Empirical Bayesian Estimation Research Articles

Bayesian estimation in generalized linear models for longitudinal data with hyperspherical coordinates

Under the framework of generalized linear models (GLM), the generalized estimating equation (GEE) method is typically applied for longitudinal data analysis. However, there are a series of problems due to the misspecification of the within-subject correlation structure, especially in Bayesian estimation. To handle these difficulties, in this paper, we construct a class of generalized estimating equations for longitudinal data with hyperspherical coordinates (HPC) and propose a Bayesian approach established through empirical likelihood (EL). Additionally, an efficient Markov chain Monte Carlo (MCMC) procedure is developed for the required computation of the posterior distribution. As proved by the simulation studies and an application to a real longitudinal data set, our method not only performs better than traditional empirical likelihood estimation and Bayesian estimation with partial autocorrelations (PAC) but also is suitable for non-Gaussian data.

Spatial extension of mixed models of the analysis of variance

ABSTRACT This study proposes a spatial extension of mixed models of the analysis of variance (MANOVA), called mixed spatial ANOVA (MS-ANOVA) models. MS-ANOVA models have been used to evaluate the spatial correlations between random effects in spatial multilevel data in which observations belong to nested clusters. Because the proposed model can be regarded as a Bayesian hierarchical model, we introduce empirical Bayesian estimation methods in which hyperparameters are estimated by quasi-maximum likelihood estimation methods in the first step, and posterior distributions for the parameters are evaluated with the estimated hyperparameters in the second step. Moreover, we justify the asymptotic properties of the first-step estimator. The proposed models were applied to happiness survey data collected in Japan. Empirical results show that social capital, which can be interpreted as ‘the beliefs and norms by which a community values collective action and pursues activities worthy for the entire community’, significantly increases people’s happiness, even after controlling for various individual characteristics and spatial correlations.

Experience rating of risk premium for Esscher premium principle

In this article, a new method is introduced under the Bayesian framework to derive the credibility estimator of risk premiums based on the Esscher premium principle. This new estimator offers desirable statistical properties, making it more useful and practical compared to existing estimators. Additionally, Bayesian models for policy portfolios are established, and empirical Bayes methods are employed to estimate the structural parameters. The empirical Bayesian estimation of risk premiums is also discussed in detail. The convergence rate and goodness of the proposed estimators are verified through simulations. The results demonstrate the effectiveness and accuracy of the new estimator and its superior performance compared to other existing methods. Finally, an empirical analysis is conducted using real insurance data, which further confirms the applicability and reliability of the proposed credibility estimator and its superiority in practical insurance applications.

Robust small area estimation for unit level model with density power divergence.

Unit level model is one of the classical models in small area estimation, which plays an important role with unit information data. Empirical Bayesian(EB) estimation, as the optimal estimation under normal assumption, is the most commonly used parameter estimation method in unit level model. However, this kind of method is sensitive to outliers, and EB estimation will lead to considerable inflation of the mean square error(MSE) when there are outliers in the responses yij. In this study, we propose a robust estimation method for the unit-level model with outliers based on the minimum density power divergence. Firstly, by introducing the minimum density power divergence function, we give the estimation equation of the parameters of the unit level model, and obtain the asymptotic distribution of the robust parameters. Considering the existence of tuning parameters in the robust estimator, an optimal parameter selection algorithm is proposed. Secondly, empirical Bayesian predictors of unit and area mean in finite populations are given, and the MSE of the proposed robust estimators of small area means is given by bootstrap method. Finally, we verify the superior performance of our proposed method through simulation data and real data. Through comparison, our proposed method can can solve the outlier situation better.

