Covariate Model Research Articles

A double sampling multivariate GLR control charting method for joint monitoring of process mean vector and covariance matrix under additive covariate model

A process is in its stable state when it has sufficient precision and accuracy. The accuracy is expressed as the deviation of the process mean from its target while the low dispersion indicates the high process precision. Therefore, joint monitoring of multivariate process mean and variability ensures a high rate of producing items. This study proposes a multivariate generalized likelihood ratio scheme called DS-MGLR with two features: (1) considering the effect of imprecise observations, and (2) enhancing the chart sensitivity in reacting to process disturbances. Since the shift magnitude in the distribution parameters is not known in advance, the expected average run length is also calculated for both the MGLR and DS-MGLR charts. To reduce the impact of measurement errors, an improved version of the DS-MGLR chart based on multiple measurement approach is developed. Three comparative studies are presented to indicate the importance of the properties of the developed charts. The first study confirms the efficiency of the double sampling strategy, while the second study indicates the error effect. The third one shows the effectiveness of the multiple measurement approach in compensating for the error contamination. Finally, the applicability of the proposed chart is highlighted using a real data example.

Defining childhood hypertension: is it too complicated? An evaluation of the potential impact of different approaches in an Australian paediatric population.

Current American Academy of Pediatrics (AAP) and European Society of Hypertension (ESH) thresholds defining hypertension in children use blood pressure (BP) normalised to age, sex and height. However, scare data exists regarding the relative importance of these variables to accurately model the 95th quantile of BP. We hypothesised that height alone may fit the population data equally well compared to more complex definitions. We also compare the potential impact of various thresholds for defining hypertension in an Australian population. Longitudinal data from the Raine Study were used, with 2248 participants contributing 7479 valid BP values across the 3/5/10/14/17-year study visits. BP was measured after 5 min rest, ≥3 times at each visit, using a Dinamap device. Quantile regression was used to predict the 95th percentile of BP, with nonlinear modelling of covariates through restricted cubic spline terms. At a single visit, 6-16% of young children exceeded the ESH threshold and 12-23% the AAP threshold. The transition to fixed thresholds (≥13 years AAP, ≥16 years ESH), increased the number of males (AAP only) and reduced the number of females considered hypertensive. A quantile regression model constructed with Raine Study data using height-only as the explanatory variable better predicted BP than the respective model using age-only (or a combination of the two). There may be large differences in the prevalence of hypertension according to AAP and ESH criteria, with a marked sex-discrepancy emerging from the point of fixed threshold application in adolescence. It may not be necessary to normalise BP by both age and height, the latter being a better predictor of childhood BP. Simpler methods may be preferable in clinical practice but require validation against clinical outcomes.

Transfer behaviors in stroke and dementia development associated with environmental risks.

Chronic diseases are influential components of stroke, one of the dominant reasons for dementia and premature mortality. Environmental risks are risk factors for transitioning from stroke to dementia. This study addresses the transition behaviors in stroke and dementia development associated with chronic diseases and environmental risks. This study is an integrated survey of medical and environmental informatics concerning stroke patients' quality of life. A total of 10 627 stroke patients diagnosed in Taiwan were surveyed in this study. A covariate model and subgroup analysis were used to evaluate the influence of chronic diseases and environmental risk factors (i.e., divorce rate, unemployment rate, solitariness rate, temperature, and air pollution rate) on stroke and the corresponding dementia transition behaviors. This study constructed a total of 98 covariate analysis models, consisting of 14 transition types [10 transitions from chronic diseases to stroke (5 metabolic risk states × 2 stroke states) and 4 transitions from stroke to dementia (2 stroke states × 2 dementia states)] by 7 covariates (i.e., sex, age, divorce rate, unemployment rate, temperature, air pollution, and solitariness rate). Among the 98 transitions, 26 were statistically significant. Sex, age, divorce rate, unemployment rate, temperature, and air pollution rate exerted a partially significant influence on the transition from chronic diseases to stroke. Sex, age, unemployment rate, and temperature partially influenced the transition from stroke to dementia. This study also considered high-risk sub-populations of stroke patients, particularly males aged 65 years and below. Geriatr Gerontol Int 2024; 24: 766-772.

