Proving Efficiency of NPMLE and Identities

A method is given for proving efficiency of NPMLE directly linked to empirical process theory. The conditions in general are appropriate consistency of the NPMLE, differentiability of the model, differentiability of the parameter of interest, local convexity of the parameter space, and a Donsker class condition for the class of efficient influence functions obtained by varying the parameters. For the case that the model is linear in the parameter and the parameter space is convex, as with most nonparametric missing data models, we show that the method leads to an identity for the NPMLE which almost says that the NPMLE is efficient and provides us straightforwardly with a consistency and efficiency proof. This identify is extended to an almost linear class of models which contain biased sampling models. To illustrate, the method is applied to the univariate censoring model, random truncation models, interval censoring case I model, the class of parametric models and to a class of semiparametric models.

MULTIPLE MODEL EVALUATION ABSENT THE GOLD STANDARD VIA MODEL COMBINATION

We describe a method for evaluating an ensemble of predictive models given a sample of observations comprising the model predictions and the outcome event measured with error. Our formulation allows us to simultaneously estimate measurement error parameters, true outcome — aka the gold standard — and a relative weighting of the predictive scores. We describe conditions necessary to estimate the gold standard and for these estimates to be calibrated and detail how our approach is related to, but distinct from, standard model combination techniques. We apply our approach to data from a study to evaluate a collection of BRCA1/BRCA2 gene mutation prediction scores. In this example, genotype is measured with error by one or more genetic assays. We estimate true genotype for each individual in the dataset, operating characteristics of the commonly used genotyping procedures and a relative weighting of the scores. Finally, we compare the scores against the gold standard genotype and find that Mendelian scores are, on average, the more refined and better calibrated of those considered and that the comparison is sensitive to measurement error in the gold standard.

Comparison of Doppler-derived aortic velocities obtained from various transducer sites in healthy dogs and cats

In normal dogs and dogs with subaortic stenosis, it is known that the subcostal transducer site provides higher left ventricular ejection velocities than does the left apical site. We hypothesized that aortic flow velocities could also be obtained from the right parasternal long-axis view, optimized for the placement of the Doppler cursor as parallel as possible into the aortic root. In 15 healthy dogs and 13 healthy cats, high-pulsed repetition frequency Doppler flow velocity measurements in the proximal aorta were performed using two-dimensional echocardiographic guidance. The mean [ +/- standard error of the mean (SEM)] peak aortic flow velocities in healthy dogs were as follows: subcostal site 1.46 +/- 0.05 m/s; apical site 1.12 +/- 0.06 m/s; right parasternal long-axis site 1.09 +/- 0.05 m/s. In healthy cats, the following peak aortic flow velocities were observed: apical site 0.87 +/- 0.03m/s; right parasternal long-axis site 0.87 +/- 0.03 m/s. Aortic flow velocities obtained from the subcostal site were significantly higher in healthy dogs than those obtained from the left apical and right parasternal long-axis site (P< 0.001). There was no statistical difference between the peak aortic flow velocities obtained from right parasternal long-axis and left apical transducer position in all groups. We conclude therefore that right parasternal long-axis and left apical-derived aortic flow velocities are similar and may be used interchangeably in healthy dogs and cats.

Clinical and microbiological analysis of subjects treated with Brånemark or AstraTech implants: a 7-year follow-up study

AIMS: To assess the impact of different implant systems on the clinical conditions and the microbiota at implants, and whether the presence of bacteria at tooth sites was predictive of the presence at implant sites. MATERIALS AND METHODS: Subjects with either AstraTech or Brånemark in function for 7 years were enrolled. Sub-gingival bacterial samples at tooth and implant sites were collected with sterile endodontic paper points, and analyzed by the checkerboard DNA-DNA hybridization method (40 species). RESULTS: Fifty-four subjects, 27 supplied with AstraTech (n=132 implants) and 27 with Brånemark (n=102) implants, were studied. Test tooth sites had significantly less evidence of bleeding on probing (P<0.001) and presence of plaque (P<0.001) than implant test sites. Implant sites presented with deeper probing pocket depth than tooth sites (mean difference: 1.1 mm, standard error of differences: 0.08, 95% confidence intervals (CI): 0.9-1.3, P<0.001). Tannerella forsythia (P<0.05), Capnocytophaga sputigena (P<0.05), Actinomyces israelii (P<0.05) and Lactobacillus acidophilus (P<0.05) were found at higher levels at tooth surfaces. No differences in bacterial load for any species were found between the two implant systems. The odds of being present/absent at tooth and implants sites were only significant for Staphylococcus aureus [odds ratio (OR): 5.2 : 1, 95% CI: 1.4-18.9, P<0.01]. CONCLUSIONS: After 7 years in function, implants presented with deeper probing depths than teeth. S. aureus was commonly present at both teeth and implants sites. S. aureus at tooth sites was predictive of also being present at implant sites.