SOCIAL BEHAVIOUR TOWARDS TAX PAYMENT: A SURVEY-BASED EVIDENCE FROM SADC COUNTRIES

Tax non-compliance and its consequences have become a subject of increasing interest in academic literature and economic forums worldwide. While most studies on this issue focus on developed countries, there is a growing trend to explore understudied developing countries. To fill this gap, we investigated tax evasion drivers in eight Southern African Development Community (SADC) countries, using the round 7 Afrobarometer survey data conducted in 2019-2020. The survey's comprehensive coverage of economic, political, and sociological questions made it one of the most extensive surveys on the continent. We used logistic regression and Empirical Bayesian estimation and found that political legitimacy significantly influences tax evasion behavior in the SADC region. Individuals residing within the SADC are more likely to engage in tax evasion activities when they perceive a lack of access to fundamental services provided by their governments or harbor doubts about the legitimacy of political institutions. Therefore, policymakers in SADC member states should prioritize reviewing and evaluating economic policies, the performance and efficiency of political institutions, and more inclusive governance. We suggest that a strong and legitimate political framework, coupled with effective service delivery, can contribute to reducing tax evasion rates and enhancing public welfare outcomes. Institutional reforms, increased transparency, accountability, and a more inclusive governance system are necessary for fostering a culture of compliance and trust, leading to improved revenue collection.

A spatial analysis of co-circulating dengue and chikungunya virus infections during an epidemic in a region of Northeastern Brazil

The aim of this study was to describe, through spatial analysis, the cases of arboviruses (dengue and chikungunya), including deaths, during the first epidemic after the circulation of the chikungunya virus (CHIKV) in the state of Pernambuco, Northeastern Brazil. This was an ecological study in both Pernambuco and the state capital, Recife, from 2015 to 2018. The odds ratios (OR) were estimated, and the statistical significance was considered p≤0.05. For the spatial analysis, Kulldorff's space-time scan statistics method was adopted to identify spatial clusters and to provide the relative risk (RR). In order to assess the significance at a level of p < 0.01 of the model, the number of Monte Carlo replications was 999 times. To perform the scan statistics we used the Poisson probability model, with a circular scanning window; annual temporal precision and retrospective analysis. A total of 227 deaths and 158,728 survivors from arboviruses was reported during the study period, with 100 deaths from dengue and 127 from CHIKV. The proportion of deaths from dengue was 0.08% and from chikungunya was 0.35%. The proportion of all those infected (deaths plus survivors) with dengue was 77.42% and with chikungunya was 22.58%. Children aged 0 to 9 years were around 3 times more likely to die than the reference group (OR 2.84; CI95% 1.16–5.00). From the age of 40, the chances of death increased significantly: 40–49 (OR 2.52; CI95% 1.19–5.29), 50–59 (OR 5.55; CI95% 2.76–11.17) and 60 or more (OR 14.90; CI95% 7.79–28.49). Males were approximately twice as likely to die as females (OR 1.77; CI95% 1.36–2.30). White-skinned people were less likely to die compared to non-white (OR 0.60; CI95% 0.41–0.87). The space-time analysis of prevalence in the state of Pernambuco revealed the presence of four clusters in the years 2015 and 2016, highlighting the Metropolitan Macro-region with a relative risk=4 and the Agreste and Hinterland macro-regions with a relative risk=3.3. The spatial distribution of the death rate in the municipality of Recife smoothed by the local empirical Bayesian estimator enabled a special pattern to be identified in the southwest and northeast of the municipality. The spatiotemporal analysis of the death rate revealed the presence of two clusters in the year 2015. In the primary cluster, it may be noted that the aforementioned aggregate presented a RR=7.2, and the secondary cluster presented a RR=6.0. The spatiotemporal analysis with Kulldorff's space-time scan statistics method, proved viable in identifying the risk areas for the occurrence of arboviruses, and could be included in surveillance routines so as to optimize prevention strategies during future epidemics.