Steroid-tacrolimus drug-drug interaction and the effect of CYP3A genotypes.

Tacrolimus, metabolized by CYP3A4 and CYP3A5 enzymes, is susceptible to drug-drug interactions (DDI). Steroids induce CYP3A genes to increase tacrolimus clearance, but the effect is variable. We hypothesized that the extent of the steroid-tacrolimus DDI differs by CYP3A4/5 genotypes. Kidney transplant recipients (n = 2462) were classified by the number of loss of function alleles (LOF) (CYP3A5*3, *6 and *7 and CYP3A4*22) and steroid use at each tacrolimus trough in the first 6months post-transplant. A population pharmacokinetic analysis was performed by nonlinear mixed-effect modelling (NONMEM) and stepwise covariate modelling to define significant covariates affecting tacrolimus clearance. A stochastic simulation was performed and translated into a Shiny application with the mrgsolve and Shiny packages in R. Steroids were associated with modestly higher (3%-11.8%) tacrolimus clearance. Patients with 0-LOF alleles receiving steroids showed the greatest increase (11.8%) in clearance compared to no steroids, whereas those with 2-LOFs had a negligible increase (2.6%) in the presence of steroids. Steroid use increased tacrolimus clearance by 5% and 10.3% in patients with 1-LOF and 3/4-LOFs, respectively. Steroids increase the clearance of tacrolimus but vary slightly by CYP3A genotype. This is important in individuals of African ancestry who are more likely to carry no LOF alleles, may more commonly receive steroid treatment, and will need higher tacrolimus doses.

Characterizing Enoxaparin's Population Pharmacokinetics to Guide Dose Individualization in the Pediatric Population.

Pediatric dosing of enoxaparin was derived based on extrapolation of the adult therapeutic range to children. However, a large fraction of children do not achieve therapeutic anticoagulation with initial dosing. We aim to use real-world anti-Xa data obtained from children receiving enoxaparin per standard of care to characterize the population pharmacokinetics (PopPK).Author names: Please confirm if the author names are presented accurately and in the correct sequence (given name, middle name/initial, family name). Also, kindly confirm the details in the metadata are correct.The author names are accurately presented and the metadata are correct. METHODS: A PopPK analysis was performed using NONMEM, and a stepwise covariate modeling approach was applied for the covariate selection. The final PopPK model, developed with data from 1293 patients ranging in age from 1 day to 18 years, was used to simulate enoxaparin subcutaneous dosing for prophylaxis and treatment based on total body weight (0-18 years, TBW) or fat-free mass (2-18 years, FFM). Simulated exposures in children with obesity (body mass index percentile ≥95th percentile) were compared with those without obesity. A linear, one-compartment PopPK model that included allometric scaling using TBW (<2 years) or FFM (≥2 years) characterized the enoxaparin pharmacokinetic data. In addition, serum creatinine was identified as a significant covariate influencing clearance. Simulations indicated that in patients aged <2 years, the recommended 1.5 mg/kg TBW-based dosing achieves therapeutic simulated concentrations. In pediatric patients aged ≥2 years, the recommended 1.0 mg/kg dose resulted in exposures more comparable in children with and without obesity when FFM weight-based dosing was applied. Using real-world data and PopPK modeling, enoxaparin's pharmacokinetics were characterized in pediatric patients. Using FFM and twice-daily dosing might reduce the risk of overdosing, especially in children with obesity.

Can the postload-fasting glucose gap be used to determine risk of developing diabetes in chinese adults: A prospective cohort study

