in Wyoming caused a total damage of $3.1 million from 2015 to 2018. HAZMAT crashes on interstate roads represented 22% of the total HAZMAT crashes causing a total damage of $2.2 million, representing 71% of the cost of total damage. Previous studies in Wyoming investigated all vehicle crashes, including large truck crashes, but none has analyzed HAZMAT-related crashes or accounted for its type as a contributing factor. This study fills the gap by analyzing crash injury severity of HAZMAT-related crashes on all interstate freeways in Wyoming. Furthermore, the study introduces the No-U-Turn (NUT) Hamiltonian Monte Carlo (HMC) method of hierarchical Bayesian analysis into HAZMAT crash injury severity analysis. In recent developments, NUT HMC has been proven to be the most efficient Markov Chain Monte Carlo (MCMC) sampling method. The results showed that 30% of the unobserved heterogeneity arises from variation in summer and winter crashes which justifies the use of hierarchical model. Among the other covariates investigated, the population-averaged effects showed that number of trucks involved, hit-and-run crashes, animal-vehicle crashes, work-zone-related crashes, collision type, percentage of females involved, drivers’ drug/alcohol use, seat-belt use, crash location, roadway curves, and surface conditions significantly impact HAZMAT crash injury severity.

PBPK modeling of impact of nonalcoholic fatty liver disease on toxicokinetics of perchloroethylene in mice

BackgroundNonalcoholic fatty liver disease (NAFLD), a major cause of chronic liver disease in the Western countries with increasing prevalence worldwide, may substantially affect chemical toxicokinetics and thereby modulate chemical toxicity. ObjectivesThis study aims to use physiologically-based pharmacokinetic (PBPK) modeling to characterize the impact of NAFLD on toxicokinetics of perchloroethylene (perc). MethodsQuantitative measures of physiological and biochemical changes associated with the presence of NAFLD induced by high-fat or methionine/choline-deficient diets in C57B1/6 J mice are incorporated into a previously developed PBPK model for perc and its oxidative and conjugative metabolites. Impacts on liver fat and volume, as well as blood:air and liver:air partition coefficients, are incorporated into the model. Hierarchical Bayesian population analysis using Markov chain Monte Carlo simulation is conducted to characterize uncertainty, as well as disease-induced variability in toxicokinetics. ResultsNAFLD has a major effect on toxicokinetics of perc, with greater oxidative and lower conjugative metabolism as compared to healthy mice. The NAFLD-updated PBPK model accurately predicts in vivo metabolism of perc through oxidative and conjugative pathways in all tissues across disease states and strains, but underestimated parent compound concentrations in blood and liver of NAFLD mice. ConclusionsWe demonstrate the application of PBPK modeling to predict the effects of pre-existing disease conditions as a variability factor in perc metabolism. These results suggest that non-genetic factors such as diet and pre-existing disease can be as influential as genetic factors in altering toxicokinetics of perc, and thus are likely contribute substantially to population variation in its adverse effects.

Nerve palsy, dislocation and revision rate among the approaches for total hip arthroplasty: a Bayesian network meta-analysis.

Total hip arthroplasty (THA) is one of the most performed intervention in orthopaedics surgery. Currently, there is no unanimous approval concerning the best approach for THA in terms of nerve palsies, dislocations and further revisions. Hence, a Bayesian network meta-analysis was conducted. The present study was conducted according to the PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of healthcare interventions. The literature search was performed in September 2019. The NMA was performed through the STATA Software/MP routine for Bayesian hierarchical random-effects model analysis. Data from 10,675 THA were collected. The mean follow-up was 10months. The anterior approach reported the lowest risk to incur a post-operative dislocation (overall inconsistency P = 0.99). The posterolateral approach reported the lowest risk to incur a nerve palsy (overall inconsistency P = 0.77). The funnel plot revealed a low risk of publication bias. The lateral approach was found to have the lowest risk of resulting in a revision surgery (overall inconsistency P = 0.90). According to our network comparisons, the posterolateral approach for THA represent the favourable exposure with regards to nerve palsy, further dislocations and revision surgeries.