Reliability analysis of bathtub-shaped distribution using empirical Bayesian and E-Bayesian estimations under progressive Type-II censoring

This paper investigates the performance of empirical Bayesian and E-Bayesian estimation of the bathtub-shaped distribution based on the progressive Type-II censored samples. These two estimations overcome the selection of hyperparameters in the Bayesian method. The empirical Bayesian analysis employs the classical approach to determine the hyperparameters. The E-Bayesian estimation uses the prior distribution of hyperparameters to derive the expectation of Bayesian estimates. The estimates of parameters are derived under the squared error loss and LINEX loss functions. The extensive simulations show the results of the empirical Bayesian estimates and the E-Bayesain estimates. Further, the empirical Bayesian estimation is also presented to analyze the parameters of bathtub-shaped distribution based on the general progressive Type-II censored samples. The two datasets from the electromechanical field and medical survival analysis are analyzed using the empirical Bayesian and E-Bayesian methods.

Exposure-response Modeling From the CLARITY Trial of Pimavanserin for Adjunctive Treatment of Major Depressive Disorder.

In the 10-week, phase 2 CLARITY study of patients with major depressive disorder, adjunctive therapy to antidepressants with pimavanserin 34mg once daily statistically significantly improved the Hamilton Depression Rating Scale (HAMD-17) total score (primary endpoint) and Sheehan Disability Scale (SDS) score (secondary endpoint) versus placebo. This analysis characterized the exposure-response (E-R) relationships of pimavanserin in this CLARITY patient population. Exposure measures were estimated for each patient based on population-pharmacokinetic empirical Bayesian estimates. E-R models were developed to describe exposure-efficacy (HAMD-17, SDS, and Clinical Global Impression-Improvement [CGI-I] scale) and exposure-safetyrelationships (Karolinska Sleepiness Scale [KSS], Massachusetts General Hospital Sexual Functioning Inventory [MGH-SFI], and adverse events [AEs] of headache, sedation, and somnolence) relationships. For the primary efficacy endpoint (HAMD-17 scores), a sigmoid maximum-effect model described the time course of response, and a linear function of pimavanserin exposure was statistically significant. HAMD-17 scores decreased steadily over time following placebo and pimavanserin treatment; separation from placebo increased as peak pimavanserin plasma concentration (Cmax ) increased. At median pimavanserin Cmax (34-mg dose), the reduction from baseline in HAMD-17 scores was -11.1 and -13.5 at weeks 5 and 10, respectively. Relative to placebo, the model predicted comparable reductions in HAMD-17 scores at weeks 5 and 10. Similar improvements in favor of pimavanserin were detected with SDS, CGI-I, MGH-SFI, and KSS scores. No E-R relationship was found for AEs. E-R modeling predicted a relationship between higher pimavanserin exposure and improvement in HAMD-17 score and improvement across multiple secondary efficacy endpoints.

Structural damage detection of the bridge under moving loads with the quasi-static displacement influence line from one sensor

As the bridges continue to age, it is imperative to identify structurally deficient bridges at an early stage to prevent unexpected maintenance needs. A quasi-static method is developed to identify structural damage based on the influence line (IL) from one sensor and an empirical Bayesian threshold estimator. Initially, the empirical Bayesian threshold estimator is utilized to preprocess the measured bridge displacement response to eliminate vehicle-induced dynamic effects. Subsequently, a regularization technique using Elastic Net and displacement response is used for IL estimation, and the Elastic Net that integrates with both L1 and L2 norms is adopted to establish the objective function of IL estimation to deal with the ill-posedness of the inverse problem for specific vehicle configurations. Finally, according to the mechanical relationship between the moving load and the displacement IL, an IL area-based difference index is established to identify the presence of the damage rapidly, raising a flag that leads to a more in-depth investigation to gather information related to structural integrity. The feasibility of the proposed method is verified with the numerical model and laboratory tests. This study introduces innovative contributions in the following three aspects: (1) Elastic Net that integrates with both L1 and L2 norms is adopted to establish the objective function for IL estimation, (2) A single sensor provides a cost-effective and less computationally expensive solution than multi-sensor methods, (3) Compared to other wavelet-based methods, the comparative studies demonstrate that empirical Bayesian threshold estimator can identify damage more clearly.