ObjectiveTo evaluate the relationship between fasting plasma glucose (FPG) and 2-hour postload plasma glucose (2hPG) measured during an oral glucose tolerance test, and the risk of developing diabetes in Chinese adults. MethodsWe followed 3,094 participants without diabetes, categorizing them based on their oral glucose tolerance test (OGTT) results into low post load (2hPG ≤ FPG) and high post load (2hPG > FPG) at baseline. We monitored the incidence of diabetes, incidence of prediabetes, disease progression from prediabetes to diabetes and disease reversal from prediabetes to normal glucose tolerance (NGT) over an average of 3.2 years of follow-up.After the Schoenfeld residual test, Cox’s time-varying covariate (Cox-TVC) models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CI) to compare the different clinical events between low and high post load groups. ResultsIn the cohort study, of the 3,094 participants, 702 (22.7 %) had low post load (2hPG ≤ FPG, mean postload-fasting gap: −0.8 ± 0.7 mmol/L) and 2,392 (77.3 %) had high post load (2hPG > FPG, mean postload-fasting gap: 1.8 ± 1.2 mmol/L). Over 3.2 ± 0.2 years of follow-up, 282 (9.1 %) developed diabetes. In the low post load group, the incidence rates per 1,000 person-years were: diabetes was 7.9, prediabetes was 70.0, disease progression from prediabetes to diabetes was 23.4 and disease reversal to NGT was 327.2. For the high post load group, incidence rates for diabetes was 13.9, prediabetes was 124.3, disease progression was 59.5 and disease reversal was 238.6 per 1,000 person-years.Participants with high post load showed higher incidence rates of diabetes, prediabetes, and progression from prediabetes to diabetes compared to those with low post load. HRs were significantly higher for incident diabetes and prediabetes, and disease progression from prediabetes to diabetes, whereas disease reversal was lower. ConclusionThe risk of developing prediabetes/diabetes after 3.2 years of follow-up was higher in the participants with high post load. It suggested that postload-fasting gap may be a simple tool to predict the risk of developing prediabetes, diabetes or reversal to NGT.

Longitudinal associations of time-varying homeboundness and life satisfaction after traumatic brain injury.

To examine the association of changes in homebound status (i.e., never/rarely leaving the home) with life satisfaction in the first 10 years after traumatic brain injury (TBI). We analyzed data from 2,294 individuals with moderate-to-severe TBI from the TBI Model Systems National Database using a longitudinal multilevel model with time-varying predictors to account for within-person changes over time as well as between-person differences. We measured homebound status (defined as leaving the home ≤ 2 days/week) and life satisfaction (defined as the total score on the Satisfaction With Life Scale) at 1, 2, 5, and 10 years post-TBI. We adjusted the models for demographic and injury-related covariates and used inverse probability weighting to account for selection bias. Forty-five (2%) individuals were homebound at all follow-up visits, 523 (22.8%) were homebound at least one follow-up visit, and 1,726 (75.2%) were never homebound. Individuals with TBI who were consistently homebound had > 1 SD lower life satisfaction compared to those who were never homebound, β = -8.07, 95% confidence interval (CI) = [-9.39, -6.76], p < .001. Individuals who became homebound experienced a significant, but modest, decline in life satisfaction (β = -2.13, 95% CI = [-2.66, -1.61], p < .001). Our results indicate that being homebound and becoming homebound are associated with decreased life satisfaction. Homeboundness is a potentially modifiable target to improve life satisfaction, and elucidation of contributing factors to homebound status will help develop interventions to ameliorate post-TBI homeboundness. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Correlation adjusted debiased Lasso: debiasing the Lasso with inaccurate covariate model

Correlation adjusted debiased Lasso: debiasing the Lasso with inaccurate covariate model

Traffic crash prediction in Kano, Nigeria: multivariate long short-term memory method

Accurate traffic crash prediction is crucial for implementing effective road safety measures. In this study, the performance of long short-term memory (LSTM) and multivariate LSTM (MLSTM) models in forecasting total crash count data in Kano state, Nigeria were compared. Human and vehicle factors, including speed violations, tyre bursts, brake failures, sign/light violations and phone use while driving, were incorporated as covariates in the MLSTM model. An autoregressive integrated moving average with exogenous variables (Arimax) model was used to investigate the effects of the covariates. The MLSTM model outperformed both the basic LSTM model and individual covariate models, emphasising the synergistic effect of considering a broad range of factors. The Arimax model results revealed that speed violation is significantly positively correlated with total crashes; the other covariates showed positive correlations but did not reach the statistical significance. The findings underscore the importance of a multivariate approach in enhancing traffic crash prediction. The MLSTM model's superior performance highlights the value of considering a comprehensive range of factors that influence crash occurrence to achieve more accurate predictions. Practical applications of these models could involve leveraging them for proactive traffic safety measures, including increased enforcement of traffic rules, improvements to road infrastructure and targeted driver education and campaigns.