Algorithms for spacecraft formation flying navigation based on wireless positioning system measurements

Spacecraft formation flying navigation continues to receive a great deal of interest. The research presented in this dissertation focuses on developing methods for estimating spacecraft absolute and relative positions, assuming measurements of only relative positions using wireless sensors. The implementation of the extended Kalman filter to the spacecraft formation navigation problem results in high estimation errors and instabilities in state estimation at times. This is due tp the high nonlinearities in the system dynamic model. Several approaches are attempted in this dissertation aiming at increasing the estimation stability and improving the estimation accuracy. A differential geometric filter is implemented for spacecraft positions estimation. The differential geometric filter avoids the linearization step (which is always carried out in the extended Kalman filter) through a mathematical transformation that converts the nonlinear system into a linear system. A linear estimator is designed in the linear domain, and then transformed back to the physical domain. This approach demonstrated better estimation stability for spacecraft formation positions estimation, as detailed in this dissertation. The constrained Kalman filter is also implemented for spacecraft formation flying absolute positions estimation. The orbital motion of a spacecraft is characterized by two range extrema (perigee and apogee). At the extremum, the rate of change of a spacecraft’s range vanishes. This motion constraint can be used to improve the position estimation accuracy. The application of the constrained Kalman filter at only two points in the orbit causes filter instability. Two variables are introduced into the constrained Kalman filter to maintain the stability and improve the estimation accuracy. An extended Kalman filter is implemented as a benchmark for comparison with the constrained Kalman filter. Simulation results show that the constrained Kalman filter provides better estimation accuracy as compared with the extended Kalman filter. A Weighted Measurement Fusion Kalman Filter (WMFKF) is proposed in this dissertation. In wireless localizing sensors, a measurement error is proportional to the distance of the signal travels and sensor noise. In this proposed Weighted Measurement Fusion Kalman Filter, the signal traveling time delay is not modeled; however, each measurement is weighted based on the measured signal travel distance. The obtained estimation performance is compared to the standard Kalman filter in two scenarios. The first scenario assumes using a wireless local positioning system in a GPS denied environment. The second scenario assumes the availability of both the wireless local positioning system and GPS measurements. The simulation results show that the WMFKF has similar accuracy performance as the standard Kalman Filter (KF) in the GPS denied environment. However, the WMFKF maintains the position estimation error within its expected error boundary when the WLPS detection range limit is above 30km. In addition, the WMFKF has a better accuracy and stability performance when GPS is available. Also, the computational cost analysis shows that the WMFKF has less computational cost than the standard KF, and the WMFKF has higher ellipsoid error probable percentage than the standard Measurement Fusion method. A method to determine the relative attitudes between three spacecraft is developed. The method requires four direction measurements between the three spacecraft. The simulation results and covariance analysis show that the method’s error falls within a three sigma boundary without exhibiting any singularity issues. A study of the accuracy of the proposed method with respect to the shape of the spacecraft formation is also presented.

Coagulation activity before and after acute psychosocial stress increases with age

OBJECTIVE: To assess whether stress further increases hypercoagulation in older individuals. We investigated whether acute stress-induced changes in coagulation parameters differ with age. It is known that hypercoagulation occurs in response to acute stress and that a shift in hemostasis toward a hypercoagulability state occurs with age. However, it is not yet known whether acute stress further increases hypercoagulation in older individuals, and thus may increase their risk for cardiovascular disease (CVD). METHODS: A total of 63 medication-free nonsmoking men, aged between 20 and 65 years (mean +/- standard error of the mean = 36.7 +/- 1.7 years), underwent an acute standardized psychosocial stress task combining public speaking and mental arithmetic in front of an audience. We measured plasma clotting factor VII activity (FVII:C), fibrinogen, and D-dimer at rest, immediately, and 20 minutes after stress. RESULTS: Increased age predicted greater increases in fibrinogen (beta = 0.26, p = 0.041; DeltaR(2) = 0.05), FVII:C (beta = 0.40, p = .006; DeltaR(2) = 0.11), and D-dimer (beta = 0.51, p < .001; DeltaR(2) = 0.18) from rest to 20 minutes after stress independent of body mass index and mean arterial blood pressure. General linear models revealed significant effects of age and stress on fibrinogen, FVII:C, and D-dimer (main effects: p < .04), and greater D-dimer stress reactivity with older age (interaction age-by-stress: F(1.5/90.4) = 4.36, p = .024; f = 0.33). CONCLUSIONS: Our results suggest that acute stress might increase vulnerability in the elderly for hypercoagulability and subsequent hemostasis-associated diseases like CVD.