Interferometric SAR Coherence Magnitude Estimation by Machine Learning

Current interferometric wide area ground motion services require the estimation of the coherence magnitude as accurately and computationally effectively as possible. However, a precise and at the same time computationally efficient method is missing. Therefore, the objective of this article is to improve the empirical Bayesian coherence magnitude estimation in terms of accuracy and computational cost. Precisely, this article proposes the interferometric coherence magnitude estimation by Machine Learning (ML). It results in a non-parametric and automated statistical inference. However, applying ML in this estimation context is not straightforward. The number and the domain of possible input processes is infinite and it is not possible to train all possible input signals. It is shown that the expected channel amplitudes and the expected interferometric phase cause redundancies in the input signals allowing to solve this issue. Similar to the empirical Bayesian methods, a single parameter for the maximum underlaying coherence is used to model the prior. However, no prior or any shape of prior probability is easy to implement within the ML framework. The article reports on the bias, the standard deviation and the root mean square error (RMSE) of the developed estimators. It was found that ML estimators improve the coherence estimation RMSE from small samples ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$2 N &lt; 30$</tex-math></inline-formula> ) and for small underlaying coherence compared to the conventional and empirical Bayes estimators. For three interferometric samples ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$N = 3$</tex-math></inline-formula> ) and a zero coherence magnitude, the bias related to the sample estimator improves from 0.53 to 0.39 by 27.8%. Assuming the maximum underlaying coherence is 0.6, the bias is reduced by 33.0% to 0.36 for the less strict and by 45.5% to 0.29 for the strict prior. The developed ML coherence magnitude estimators are suitable and recommended for operational InSAR systems. For the estimation, the ML model is extremely fast evaluated because no iteration, numeric integration or Bootstrapping is needed.

232. Population Pharmacokinetics (PopPK) of Ceftazidime-Avibactam (CZA) in Combination with Aztreonam (ATM) in a Phase I, Open-Label Study in Healthy Adult Subjects (COMBINE)