Vancomycin trough concentration in adult patients with periprosthetic joint infection: A machine learning-based covariate model.

Although there are various model-based approaches to individualized vancomycin (VCM) administration, few have been reported for adult patients with periprosthetic joint infection (PJI). This work attempted to develop a machine learning (ML)-based model for predicting VCM trough concentration in adult PJI patients. The dataset of 287 VCM trough concentrations from 130 adult PJI patients was split into a training set (229) and a testing set (58) at a ratio of 8:2, and an independent external 32 concentrations were collected as a validation set. A total of 13 covariates and the target variable (VCM trough concentration) were included in the dataset. A covariate model was respectively constructed by support vector regression, random forest regression and gradient boosted regression trees and interpreted by SHapley Additive exPlanation (SHAP). The SHAP plots visualized the weight of the covariates in the models, with estimated glomerular filtration rate and VCM daily dose as the 2 most important factors, which were adopted for the model construction. Random forest regression was the optimal ML algorithm with a relative accuracy of 82.8% and absolute accuracy of 67.2% (R2=.61, mean absolute error= 2.4, mean square error= 10.1), and its prediction performance was verified in the validation set. The proposed ML-based model can satisfactorily predict the VCM trough concentration in adult PJI patients. Its construction can be facilitated with only 2 clinical parameters (estimated glomerular filtration rate and VCM daily dose), and prediction accuracy can be rationalized by SHAP values, which highlights a profound practical value for clinical dosing guidance and timely treatment.

Vancomycin population pharmacokinetics and dosing proposal for the initial treatment in obese adult patients.

The aim of this study was to develop a vancomycin population pharmacokinetic model in adult obese patients and propose covariate-based dosing individualization in order to maximize the achievement of the newly recommended PK/PD target, according to a revised consensus guideline from 2020. Therapeutic drug monitoring data from initial vancomycin therapy (first 3days of treatment) in adult obese (BMI ≥ 30kg/m2) patients from 2013 to 2022 were analyzed using a non-linear mixed-effects modeling method, and Monte Carlo simulations were then used to find the optimal dosage maximizing the PK/PD target attainment. A total of 147 vancomycin serum levels obtained from 138 patients were included in the analysis. Based on the covariate model diagnosis among all tested variables, no reliable predictor of vancomycin volume of distribution (Vd) was identified, while clearance (CL) was positively correlated with eGFR and lean body mass. Creatinine-based eGFR predicted vancomycin CL better than cystatin C-based eGFR. The median (interquartile range) value from conditional modes of individual estimates of Vd, CL, and elimination half-life in our population was 74.0 (70.5-75.4) L, 6.65 (4.95-8.42) L/h, and 7.7 (6.0-10.0) h, respectively. We proposed dosing individualization based on the covariate found in order to maximize the achievement of the newly recommended PK/PD target of the AUC/MIC ratio of 400-600. Clinical pharmacy/pharmacology interventions may lead to an improvement in vancomycin dosing with a reflection in PK/PD target attainment.

Population Pharmacokinetics and Absolute Oral Bioavailability of Lasalocid after Single Intravenous and Intracrop Administration in Laying Hens.

Lasalocid sodium is a polyether carboxylic ionophore agent authorized by the EU for use as a coccidiostat in broilers, turkeys, and pullets up to 16 weeks of age, except for laying hens. However, laying hens are the most common nontarget species exposed to lasalocid sodium, mainly due to cross-contamination from feed mills. This exposure may result in potential drug residue deposition in eggs, which could potentially expose consumers to the drug. The breeds commonly used for commercial egg production in Poland are Isa Brown and Green-legged Partridge hens, which have been found to significantly differ in egg-laying performance. This variability may also affect the pharmacokinetics of lasalocid. Data on lasalocid plasma pharmacokinetics in laying hens are lacking. In this study, we aimed to determine typical population pharmacokinetic parameters, absolute oral bioavailability, and how breed may influence the pharmacokinetics of lasalocid. Twenty-layer hens of the two breeds were used in this study. Lasalocid was administered orally at a single dose of either 1 mg or 5 mg/kg body weight or intravenously at a dose of 1 mg/kg body weight, in a crossover design with a three-week washout period between study periods. Blood samples were collected for 72 h, and lasalocid concentrations were measured using high-performance liquid chromatography with fluorescence detection. A population pharmacokinetic analysis was conducted using nonlinear mixed effects modeling. Standard numerical and graphical criteria were used to select the best model, and a stepwise covariate modeling approach was used to determine any influencing factors. The best model was a three-compartment mammillary model with first-order absorption, transit compartments, and linear elimination. The estimated absolute oral bioavailability was low (36%). It was found that breed significantly influenced not only absorption but also the elimination of lasalocid. This study revealed that lasalocid absorption and elimination varied between the two breeds. This variability in pharmacokinetics may result in breed-related differences in drug residue accumulation in eggs, and ultimately, the risk associated with consumer exposure to drug residues may also vary.