Association between statin-associated myopathy and skeletal muscle damage

BACKGROUND: Many patients taking statins often complain of muscle pain and weakness. The extent to which muscle pain reflects muscle injury is unknown. METHODS: We obtained biopsy samples from the vastus lateralis muscle of 83 patients. Of the 44 patients with clinically diagnosed statin-associated myopathy, 29 were currently taking a statin, and 15 had discontinued statin therapy before the biopsy (minimal duration of discontinuation 3 weeks). We also included 19 patients who were taking statins and had no myopathy, and 20 patients who had never taken statins and had no myopathy. We classified the muscles as injured if 2% or more of the muscle fibres in a biopsy sample showed damage. Using reverse transcriptase polymerase chain reaction, we evaluated the expression levels of candidate genes potentially related to myocyte injury. RESULTS: Muscle injury was observed in 25 (of 44) patients with myopathy and in 1 patient without myopathy. Only 1 patient with structural injury had a circulating level of creatine phosphokinase that was elevated more than 1950 U/L (10x the upper limit of normal). Expression of ryanodine receptor 3 was significantly upregulated in patients with biopsy evidence of structural damage (1.7, standard error of the mean 0.3). INTERPRETATION: Persistent myopathy in patients taking statins reflects structural muscle damage. A lack of elevated levels of circulating creatine phosphokinase does not rule out structural muscle injury. Upregulation of the expression of ryanodine receptor 3 is suggestive of an intracellular calcium leak.

Independent association between lower level of social support and higher coagulation activity before and after acute psychosocial stress

OBJECTIVE: To investigate the relationship between social support and coagulation parameter reactivity to mental stress in men and to determine if norepinephrine is involved. Lower social support is associated with higher basal coagulation activity and greater norepinephrine stress reactivity, which in turn, is linked with hypercoagulability. However, it is not known if low social support interacts with stress to further increase coagulation reactivity or if norepinephrine affects this association. These findings may be important for determining if low social support influences thrombosis and possible acute coronary events in response to acute stress. We investigated the relationship between social support and coagulation parameter reactivity to mental stress in men and determined if norepinephrine is involved. METHODS: We measured perceived social support in 63 medication-free nonsmoking men (age (mean +/- standard error of the mean) = 36.7 +/- 1.7 years) who underwent an acute standardized psychosocial stress task combining public speaking and mental arithmetic in front of an audience. We measured plasma D-dimer, fibrinogen, clotting Factor VII activity (FVII:C), and plasma norepinephrine at rest as well as immediately after stress and 20 minutes after stress. RESULTS: Independent of body mass index, mean arterial pressure, and age, lower social support was associated with higher D-dimer and fibrinogen levels at baseline (p < .012) and with greater increases in fibrinogen (beta = -0.36, p = .001; DeltaR(2) = .12), and D-dimer (beta = -0.21, p = .017; DeltaR(2) = .04), but not in FVII:C (p = .83) from baseline to 20 minutes after stress. General linear models revealed significant main effects of social support and stress on fibrinogen, D-dimer, and norepinephrine (p < .035). Controlling for norepinephrine did not change the significance of the reported associations between social support and the coagulation measures D-dimer and fibrinogen. CONCLUSIONS: Our results suggest that lower social support is associated with greater coagulation activity before and after acute stress, which was unrelated to norepinephrine reactivity.

Vitamin D and central hypersensitivity in patients with chronic pain.