Abstract Background Scant PK data are available with CZA-ATM in combination. Occurrence of ALT/AST elevations are common with ATM, and it is unknown if it is exacerbated by use of CZA-ATM. This analysis of COMBINE sought to describe the popPK of CZA-ATM and to assess the association between ATM exposures and ALT/AST elevations. Methods COMBINE was a Phase I study of 48 healthy subjects aged 18-45 years (NCT03978091). Subjects were enrolled into 1 of 6 Cohorts (Table 1). Drug(s) were administered for 7 D and intensive plasma and urine PK sampling was performed. PopPK models were developed for ceftazidime (CAZ), avibactam (AVI), and ATM. Empirical Bayesian estimates from the ATM PopPK model were used to simulate ATM day 1 (D1) exposures. Associations between D1 ATM exposures and highest observed ALT/AST were assessed with curvilinear regression (CR) and generalized linear models (GLM). Results Of enrolled 48 subjects, 35% had ALT/AST elevations; 94% occurred in ATM/CZA-ATM Cohorts. Two severe study product related ALT/AST AEs were observed with in the continuous infusion (CI) ATM Cohort. All subjects with ALT/ATS elevations were asymptomatic with no other signs of liver injury. In the ATM PopPK model, CZA-ATM administration reduced ATM non-renal clearance (CLNR) by 1 L/hr (16% of total CL) (Table 2). Administration of CAZ-ATM had a negligible effect on total CAZ CL in the CAZ PopPK model and CZA-ATM was not a covariate in final AVI PopPK model. In the CR, no ATM exposure-ALT/AST associations were identified in overall analyses. Associations (R2=0.18-0.22, p&amp;lt; 0.05) between D1 AUC and ALT/AST were observed in CR analyses restricted to the intermittent infusion (II) ATM Cohorts (Figure 1). In GLM, D1 AUC was significantly associated with ALT/AST in the ATM II Cohort analyses. Administration of CZA-ATM was not associated with ALT/AST in the GLM analyses. Figure 1 Conclusion Administration of CZA-ATM was found to reduce ATM CL, resulting in higher daily ATM AUCs, but did not exacerbate AST/ALT elevations relative to ATM alone. The observed association between ATM AUC and ALT/AST elevations in subjects who received II ATM suggest the risks vs benefits of using II of ATM 8 g/daily vs 6 g/daily with CZA should be considered. Use of CI ATM should be used with caution given the 2 severe ALT/ASTs in the CI ATM Cohort. Disclosures Thomas Lodise, Jr., Pharm.D., PhD, BioFire Diagnostics: Grant/Research Support|cidara: Advisor/Consultant|cidara: Honoraria|Entasis: Grant/Research Support|Merck: Advisor/Consultant|Merck: Grant/Research Support|Paratek: Advisor/Consultant|Shionogi: Advisor/Consultant|Spero: Advisor/Consultant|Venatrox: Advisor/Consultant J Nicholas O'Donnell, Pharm.D., Merck &amp; Co, Inc: Grant/Research Support|Paratek Pharmaceuticals: Grant/Research Support Stephen Balevic, MD, Purdue Pharma: Grant/Research Support|UCB: Advisor/Consultant Jeffrey Guptill, MD, argenx: Stocks/Bonds Vance G. Fowler, Jr, MD, MHS, Affinergy: Grant/Research Support|Affinergy: Honoraria|Affinium: Honoraria|Amphliphi Biosciences: Honoraria|ArcBio: Stocks/Bonds|Basilea: Grant/Research Support|Basilea: Honoraria|Bayer: Honoraria|C3J: Honoraria|Cerexa/Forest/Actavis/Allergan: Grant/Research Support|Contrafect: Grant/Research Support|Contrafect: Honoraria|Cubist/Merck: Grant/Research Support|Debiopharm: Grant/Research Support|Deep Blue: Grant/Research Support|Destiny: Honoraria|Genentech: Grant/Research Support|Genentech: Honoraria|Integrated Biotherapeutics: Honoraria|Janssen: Grant/Research Support|Janssen: Honoraria|Karius: Grant/Research Support|Medicines Co.: Honoraria|MedImmune: Grant/Research Support|MedImmune: Honoraria|NIH: Grant/Research Support|Novartis: Grant/Research Support|Novartis: Honoraria|Pfizer: Grant/Research Support|Regeneron: Grant/Research Support|Regeneron: Honoraria|Sepsis diagnostics: Sepsis diagnostics patent pending|UpToDate: Royalties|Valanbio: Stocks/Bonds henry chambers, MD, Merck: DSMB member|Merck: Stocks/Bonds|Moderna: Stocks/Bonds.

Evaluation of the accessibility and its equity of the national public-private mix program for tuberculosis in Korea: a multilevel analysis.

To examine the effect of individual and area-level characteristics on the probability of public-private mix (PPM) support (PPM coverage) for tuberculosis (TB). This study is a retrospective cohort design using TB reporting and treatment management data in Korea. We analyzed PPM coverage through multilevel logistic regression and empirical Bayesian estimation according to individual and area-level characteristics and their interaction. Patients aged 0-29 years, women, of Korean nationality, treated at a general hospital, a one-time reporting, urban areas, and the lowest deprivation index (DI) showed higher PPM coverage. Due to the cross-level interaction, PPM coverage in the urban areas (slope=-0.048, p<0.001) had a higher level but a steeper negative deprivation gradient than in rural areas (slope= -0.015, p<0.001). For a general hospital, the PPM coverage in urban is high but more significantly decreased than in rural areas with the higher DI (urban: slope=-0.047, p<0.001; rural: slope=-0.031, p<0.001). For clinics and hospitals, the effect of DI did not appear in urban areas, but in rural areas, the higher the DI, the higher the PPM coverage with a slope of 0.046 (p<0.001) and 0.063 (p<0.001), respectively. The PPM program created a significant disparity in PPM coverage between urban-rural areas and type of healthcare provider according to DI. Considering the high risk of TB incidence in areas with higher DI, institutional improvement and program redesign are needed to improve accessibility and equity.

