Nonlinear Mixed Model Analysis Research Articles

Performance of specific morphologic features in distinguishing low-grade squamous intraepithelial lesions from high-grade squamous intraepithelial lesions in borderline cases: a College of American Pathologists Cytopathology Committee multiobserver study.

Distinguishing between low-grade squamous intraepithelial lesions (LSIL) and high-grade squamous intraepithelial lesions (HSIL) can be difficult on certain Papanicolaou (Pap) tests, hindering interobserver concordance. We investigated the variables influencing the interpretation of LSIL versus HSIL in Pap test slides rejected from the College of American Pathologists PAP education program. Eleven cytologists, who were unaware of the reference interpretation, examined 21 Pap slides (11 submitted as LSIL and 10 as HSIL) rejected from the PAP education program and recorded the number of LSIL cells, HSIL cells, keratinized dysplastic cells, LSIL clusters with mixed HSIL cells, atypical squamous metaplasia, atypical glandular cells, the presence of inflammation or infectious organisms, and the overall interpretation (LSIL or HSIL). We evaluated the significance of these 11 variables using a nonlinear mixed model analysis. LSIL had greater concordance (92 of 121 responses; 76.0% concordance) than HSIL (68 of 110 responses; 61.8% concordance; P < 0.001). The only predictors of misclassified cases were the number of atypical squamous metaplastic cells and the number of HSIL cells (P < 0.001). The more of these cells identified, the more likely the reviewers were to classify the slide as HSIL. The reproducibility of the diagnosis was fair (Gwet's agreement coefficient, 0.33). Interobserver reproducibility is a challenge for a subset of cases with features intermediate between LSIL and HSIL. Atypical squamous metaplasia and dysplastic nuclei with a nuclear/cytoplasmic ratio greater than one half of the cell volume (HSIL) present on a Pap test influenced the likelihood that a reviewer would interpret the case as HSIL rather than LSIL.

Development of a decision-making application for optimum soybean and maize fertilization strategies in Mato Grosso

Mato Grosso accounts for 25 and 31% of soybean and maize grain production in Brazil, respectively. Despite the importance of this region, there is limited information to optimize fertilization management decisions for these crops. Our objectives were to i) quantify soybean and maize yield response to fertilization, ii) develop prediction algorithms to prescribe recommendations based on site characteristics, and iii) incorporate these algorithms in a decision-making application. A total of 37 soybean and 27 maize field trials (year by site combinations) were conducted to test yield responses of soybean to phosphorus (P) and potassium (K), and maize responses to nitrogen (N) and K. Soybean and maize showed significant yield increments to applied P and N, respectively, but K fertilization showed minor yield effects in both crops (and was not considered further). Yield responses to P and N in soybean and maize, respectively, were highly variable across environments, and non-linear mixed model analysis was used to test co-variates to optimize yield response predictions. Including soil pH and initial soil P for soybean, and planting date and soil clay for maize, helped reduce a significant fraction of the site-to-site random variation. Soybean yield response to P increased in soils with higher pH values, agreeing with the concept of more mineral P retention in acidic soils. Maize showed lower N fertilization yield responses in later plantings dates, agreeing with higher water stress typical of late plantings. Out-of-sample relative root mean square error (rRMSE) of predictions were 15.9 and 19.5% for soybean and maize, respectively, showing the model has adequate accuracy to predict crop yield responses to nutrient fertilizations across the region. These algorithms were incorporated in a user-friendly web application to facilitate utilization, and dynamically calculates optimum fertilization rates including economic inputs such as grain and fertilizer prices. Our field results provide basic information to optimize fertilization rates for soybean and maize in Mato Grosso and are available for management decisions through a web application.