BACKGROUND Low vitamin D is implicated in various chronic pain conditions with, however, inconclusive findings. Vitamin D might play an important role in mechanisms being involved in central processing of evoked pain stimuli but less so for spontaneous clinical pain. OBJECTIVE This study aims to examine the relation between low serum levels of 25-hydroxyvitamin D3 (25-OH D) and mechanical pain sensitivity. DESIGN We studied 174 patients (mean age 48 years, 53% women) with chronic pain. A standardized pain provocation test was applied, and pain intensity was rated on a numerical analogue scale (0-10). The widespread pain index and symptom severity score (including fatigue, waking unrefreshed, and cognitive symptoms) following the 2010 American College of Rheumatology preliminary diagnostic criteria for fibromyalgia were also assessed. Serum 25-OH D levels were measured with a chemiluminescent immunoassay. RESULTS Vitamin deficiency (25-OH D < 50 nmol/L) was present in 71% of chronic pain patients; another 21% had insufficient vitamin D (25-OH D < 75 nmol/L). After adjustment for demographic and clinical variables, there was a mean ± standard error of the mean increase in pain intensity of 0.61 ± 0.25 for each 25 nmol/L decrease in 25-OH D (P = 0.011). Lower 25-OH D levels were also related to greater symptom severity (r = -0.21, P = 0.008) but not to the widespread pain index (P = 0.83) and fibromyalgia (P = 0.51). CONCLUSIONS The findings suggest a role of low vitamin D levels for heightened central sensitivity, particularly augmented pain processing upon mechanical stimulation in chronic pain patients. Vitamin D seems comparably less important for self-reports of spontaneous chronic pain.

ROBREG: Stata module providing robust regression estimators

robreg provides a number of robust estimators for linear regression models. Among them are the high breakdown-point and high efficiency MM-estimator, the Huber and bisquare M-estimator, and the S-estimator, each supporting classic or robust standard errors. Furthermore, basic versions of the LMS/LQS (least median of squares) and LTS (least trimmed squares) estimators are provided. Note that the moremata package, also available from SSC, is required.

Development and validation of a novel pedometer algorithm to quantify extended characteristics of the locomotor behavior of dairy cows

Behavior is one of the most important indicators for assessing cattle health and well-being. The objective of this study was to develop and validate a novel algorithm to monitor locomotor behavior of loose-housed dairy cows based on the output of the RumiWatch pedometer (ITIN+HOCH GmbH, Fütterungstechnik, Liestal, Switzerland). Data of locomotion were acquired by simultaneous pedometer measurements at a sampling rate of 10 Hz and video recordings for manual observation later. The study consisted of 3 independent experiments. Experiment 1 was carried out to develop and validate the algorithm for lying behavior, experiment 2 for walking and standing behavior, and experiment 3 for stride duration and stride length. The final version was validated, using the raw data, collected from cows not included in the development of the algorithm. Spearman correlation coefficients were calculated between accelerometer variables and respective data derived from the video recordings (gold standard). Dichotomous data were expressed as the proportion of correctly detected events, and the overall difference for continuous data was expressed as the relative measurement error. The proportions for correctly detected events or bouts were 1 for stand ups, lie downs, standing bouts, and lying bouts and 0.99 for walking bouts. The relative measurement error and Spearman correlation coefficient for lying time were 0.09% and 1; for standing time, 4.7% and 0.96; for walking time, 17.12% and 0.96; for number of strides, 6.23% and 0.98; for stride duration, 6.65% and 0.75; and for stride length, 11.92% and 0.81, respectively. The strong to very high correlations of the variables between visual observation and converted pedometer data indicate that the novel RumiWatch algorithm may markedly improve automated livestock management systems for efficient health monitoring of dairy cows.

Development of a Tumour Growth Inhibition Model to Elucidate the Effects of Ritonavir on Intratumoural Metabolism and Anti-tumour Effect of Docetaxel in a Mouse Model for Hereditary Breast Cancer