GIS-Based Study of Dental Accessibility and Caries in 3-Year-Old Japanese Children

This study aimed to reveal the geographic accessibility of dental clinics for most municipalities in Japan in 2015 and to explore the association between dental accessibility and dental caries status in 3-year-old children. We computed the accessibility index and accessibility index rate for the population outside a 1-km radius of dental clinics using a geographic information system. We also used spatial autocorrelation analysis (Moran's I statistic) to examine the spatial clustering patterns of dental accessibility in Japanese municipalities. In addition, we adjusted the prevalence of dental caries for most municipalities using empirical Bayesian estimation. Finally, we applied multiple linear regression to scrutinise the associations between dental caries status, including the prevalence of dental caries and decayed and filled teeth (dft), and dental accessibility, with adjustments made for other sociodemographic variables. The distribution of dental accessibility in Japanese municipalities is relatively unequal. Dental accessibility is decent in the 3 metropolitan areas around Tokyo, Osaka, and Nagoya but poor in the Tohoku and Kyushu regions. In addition, dental accessibility is significantly related to the prevalence of dental caries and dft after adjusting for other sociodemographic variables (P < .005). This study suggests that dental accessibility is considerably connected to the dental caries status of 3-year-old children after excluding financial burden. Preschool children in areas with poor dental accessibility are likely to have poor dental caries status. We also verified the inequality of dental accessibility amongst Japanese municipalities. For the future development of primary oral health care, more attention should be paid to people with a disadvantage in terms of dental accessibility.

Safety of sildenafil in premature infants at risk of bronchopulmonary dysplasia: Rationale and methods of a phase II randomized trial

Bronchopulmonary dysplasia (BPD) is a disease of chronic respiratory insufficiency stemming from premature birth and iatrogenic lung injury leading to alveolar simplification, impaired alveolar-capillary development, interstitial fibrosis, and often pulmonary hypertension. BPD is the most common pulmonary sequela of prematurity and is often fatal; however, there remains no FDA-approved therapies to treat or prevent BPD. Sildenafil is increasingly used off-label in premature infants despite scant safety and efficacy data. Sildenafil reduces lung injury and preserves normal vasculature in preclinical models, and improves outcomes in children with pulmonary hypertension, and thus is a promising candidate for BPD. Following phase I studies, we developed the phase II SIL02 trial to describe the safety, pharmacokinetics and preliminary effectiveness of intravenous and enteral sildenafil in premature infants at risk for BPD. SIL02 is a randomized, double-blind, placebo-controlled, 3-cohort, sequential dose-escalating trial of enteral or intravenous (IV) sildenafil dosed every 8 h for up to 34 days. The target IV doses were 0.125, 0.5 and 1 mg/kg/dose in cohorts 1, 2 and 3, respectively; while the enteral doses will be double the IV doses. Eligible infants must be < 29 weeks' gestation at birth and requiring respiratory support at 7–28 days' postnatal age. Adverse events and preliminary effectiveness will be compared by treatment group. Using the final population PK model, empirical Bayesian estimates will be generated for each patient. Preliminary effectiveness will be measured by the incidence of moderate to severe BPD or death at 36 weeks and change in the BPD risk estimation.

Space-Time Clustering and Socioeconomic Factors Associated with Mortality from Diarrhea in Alagoas, Northeastern Brazil: A 20-Year Population-Based Study.

Acute diarrhea is the second leading cause of death among children in developing countries and is strongly related with the socioeconomic conditions of the population. In Brazil, data show a drop in the diarrhea mortality rate. Nevertheless, the northeastern region still has the most deaths. Considering this, we analyze high-risk areas for diarrhea- and gastroenteritis-related deaths, and their association with social determinants of health (SDH) in the state with one of the worst human development indicators in Brazil (Alagoas) between 2000 and 2019. We applied temporal, spatial, and space–time risk modelling. We used a log-linear regression model to assess temporal trends and the local empirical Bayesian estimator, the global and local Moran indices for spatial analysis. Spearman’s correlation was used to correlate mortality rates with SDH. A total of 3472 diarrhea-related deaths were reported during this period in Alagoas. We observed a decreasing time trend of deaths in the state (9.41/100,000 in 2000 to 2.21 in 2019; APC = −6.7; p-value < 0.001), especially in children under one year of age. However, there was stability among adults and the elderly. We identified two high-risk spatiotemporal clusters of mortality in inland municipalities. Lastly, mortality rates correlated significantly with 90% of SDH. Taken together, these findings indicate that diarrhea diseases remain a serious public health concern in Alagoas, mainly in the poorest and inland municipalities. Thereby, it is urgently necessary to invest in measures to control and prevent cases, and improve the living conditions of the poorest populations and those with the highest social vulnerability index.