Implant positioning among the surgical approaches for total hip arthroplasty: a Bayesian network meta-analysis.

Although there is anatomical individuality among patients, some standardized requirements for component orientation for total hip arthroplasty (THA) exist. To date, no study has compared implant positioning using a network meta-analysis (NMA). Hence, the purpose of this study was to perform a NMA comparing implant positioning among the various approaches for THA. This study was conducted according to the PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions. All randomized (RCT) and non-randomized (nRCT) clinical trials comparing two or more different approaches for THA were considered for inclusion. For the statistical analysis, the STATA Software/ MP, Version 14.1 (Stata Corporation, College Station, Texas, USA) was used. The NMA was performed through a Stata routine for Bayesian hierarchical random-effects model analysis. We enrolled in the present study, a total of 4060 patients, with a mean of 17.84 ± 16.41months follow-up. The patient baseline data were very homogeneous across the groups concerning age (P = 0.91), sex (P = 0.52), and BMI (P = 0.83). Concerning the cup anteversion, the anterolateral approach reported the value closest to the reference value. The test for overall inconsistency, however, was not statistically significant (P = 0.30). Concerning cup inclination, the lateral approach reported the value closest to the reference value. The test for overall inconsistency was statistically significant (P = 0.01). The anterolateral approach for total hip arthroplasty detected the best femoral stem alignment and cup anteversion. Data concerning cup inclination detected no sufficient transitivity to draw consistent conclusions.

PMH51 WHAT ARE THE ATTRIBUTES THAT PSYCHIATRISTS CONSIDER IN THE DECISION TO DISCHARGE ADULTS WITH MAJOR DEPRESSIVE DISORDER WHO HAVE BEEN HOSPITALIZED WITH ACTIVE SUICIDAL IDEATION WITH INTENT?

To ascertain the relative importance of various attributes that US-based psychiatrists consider when deciding to discharge adults hospitalized with major depressive disorder and active suicidal ideation with intent (MDSI). A choice-based conjoint (CBC) analysis was conducted via online survey among psychiatrists actively managing patients with MDSI. Potential attributes and levels within each attribute were identified through literature review with expert consultation and validated through qualitative interviews with 10 US-based psychiatrists. In the final CBC analysis, psychiatrists assessed which of three patients in hypothetical scenarios (in which levels for each attribute varied across patients) were most appropriate for discharge. A hierarchical Bayesian multinomial-logit analysis using a Markov Chain Monte Carlo algorithm was used to calculate model parameters (utilities). Ranges of utilities were used to determine the relative importance of attributes (scaled to sum to 100). 100 psychiatrists completed the survey; 48% were affiliated with a teaching/university hospital, and participants reported a mean of 14.5 years (standard deviation: 8.5) in practice. The relative importance of attributes considered in determining appropriateness for discharge were (from most to least important): current major depressive disorder (MDD) severity (24.8 [95% confidence interval: 23.3, 26.3]); clinician assessment of current suicidal ideation (SI) (20.8 [18.5, 23.0]); previous history of suicide attempts (16.7 [15.9, 17.6]); psychosocial support at discharge (13.0 [11.7, 14.4]); post-discharge outpatient follow-up plan (immediacy and frequency of visits) (9.8 [8.8, 10.8]); current length of hospital stay (9.2 [8.1, 10.3]); and clinician assessment of SI at admission (5.7 [4.8, 6.6]). Current level of symptoms of MDD, including overall depression severity and SI, were the most important attributes that psychiatrists reported considering when discharging patients hospitalized with MDSI. Further research should be conducted to determine whether treatment approaches targeting modifiable attributes (e.g. MDD symptom severity and SI) could reduce length of stay for these patients.