Obesity and fatty liver are prevented by inhibition of the aryl hydrocarbon receptor in both female and male mice

Inhibition of the aryl hydrocarbon receptor (AHR) prevents Western diet–induced obesity and fatty liver in C57Bl/6J (B6) male mice. The AHR is a ligand-activated nuclear receptor that regulates genes involved in xenobiotic metabolism and T-cell differentiation. Here, we tested the hypothesis that AHR antagonism would also prevent obesity and fatty liver in female mice and that B6 mice (higher-affinity AHR) and congenic B6.D2 mice (lower-affinity AHR) would differentially respond to AHR inhibition. Female and male adult B6 and B6.D2 mice were fed control and Western diets with and without α-naphthoflavone (NF), an AHR inhibitor. A nonlinear mixed-model analysis was developed to project asymptote body mass. We found that obesity, adiposity, and liver steatosis were reduced to near control levels in all female and male B6 and B6.D2 experimental groups fed Western diet with NF. However, differences were noted in that female B6.D2 vs B6 mice on Western diet became more obese; and in general, female mice compared with male mice had a greater fat mass to body mass ratio, were less responsive to NF, and had reduced liver steatosis and hepatomegaly. We report that male mice fed Western diet containing NF or CH-223191, another AHR inhibitor, caused reduced mRNA levels of several liver genes involved in metabolism, including Cyp1b1 and Scd1, offering evidence for a possible mechanism by which the AHR regulates obesity. In conclusion, although there are some sex- and Ahr allelic-dependent differences, AHR inhibition prevents obesity and liver steatosis in both males and females regardless of the ligand-binding capacity of the AHR. We also present evidence consistent with the notion that an AHR-CYP1B1-SCD1 axis is involved in obesity, providing potentially convenient and effective targets for treatment.

Linear and nonlinear mixed model analyses of growth performance of commercial U.S. dairy buck kids

The objectives of this research were: (1) to describe the growth of crossbred, intact, dairy buck kids from shortly after weaning to slaughter weight and (2) to determine whether raising them can be profitable. The kids in this study were purchased from three commercial dairy goat farms and were fed to slaughter weight in two separate environments. Individual kid body weight (BW) and pen feed disappearance was recorded. Both a linear mixed model and a nonlinear Gompertz mixed model were implemented to describe BW over the time of the feeding trial. The fits of the models were compared by −2 log likelihood, Akaike information criterion, and Bayesian information criterion. All fit statistics favored the nonlinear model, though upon visual inspection, growth was near linear. Although groups of kids were sold at varying weights and for different prices, revenues over kid purchase price and feed costs were positive. Feeding cull dairy kids can be profitable under many production scenarios.

Analysis of Reliability Experiments with Random Blocks and Subsampling

Reliability experiments provide important information regarding the life of an item, including how various factors affect item life. Current analyses of reliability data usually assume a completely randomized design. However, reliability experiments are frequently executed with restrictions on randomization. Common restrictions on randomization include subsampling and blocking. This paper proposes a nonlinear mixed-model analysis for reliability experiments with random blocks and subsampling. A simulation study compares this nonlinear mixed model analysis with a two-stage analysis and traditional reliability analyses.

Short communication: Prediction of intake in dairy cows under tropical conditions

A meta-analysis was conducted to develop a model for predicting dry matter intake (DMI) in dairy cows under the tropical conditions of Brazil and to assess its adequacy compared with 5 currently available DMI prediction models: Agricultural and Food Research Council (AFRC); National Research Council (NRC); Cornell Net Carbohydrate and Protein System (CNCPS; version 6); and 2 other Brazilian models. The data set was created using 457 observations (n=1,655 cows) from 100 studies, and it was randomly divided into 2 subsets for statistical analysis. The first subset was used to develop a DMI prediction equation (60 studies; 309 treatment means) and the second subset was used to assess the adequacy of DMI predictive models (40 studies; 148 treatment means). The DMI prediction model proposed in the current study was developed using a nonlinear mixed model analysis after reparameterizing the NRC equation but including study as a random effect in the model. Body weight (mean=540±57.6kg), 4% fat-corrected milk (mean=21.3±7.7kg/d), and days in milk (mean=110±62d) were used as independent variables in the model. The adequacy of the DMI prediction models was evaluated based on coefficient of determination, mean square prediction error (MSPE), root MSPE (RMSPE), and concordance correlation coefficient (CCC). The observed DMI obtained from the data set used to evaluate the prediction models averaged 17.6±3.2kg/d. The following model was proposed: DMI (kg/d)=[0.4762 (±0.0358) × 4% fat-corrected milk + 0.07219 (±0.00605) × body weight0.75] × (1 – e−0.03202 (±0.00615) × [days in milk + 24.9576 (±5.909)]). This model explained 93.0% of the variation in DMI, predicting it with the lowest mean bias (0.11kg/d) and RMSPE (4.9% of the observed DMI) and the highest precision [correlation coefficient estimate (ρ)=0.97] and accuracy [bias correction factor (Cb)=0.99]. The NRC model prediction equation explained 92.0% of the variation in DMI and had the second lowest mean bias (0.42kg/d) and RMSPE (5.8% of the observed DMI), as well as the second highest precision (ρ=0.94) and accuracy (Cb=0.98). The CNCPS and AFRC DMI prediction models explained 93.0 and 85.0% of the variation in DMI but underpredicted DMI by 1.8 and 1.4kg/d, respectively. These 2 models (CNCPS and AFRC) resulted, respectively, in RMSPE of 11.3 and 10.7% of the observed DMI, with moderate to high precision (ρ=0.81 and 0.82) and accuracy (Cb=0.84 and 0.89). The remaining 2 models resulted in the poorest results, underpredicting DMI by 2.3 and 1.9kg/d, with RMSPE of 22.8 and 14.9% of the observed DMI and moderate to low precision (ρ=0.49 and 0.76) and accuracy (Cb=0.81 and 0.86). The new model derived from the current meta-analytical approach provided the best accuracy and precision for predicting DMI in lactating dairy cows under Brazilian conditions.