Leveraging Predictive Pharmacometrics-Based Algorithms to Enhance Perinatal Care-Application to Neonatal Jaundice.

The field of medicine is undergoing a fundamental change, transforming towards a modern data-driven patient-oriented approach. This paradigm shift also affects perinatal medicine as predictive algorithms and artificial intelligence are applied to enhance and individualize maternal, neonatal and perinatal care. Here, we introduce a pharmacometrics-based mathematical-statistical computer program (PMX-based algorithm) focusing on hyperbilirubinemia, a medical condition affecting half of all newborns. Independent datasets from two different centers consisting of total serum bilirubin measurements were utilized for model development (342 neonates, 1,478 bilirubin measurements) and validation (1,101 neonates, 3,081 bilirubin measurements), respectively. The mathematical-statistical structure of the PMX-based algorithm is a differential equation in the context of non-linear mixed effects modeling, together with Empirical Bayesian Estimation to predict bilirubin kinetics for a new patient. Several clinically relevant prediction scenarios were validated, i.e., prediction up to 24 h based on one bilirubin measurement, and prediction up to 48 h based on two bilirubin measurements. The PMX-based algorithm can be applied in two different clinical scenarios. First, bilirubin kinetics can be predicted up to 24 h based on one single bilirubin measurement with a median relative (absolute) prediction difference of 8.5% (median absolute prediction difference 17.4 μmol/l), and sensitivity and specificity of 95.7 and 96.3%, respectively. Second, bilirubin kinetics can be predicted up to 48 h based on two bilirubin measurements with a median relative (absolute) prediction difference of 9.2% (median absolute prediction difference 21.5 μmol/l), and sensitivity and specificity of 93.0 and 92.1%, respectively. In contrast to currently available nomogram-based static bilirubin stratification, the PMX-based algorithm presented here is a dynamic approach predicting individual bilirubin kinetics up to 48 h, an intelligent, predictive algorithm that can be incorporated in a clinical decision support tool. Such clinical decision support tools have the potential to benefit perinatal medicine facilitating personalized care of mothers and their born and unborn infants.

An Evaluation of Gene-Diet Interaction Statistical Methods and Discovery of rs7175421-Whole Grain Interaction in Lung Cancer

Dietary factors show different effects on genetically diverse populations. Scientific research uses gene-environment interaction models to study the effects of dietary factors on genetically diverse populations for lung cancer risk. However, previous study designs have not investigated the degree of type I error inflation and, in some instances, have not corrected for multiple testing. Using a motivating investigation of diet-gene interaction and lung cancer risk, we propose a training and testing strategy and perform real-world simulations to select the appropriate statistical methods to reduce false-positive discoveries. The simulation results show that the unconstrained maximum likelihood (UML) method controls the type I error better than the constrained maximum likelihood (CML). The empirical Bayesian (EB) method can compete with the UML method in achieving statistical power and controlling type I error. We observed a significant interaction between SNP rs7175421 with dietary whole grain in lung cancer prevention, with an effect size (standard error) of −0.312 (0.112) for EB estimate. SNP rs7175421 may interact with dietary whole grains in modulating lung cancer risk. Evaluating statistical methods for gene-diet interaction analysis can help balance the statistical power and type I error.