Machine-Learning Assisted Screening of Correlated Covariates: Application to Clinical Data of Desipramine

Stepwise covariate modeling (SCM) has a high computational burden and can select the wrong covariates. Machine learning (ML) has been proposed as a screening tool to improve the efficiency of covariate selection, but little is known about how to apply ML on actual clinical data. First, we simulated datasets based on clinical data to compare the performance of various ML and traditional pharmacometrics (PMX) techniques with and without accounting for highly-correlated covariates. This simulation step identified the ML algorithm and the number of top covariates to select when using the actual clinical data. A previously developed desipramine population-pharmacokinetic model was used to simulate virtual subjects. Fifteen covariates were considered with four having an effect included. Based on the F1 score (an accuracy measure), ridge regression was the most accurate ML technique on 200 simulated datasets (F1 score = 0.475 ± 0.231), a performance which almost doubled when highly-correlated covariates were accounted for (F1 score = 0.860 ± 0.158). These performances were better than forwards selection with SCM (F1 score = 0.251 ± 0.274 and 0.499 ± 0.381 without/with correlations respectively). In terms of computational cost, ridge regression (0.42 ± 0.07 seconds/simulated dataset, 1 thread) was ~20,000 times faster than SCM (2.30 ± 2.29 hours, 15 threads). On the clinical dataset, prescreening with the selected ML algorithm reduced SCM runtime by 42.86% (from 1.75 to 1.00 days) and produced the same final model as SCM only. In conclusion, we have demonstrated that accounting for highly-correlated covariates improves ML prescreening accuracy. The choice of ML method and the proportion of important covariates (unknown a priori) can be guided by simulations.Graphical

Assessment of body mass-related covariates for rifampicin pharmacokinetics in healthy Caucasian volunteers

PurposeCurrently, body weight-based dosing of rifampicin is recommended. But lately, fat-free mass (FFM) was reported to be superior to body weight (BW). The present evaluation aimed to assess the influence of body mass-related covariates on rifampicin’s pharmacokinetics (PK) parameters in more detail using non-linear mixed effects modeling (NLMEM).MethodsTwenty-four healthy Caucasian volunteers were enrolled in a bioequivalence study, each receiving a test and a reference tablet of 600 mg of rifampicin separated by a wash-out period of at least 9 days. Monolix version 2023R1 was used for NLMEM. Monte Carlo simulations (MCS) were performed to visualize the relationship of body size descriptors to the exposure to rifampicin.ResultsA one-compartment model with nonlinear (Michaelis–Menten) elimination and zero-order absorption kinetics with a lag time best described the data. The covariate model including fat-free mass (FFM) on volume of distribution (V/F) and on maximum elimination rate (Vmax/F) lowered the objective function value (OFV) by 56.4. The second-best covariate model of sex on V/F and Vmax/F and BW on V/F reduced the OFV by 51.2. The decrease in unexplained inter-individual variability on Vmax/F in both covariate models was similar. For a given dose, MCS showed lower exposure to rifampicin with higher FFM and accordingly in males compared to females with the same BW and body height.ConclusionOur results indicate that beyond BW, body composition as reflected by FFM could also be relevant for optimized dosing of rifampicin. This assumption needs to be studied further in patients treated with rifampicin.

Population kinetic/pharmacodynamic modelling of the haemodynamic effects of cafedrine/theodrenaline (Akrinor) under general anaesthesia.