Primary Pulmonary Non–Small Cell Carcinomas: The College of American Pathologists Interlaboratory Comparison Program Confirms a Significant Trend Toward Subcategorization Based Upon Fine-Needle Aspiration Cytomorphology Alone

Context.-Subtyping of non-small cell lung carcinomas (NSCLCs) is necessary for optimal patient management with specific diagnoses triggering specific molecular tests and affecting therapy. Objective.-To assess the accuracy of the participants of the College of American Pathologists Interlaboratory Comparison Program in diagnosing and subtyping NSCLC fine-needle aspiration (FNA) slides, based on morphology alone, considering preparation and participant type and trends over time. Design.-The performance of program participants was reviewed for the 5-year period spanning 2007-2011. Lung FNA challenges with reference diagnoses of adenocarcinoma and squamous cell carcinoma (SCC) were evaluated for diagnostic concordance by using a nonlinear mixed model analysis. Results.-There were 10 493 pathologist and 6378 cytotechnologist responses with concordance rates of 97.4% and 97.9% for malignancy, respectively. Overall concordance rates for subcategorization were 54.6% for adenocarcinoma and 74.9% for SCC. For the exact reference diagnoses, pathologists performed better for adenocarcinoma and cytotechnologists performed better for SCC. Accurate subcategorization of adenocarcinomas significantly increased over time with 31.5% of adenocarcinomas classified as NSCLC in 2007 and 25.5% of adenocarcinomas classified as NSCLC in 2011 (P < .001). In comparing preparation types, modified Giemsa-stained smears showed the lowest overall concordance (46.8%). Modified Giemsa-stained smears with SCCs were the least likely to be accurately subcategorized (36.4%). Conclusions.-Participants are proficient at interpreting NSCLCs as malignant by FNA but are less successful at subcategorization with cytomorphology alone. During the study period, a statistically significant trend was confirmed toward greater accuracy of subcategorization of adenocarcinomas, suggesting that participants are cognizant of the impact that more specific cytomorphologic interpretations have in directing molecular triage and therapy.

Investigating the appropriate mode of expressing lysine requirement of fish through non-linear mixed model analysis and multilevel analysis

Accurate estimates of lysine requirement are essential to fish feed formulation. However, controversy exists regarding the most appropriate mode to express lysine requirement. In the fish nutrition literature, essential amino acid (AA) requirement has been expressed as a percentage of diet, a percentage of dietary crude protein or a ratio to dietary digestible energy (DE). The controversy lies in the different assumptions regarding the effects of dietary protein and DE on lysine requirement. Non-linear mixed model analysis and multilevel analysis were carried out to investigate whether dietary protein or DE affected lysine requirement of fish. The non-linear mixed model analysis suggests that expressing lysine requirement as a percentage of dietary protein provides a better goodness of fit to the modelling dataset than expressing requirement as a fixed concentration of diet, which in turn is generally better than expressing requirement as a ratio to DE. Results from the multilevel analysis confirm that dietary protein content has a significant effect on lysine requirement, while DE does not. The findings of the present study could contribute to a better understanding of the underlying dietary factors that affect AA requirements of fish. The results of the present study could also be useful for developing nutritional guidelines and feed formulations for fish.