Bias correction for multiple covariate analysis using empirical bayesian estimation in mixed-effects models for longitudinal data

The naïve empirical Bayes method has been widely used as an ad hoc tool in fitting linear mixed-effect models, which is much computationally efficient than the maximum likelihood estimation method. However, the shrinkage effect of the empirical Bayes method causes bias in the estimates of the fixed effects. Bias-correction has been proposed for the mixed-effects model when only one covariate is present. In this paper, we derive the shrinkage factor of the empirical Bayes predictors of the random effects and the variance-covariance matrix of the corrected estimates when the model has more than one covariate. The empirical Bayes estimates and test statistics are then corrected using the derived factor. Theoretical derivations, simulation studies and a real data application demonstrate the validity of the proposed method in that the corrected estimates are unbiased and the corrected tests have correct p-values.

Serotonin transporter binding in major depressive disorder: impact of serotonin system anatomy.

Serotonin transporter (5-HTT) binding deficits are reported in major depressive disorder (MDD). However, most studies have not considered serotonin system anatomy when parcellating brain regions of interest (ROIs). We now investigate 5-HTT binding in MDD in two novel ways: (1) use of a 5-HTT tract-based analysis examining binding along serotonergic axons; and (2) using the Copenhagen University Hospital Neurobiology Research Unit (NRU) 5-HT Atlas, based on brain-wide binding patterns of multiple serotonin receptor types. [11C]DASB 5-HTT PET scans were obtained in 60 unmedicated participants with MDD in a current depressive episode and 31 healthy volunteers (HVs). Binding potential (BPP) was quantified with empirical Bayesian estimation in graphical analysis (EBEGA). Within the [11C]DASB tract, the MDD group showed significantly lower BPP compared with HVs (p = 0.02). This BPP diagnosis difference also significantly varied by tract location (p = 0.02), with the strongest MDD binding deficit most proximal to brainstem raphe nuclei. NRU 5-HT Atlas ROIs showed a BPP diagnosis difference that varied by region (p < 0.001). BPP was lower in MDD in 3/10 regions (p-values < 0.05). Neither [11C]DASB tract or NRU 5-HT Atlas BPP correlated with depression severity, suicidal ideation, suicide attempt history, or antidepressant medication exposure. Future studies are needed to determine the causes of this deficit in 5-HTT binding being more pronounced in proximal axon segments and in only a subset of ROIs for the pathogenesis of MDD. Such regional specificity may have implications for targeting antidepressant treatment, and may extend to other serotonin-related disorders.

Can First-Dose Therapeutic Drug Monitoring Predict the Steady State Area Under the Blood Concentration-Time Curve of Busulfan in Pediatric Patients Undergoing Hematopoietic Stem Cell Transplantation?

Busulfan has high intra-individual variability and possible time-dependent changes in clearance, which complicates therapeutic drug monitoring (TDM), as first dose sampling may not predict the steady state concentrations. In this study, we aimed to use Bayesian pharmacokinetic parameters estimated from the first dose to predict the steady state AUC for busulfan. This observational study was conducted among pediatric patients at King Abdullah Specialist Children’s Hospital. From each patient, we collected six blood samples (2, 2.25, 2.5, 3, 4, and 6 h after the start of IV infusion of the first dose). A subset of patients were also sampled at the steady state. First, we modeled the data using only the first dose. The model was used to estimate the empirical Bayesian estimates of clearance for each individual patient, then we used the empirical Bayesian estimates of clearance to predict the AUC0–tau at steady state (i.e., predicted AUC0–tau). Steady state AUC0–tau was also calculated for patients sampled at steady state using the trapezoidal method using raw time concentration data; this was considered the reference AUC0–tau.. Then, we compared the AUC0–tau predicted using the Bayesian approach with the reference AUC0–tau values. We calculated bias and precision to assess predictability. In total we had 33 patients sampled after first dose and at steady state. Using the Bayesian approach to predict the AUC0–tau, bias was −2.8% and precision was 33%. This indicates that first dose concentrations cannot accurately predict steady state busulfan concentrations; therefore, follow-up TDM may be required for optimal dosing.