The 20:1 combination of cafedrine and theodrenaline (C/T) is widely used in Germany for the treatment of arterial hypotension. Since there is little knowledge about the impact of covariates on the effect, the aim was to develop a kinetic/pharmacodynamic covariate model describing mean arterial pressure (MAP), systolic (SBP) and diastolic blood pressure (DBP), and heart rate (HR) for 30 min after the administration of C/T. Data of patients receiving C/T from the HYPOTENS study (NCT02893241, DRKS00010740) were analysed using nonlinear mixed-effects modelling techniques. Overall, 16 579 measurements from 315 patients were analysed. The combination of two kinetic compartments and a delayed effect model, coupled with distinct Emax models for HR, SBP and DBP, described the data best. The model included age, sex, body mass index (BMI), antihypertensive medication, American Society of Anaesthesiologists (ASA) physical status classification grade, baseline SBP at the time of hypotension and pre-surgery HR as covariates (all P < .001). A higher baseline SBP led to a lower absolute increase in MAP. Patients with higher age, higher BMI and lower ASA grade showed smaller increases in MAP. The initial increase was similar for male and female patients. The long-term effect was higher in women. Concomitant antihypertensive medication caused a delayed effect and a lower maximum MAP. The HR increased only slightly (median increase 2.6bpm, P < .001). Seven covariates with an impact on the effect of C/T could be identified. The results will enable physicians to optimize the dose with respect to individual patients.

Serum klotho associated with thyroid hormone in adults: A population-based cross-sectional research.

The klotho protein, a multifunctional protein, has been shown to be associated with a wide range of endocrine diseases and has been linked to thyroid tumourigenesis. However, the relationship between serum klotho levels and thyroid hormones remains poorly understood. This study aimed to explore the correlation between serum klotho levels and thyroid hormones. Data was obtained from the NHANES cycles 2007-2008, 2009-2010, and 2011-2012. A total of 4674 participants were recruited for this study. Statistical analysis was using multiple linear regression analyses, and restricted cubic spline plots (RCS) to investigate the association between serum klotho levels and serum levels of thyroid hormones. In the unadjusted covariate model, ln(klotho) significantly positively correlated with tT3, tT4, fT3, tT4/fT4, and tT3/fT3 (all P<0.01) and negatively correlated with TSH, tT4/tT3, and fT4/fT3 (all P<0.05). Furthermore, tT3, tT4, fT3and tT3/fT3 (P < 0.05) were still significant in the adjusted model. And it is worth noting that there is an approximately L-shaped nonlinear relationship between ln(klotho) and fT3,tT3 with a cut-off point of 6.697 (P-non-linear < 0.05). The stratification analysis showed gender and iodine level differences in the relationship between serum Klotho levels and thyroid hormones. There is an L-shaped nonlinear relationship between ln(klotho) and fT3, tT3, suggesting that klotho could be involved in the physiological regulation of thyroid function.

Evaluation of ChatGPT and Gemini large language models for pharmacometrics with NONMEM.

To assess ChatGPT 4.0 (ChatGPT) and Gemini Ultra 1.0 (Gemini) large language models on NONMEM coding tasks relevant to pharmacometrics and clinical pharmacology. ChatGPT and Gemini were assessed on tasks mimicking real-world applications of NONMEM. The tasks ranged from providing a curriculum for learning NONMEM, an overview of NONMEM code structure to generating code. Prompts in lay language to elicit NONMEM code for a linear pharmacokinetic (PK) model with oral administration and a more complex model with two parallel first-order absorption mechanisms were investigated. Reproducibility and the impact of "temperature" hyperparameter settings were assessed. The code was reviewed by two NONMEM experts. ChatGPT and Gemini provided NONMEM curriculum structures combining foundational knowledge with advanced concepts (e.g., covariate modeling and Bayesian approaches) and practical skills including NONMEM code structure and syntax. ChatGPT provided an informative summary of the NONMEM control stream structure and outlined the key NONMEM Translator (NM-TRAN) records needed. ChatGPT and Gemini were able to generate code blocks for the NONMEM control stream from the lay language prompts for the two coding tasks. The control streams contained focal structural and syntax errors that required revision before they could be executed without errors and warnings. The code output from ChatGPT and Gemini was not reproducible, and varying the temperature hyperparameter did not reduce the errors and omissions substantively. Large language models may be useful in pharmacometrics for efficiently generating an initial coding template for modeling projects. However, the output can contain errors and omissions that require correction.