Modeling the efficiency of phosphorus utilization in growing pigs1

Microbial phytase has been used to reduce P excretion from swine to mitigate environmental pollution. The objective of the study was to quantify the effect of feeding a low-P phytase-supplemented diet on growth and P utilization in growing pigs using mathematical models. A total of 20 weaned piglets (BW = 6.5 kg) housed in metabolism cages were randomly assigned to a standard diet (STD) or P-amended diet containing reduced P content and supplemented with phytase (AMN) with 10 pigs/diet. Body weight and feed consumption were recorded weekly so complete growth and cumulative P intake (cPI) curves could be modeled. A function with fixed point of inflexion (Gompertz) and a variable point of inflexion (generalized Michaelis-Menten) were considered in determining bioequivalence by analyzing BW vs. age relationships, whereas the monomolecular function was used to describe BW vs. cPI. All functions were incorporated into a nonlinear mixed effects model, and a first-order autoregressive correlation structure was implemented to take into account repeated measures. There was no difference between the 2 groups in final BW when the Gompertz equation was fitted (176 vs. 178 kg with SE of 7 kg for the STD and AMN, respectively) or the rate parameter (0.0140 vs. 0.0139 with SE of 0.0004 for the STD and AMN, respectively). The generalized Michaelis-Menten equation also showed a similar trend. When BW was expressed as a function of cPI the derivative with respect to cPI represented P efficiency, so it was possible to analyze the expected difference of the 2 diets in using P for BW gain and express it as a continuous function of cPI. The analysis showed through the entire growth period the difference in P efficiency was different from zero. On average, 56 g of supplemented inorganic P was consumed by a pig fed the AMN to reach market weight. In contrast, 309 g of supplemented inorganic P was consumed by the group fed the STD to reach similar BW. It would depend on other factors, but feeding pigs the AMN can result in economic benefit. Pigs fed the AMN excreted 19% less P compared with those fed the STD. In conclusion, nonlinear mixed model analysis (with repeated measures) was suitable for growth and efficiency analysis and showed that pigs fed the AMN consumed less than 20% of the inorganic P and performed as well as those fed the traditional inorganic P supplemented diet. The implications for mitigating P pollution, especially in areas where P loading is already problematic, are substantial.

Predictors of perceived susceptibility of breast cancer and changes over time: a mixed modeling approach.

To examine predictors of perceived susceptibility to breast cancer and assess differences across three dependent measures. Annual surveys were completed by US women veterans (N = 3,758) participating in a repeat mammography intervention trial. Multivariable non-linear mixed model analyses examined individual- and group-level changes in perceived susceptibility to breast cancer. Three single-item measures of perceived susceptibility to breast cancer (percent risk, ordinal risk, and comparative risk likelihood). Predictors included demographic, health status, health behavior, affect, knowledge, and subjective norm variables. Breast symptoms and greater cancer worry increased perceived susceptibility for all three measures. Other predictors varied by dependent measure. Random change, indicating individual variability, was observed for percent risk only. Despite small model effect sizes, breast symptoms and cancer worry were consistent predictors and may be good targets for messages designed to influence women's perceived susceptibility to breast cancer. Researchers may benefit from using measures of perceived susceptibility with larger response scales, but additional measurement research is needed. Combining indicators of perceived susceptibility may be undesirable when different predictors are associated with different measures.

Individual differences in adjustment to spousal loss: A nonlinear mixed model analysis

The study of within-person change lies at the core of developmental research. Theory and empirical data suggest that many of these developmental processes are not linear. We describe a broad class of multilevel models that allows for nonlinear change — nonlinear mixed models. To demonstrate the utility of these models, we present a nonlinear mixed model analysis of adjustment to conjugal loss. Coming from a perspective of the individual as a regulatory system, our model predicts a faster rate of adjustment immediately following the loss and diminished adjustment as time since the loss increases, approaching an equilibrium level of well-being. This model allows us to estimate various aspects of the adjustment trajectory and individual differences in these trajectories, including multiple ways that pre- and post-loss factors can explain variability in the adjustment process. The model provides new insights into an important phenomenon that cannot be gleaned from linear models and other methods of trajectory analysis. We discuss the strengths and limitations of this type of analysis relative to other methods.