In a mouse tumour model for hereditary breast cancer, we previously explored the anti-cancer effects of docetaxel, ritonavir and the combination of both and studied the effect of ritonavir on the intratumoural concentration of docetaxel. The objective of the current study was to apply pharmacokinetic (PK)-pharmacodynamic (PD) modelling on this previous study to further elucidate and quantify the effects of docetaxel when co-administered with ritonavir. PK models of docetaxel and ritonavir in plasma and in tumour were developed. The effect of ritonavir on docetaxel concentration in the systemic circulation of Cyp3a knock-out mice and in the implanted tumour (with inherent Cyp3a expression) was studied, respectively. Subsequently, we designed a tumour growth inhibition model that included the inhibitory effects of both docetaxel and ritonavir. Ritonavir decreased docetaxel systemic clearance with 8% (relative standard error 0.4%) in the co-treated group compared to that in the docetaxel only-treated group. The docetaxel concentration in tumour tissues was significantly increased by ritonavir with mean area under the concentration-time curve 2.5-fold higher when combined with ritonavir. Observed tumour volume profiles in mice could be properly described by the PK/PD model. In the co-treated group, the enhanced anti-tumour effect was mainly due to increased docetaxel tumour concentration; however, we demonstrated a small but significant anti-tumour effect of ritonavir addition (p value <0.001). In conclusion, we showed that the increased anti-tumour effect observed when docetaxel is combined with ritonavir is mainly caused by enhanced docetaxel tumour concentration and to a minor extent by a direct anti-tumour effect of ritonavir.

Heritability of ambulatory and office blood pressure in the Swiss population

BACKGROUND Blood pressure (BP) is known to aggregate in families. Yet, heritability estimates are population-specific and no Swiss data have been published so far. We estimated the heritability of ambulatory and office BP in a Swiss population-based sample. METHODS The Swiss Kidney Project on Genes in Hypertension is a population-based family study focusing on BP genetics. Office and ambulatory BP were measured in 1009 individuals from 271 nuclear families. Heritability was estimated for SBP, DBP, and pulse pressure using a maximum likelihood method implanted in the Statistical Analysis in Genetic Epidemiology software. RESULTS The 518 women and 491 men included in this analysis had a mean (±SD) age of 48.3 (±17.4) and 47.3 (±17.7) years, and a mean BMI of 23.8 (±4.2) and 25.9 (±4.1) kg/m, respectively. Narrow-sense heritability estimates (±standard error) for ambulatory SBP, DBP, and pulse pressure were 0.37 ± 0.07, 0.26 ± 0.07, and 0.29 ± 0.07 for 24-h BP; 0.39 ± 0.07, 0.28 ± 0.07, and 0.27 ± 0.07 for day BP; and 0.25 ± 0.07, 0.20 ± 0.07, and 0.30 ± 0.07 for night BP, respectively (all P < 0.001). Heritability estimates for office SBP, DBP, and pulse pressure were 0.21 ± 0.08, 0.25 ± 0.08, and 0.18 ± 0.07 (all P < 0.01). CONCLUSIONS We found significant heritability estimates for both ambulatory and office BP in this Swiss population-based study. Our findings justify the ongoing search for the genetic determinants of BP.

Multiple imputation for missing continuous and dichotomous data: The effects of patterns of missingness and variable correlations on type I error with small samples

The purpose of this study is to investigate the effects of predictor variable correlations and patterns of missingness with dichotomous and/or continuous data in small samples when missing data is multiply imputed. Missing data of predictor variables is multiply imputed under three different multivariate models: the multivariate normal model for continuous data, the multinomial model for dichotomous data and the general location model for mixed dichotomous and continuous data. Subsequent to the multiple imputation process, Type I error rates of the regression coefficients obtained with logistic regression analysis are estimated under various conditions of correlation structure, sample size, type of data and patterns of missing data. The distributional properties of average mean, variance and correlations among the predictor variables are assessed after the multiple imputation process. ^ For continuous predictor data under the multivariate normal model, Type I error rates are generally within the nominal values with samples of size n = 100. Smaller samples of size n = 50 resulted in more conservative estimates (i.e., lower than the nominal value). Correlation and variance estimates of the original data are retained after multiple imputation with less than 50% missing continuous predictor data. For dichotomous predictor data under the multinomial model, Type I error rates are generally conservative, which in part is due to the sparseness of the data. The correlation structure for the predictor variables is not well retained on multiply-imputed data from small samples with more than 50% missing data with this model. For mixed continuous and dichotomous predictor data, the results are similar to those found under the multivariate normal model for continuous data and under the multinomial model for dichotomous data. With all data types, a fully-observed variable included with variables subject to missingness in the multiple imputation process and subsequent statistical analysis provided liberal (larger than nominal values) Type I error rates under a specific pattern of missing data. It is suggested that future studies focus on the effects of multiple imputation in multivariate settings with more realistic data characteristics and a variety of multivariate analyses, assessing both Type I error and power. ^