Sick leave and disability pension following delivery in women with systemic lupus erythematosus

Objective To investigate sickness benefits following delivery in mothers with systemic lupus erythematosus (SLE) and mothers without SLE. Method SLE and non-SLE mothers, matched by age and month of delivery, with a singleton liveborn (2004–2008), were identified from the Swedish Lupus Linkage cohort. Work loss (sum of sick leave and disability pension) was studied from 1 year prenatally to 3 years postpartum. Adjusted logistic regression models of covariates associated with > 30 days of work loss in the first and second years postpartum were estimated in SLE mothers. Results Among 130 SLE mothers and 440 non-SLE mothers, SLE mothers were more likely to have work loss from the prenatal year (42% vs 16%) to 3 years postpartum (49% vs 15%). In SLE mothers, work loss was on average 61 ± 112 days (mean ± sd) in the prenatal year and 38 ± 83 days in the first year postpartum, which increased to 71 ± 114 days in the third year postpartum. Having > 30 days of sick leave in the year of delivery [odds ratio (OR) 4.4, 95% confidence interval (CI) 1.5–12.9] and ≤ 12 years of education (OR 2.6, 95% CI 1.1–6.0) were associated with work loss in the first year postpartum. No covariates were associated with work loss in the second year postpartum. Conclusion SLE mothers more often had work loss in the prenatal year to 3 years postpartum compared to non-SLE mothers. Lower education and sick leave in the year of delivery were associated with a higher odds of work loss in the first year postpartum in SLE.

Statistical Assessment of Water-quality Parameters for Different Flow Regimes

Groundwater and drainage-water sustainability is crucial to water management. Controlling the interaction between the groundwater and drain-water salinity at the watershed level has only been the topic of a limited number of studies. This study statistically compares salt concentration rate across time and four different flow regimes. We sampled water from 2014 to 2018. The study used R software to analyze water-quality parameters utilizing PCA and indices. Multiple water-quality indices were used to test irrigation water. A linear model with TDS as the response was fitted to the data. Another two-factor repeated-measurement covariate model was built. Water-quality indices suggest that measured water from varied flow regimes in the study area is suitable for irrigation. Based on the statistical models, Tukey's adjustment pairwise demonstrates considerable salinity variations between drain water and surface water. However, the main effect of the flow regime on TDS is significant. This indicates that the TDS value varies depending on the flow regime, not on time plus the interactions between these regimes. Keywords: Total dissolved solids (TDSs), Groundwater, ANCOVA, Flow condition, Tile drain, PCA

Impact of covariate model building methods on their clinical relevance evaluation in population pharmacokinetic analyses: comparison of the full model, stepwise covariate model (SCM) and SCM+ approaches.

Covariate analysis in population pharmacokinetics is key for adjusting doses for patients. The main objective of this work was to compare the adequacy of various modeling approaches on covariate clinical relevance decision-making. The full model, stepwise covariate model (SCM) and SCM+ PsN algorithms were compared in a clinical trial simulation of a 383-patient population pharmacokinetic study mixing rich and sparse designs. A one-compartment model with first-order absorption was used. A base model including a body weight effect on CL/F and V/F and a covariate model including 4 additional covariates-parameters relationships were simulated. As for forest plots, ratios between covariates at a specific value and that of a typical individual were calculated with their 90% confidence interval (CI90) using standard errors. Covariates on CL, V and KA were considered relevant if their CI90 fell completely outside the reference area [0.8-1.2]. All approaches provided unbiased covariate ratio estimates. For covariates with a simulated effect, the 3 approaches correctly identify their clinical relevance. However, significant covariates were missed in up to 15% of cases with SCM/SCM+. For covariate with no simulated effects, the full model mainly identified them as non-relevant or with insufficient information while SCM/SCM+ mainly did not select them. SCM/SCM+ assume that non-selected covariates are non-relevant when it could be due to insufficient information, whereas the full model does not make this assumption and is faster. This study must be extended to other methods and completed by a more complex high-dimensional simulation framework.