Conscious Sedation Reduces Distress in Children Undergoing Voiding Cystourethrography and Does Not Interfere with the Diagnosis of Vesicoureteric Reflux: A Randomized Controlled Study

Voiding cystourethrography (VCU) is a distressing procedure for children. Conscious sedation using oral midazolam may reduce this distress, but its use may also alter the ability of the VCU to show vesicoureteric reflux (VUR). The objectives of our study were to assess the effectiveness of conscious sedation using oral midazolam when administered routinely in children undergoing VCU and to ensure that conscious sedation using oral midazolam does not alter the ability of VCU to show VUR. Our study was a randomized double-blind controlled trial performed at a university teaching hospital; our study group consisted of children over the age of 1 year who been referred for their first VCU examination from July 2001 to July 2003. Participants were randomized to receive a placebo or midazolam syrup (0.5 mg/kg) before the examination. The primary outcome measures were the Groningen Distress Rating Scale (GDRS) and grading of VUR, as defined by the international grading system established by the International Reflux Study Group. There were no serious adverse events. One hundred thirty-nine children were randomized in the study, and 117 underwent complete assessment. Eight who underwent VCU after the study day were included in a "complete case" intention-to-treat analysis. In the placebo group, 34 children (61%) experienced serious distress or severe distress (GDRS score, 3 or 4). In the midazolam group, 16 children (26%) experienced the same degree of distress. There was a significant difference between the GDRS scores (nonlinear mixed-model analysis, p < 0.001) of the two study groups. The number needed to treat to reduce serious or severe distress in one child was 2.9 (95% CI, 1.9-5.5). VUR was identified in 16% of all children. There was no difference in VUR grading between the groups (nonlinear mixed-model analysis, p = 0.31). Routine use of oral midazolam (0.5 mg/kg) for conscious sedation of children undergoing VCU reduces distress and does not alter the ability of VCU to show VUR well enough to allow diagnosis.

Growth of Container-grown Cork Oak Seedlings as Affected by Foliar and Soil Application of Paclobutrazol

Four concentrations of paclobutrazol (PBZ) and two application methods (foliar spray–soil drench) were tested on 3-month-old container-grown cork oak (Quercus suber L.) seedlings. Shoot height was affected by PBZ concentration and application technique. Time course of height was modelled by nonlinear mixed model analysis. Reduced shoot height was the result of shortened and fewer internodes. A 25 to 625 ppm soil drench or 625 to 15,000 ppm foliar spray controlled over vegetative growth, resulting in a reduction between 13% and 90% in plant size relative to untreated controls. Seedlings responded to soil drench at a lower PBZ concentration than for a foliar spray, but the danger of irreversible overdosing seems greater for a soil drench. Soil drench—and foliar spray to a lesser extent—also affected other shoot and root growth traits. PBZ dose decreased root volume and dry weights in a quadratic fashion with increasing dose, but decreased diameter and number of leaves linearly.

Nonlinear Mixed Model Analyses of Five Production Disorders of Dairy Cattle

A standardized threshold model was used to estimate herd variability and the impact of some risk factors on the incidence of retained placenta, metritis, ketosis, ovulatory dysfunction, and mastitis. Data consisted of records for 18,110 Swedish Red and White cows in 924 herds and 14,940 Swedish Friesian cows in 772 herds. Only first, second, and third parity cows were included, and the breeds were analyzed separately. Herds were considered to be random effects, and variance components were estimated by the counterpart of REML for a threshold model. Estimated intraherd correlations varied from .04 to .18 and were lowest for the incidence of retained placenta (for both breeds) and highest for ketosis and ovulatory dysfunction for Swedish Red and White and Swedish Friesian breeds, respectively. Parity increased predicted lactational incidence risks for all disorders except metritis, but season of calving had no profound effect. Dystocia and stillbirth increased the risk of retained placenta; dystocia and retained placenta increased the risk of metritis. The risk of ketosis was not influenced by any of the considered risk factor diseases, and the risk of mastitis was influenced only by retained placenta for Swedish Red and White cows. For both breeds, the risk of ovulatory dysfunction was increased by metritis and ketosis.