LO71: The effect of boarding time in ED on length of stay for psychiatric patients

Introduction: The Emergency Departments (ED) is a gateway to the health care system for many psychiatric patients. As a consequence of hospital administrative factors and overcrowding, admitted psychiatric patients are often boarded in the ED while waiting for an inpatient bed. There is currently a lack of evidence to quantify the effect that ED boarding has on psychiatric patients. The primary objective of this study is to determine whether a patient's length of stay is related to longer ED boarding time. Methods: This study is a retrospective cohort using data from an administrative source, which was obtained from patient records captured in the Sunrise Clinical Manager EMR used across Calgary, Alberta EDs from 2014-2018. A hierarchical Bayesian regression analysis was used to model the several patient-level and hospital-level factors. The mean and variance was defined by the exposure of interest, namely hours in the Emergency Department after admission to psychiatry unit expressed as a continuous variable. An interaction between this exposure and patient-level confounders was used to model the changing effect of a patient's severity in the ED on their boarding time. Results: The median boarding time for patients in our study was 6.6 hours (standard deviation 17.3), while the average was 13.6 hours. Patients who were boarded for greater than 6 hours more frequently required an antipsychotic (37% vs 11%; SMD 0.651), sedative (52% vs 29%; SMD 0.483) or restraints (18% vs. 14%; SMD 0.102). In crude analysis there was no difference in median length of stay for patients that were boarding more than 6 hours compared to those boarded for less than 6 hours (8 days vs 9 days; SMD 0.012).The rate ratio for length of stay is 1.05 with 95% posterior interval 1.04 - 1.06 for each 24 hour increase in boarding time. This means that for each 1 day worth of boarding time, the length of stay (in days) increases 1.05 times (or 0.05 days/day boarding time). Conclusion: Boarding time is associated with a small but absolute increase in length of stay for psychiatric patients. Decreasing boarding time could have ripple effects for ED efficiency and overall patient outcomes.

River dams and the stability of bird communities: A hierarchical Bayesian analysis in a tropical hydroelectric power plant

Abstract The effects of anthropogenic disturbance upon the stability of wildlife communities depend on the heterogeneity and connectivity of habitat remnants on multiple scales. The number of hydroelectric dams in biodiversity hotspots (Africa, South America and Asia) is growing rapidly. To establish their environmental impact, it is essential to understand the dynamics of wildlife communities before and following the establishment of dams. We evaluated the impacts of the filling of the Serra do Facão hydroelectric reservoir in the São Marcos river, central Brazil, on the bird community. Using data from 1,145 surveys across 20 sampling sites over 8 years, 2 years before and 6 years after the filling of the reservoir, we assessed the resistance, i.e. maintenance close to an equilibrium state during the disturbance, and resilience, i.e. ability to return to the original state following the disturbance, of the bird community. We used spatiotemporal hierarchical Bayesian models to assess the effects of reservoir filling on five community parameters: abundance, richness, phylogenetic diversity, functional diversity and species composition. In the period subsequent to reservoir filling, there was (a) a marked reduction in bird abundance, richness, phylogenetic diversity and functional diversity, and (b) a reduction in the proportion of forest species, coupled with an increase in the proportion of savanna species. Except for bird abundance, none of the other community attributes returned to their original levels, even after 6 years. Our findings indicate that Cerrado bird communities have both low resistance and low resilience to habitat loss associated with the establishment of hydroelectric reservoirs. Synthesis and applications. The environmental costs of hydroelectric dams are still underestimated or neglected in Brazil. A new paradigm in the assessment of their environmental impacts is warranted, incorporating (a) models of spatiotemporal variations based on long‐term monitoring with surveys initiated before disturbances and (b) measures of functional and phylogenetic diversity, such that society can understand the costs and benefits of the establishment of new hydroelectric dams and make informed decisions. Biodiversity loss could be minimized by ensuring the preservation and connectivity of alluvial habitats, capable of maintaining the supply of resources and the functional and phylogenetic attributes of bird communities associated with such habitats.