Robust Factor Analysis for Compositional Data

Factor analysis as frequent technique for multivariate data inspection is widely used also for compositional data analysis. The usual way is to use a centered logratio (clr)transformation to obtain the random vector y of dimension D. The factor model istheny = Λf + e (1)with the factors f of dimension k & D, the error term e, and the loadings matrix Λ.Using the usual model assumptions (see, e.g., Basilevsky, 1994), the factor analysismodel (1) can be written asCov(y) = ΛΛT + ψ (2)where ψ = Cov(e) has a diagonal form. The diagonal elements of ψ as well as theloadings matrix Λ are estimated from an estimation of Cov(y).Given observed clr transformed data Y as realizations of the random vectory. Outliers or deviations from the idealized model assumptions of factor analysiscan severely effect the parameter estimation. As a way out, robust estimation ofthe covariance matrix of Y will lead to robust estimates of Λ and ψ in (2), seePison et al. (2003). Well known robust covariance estimators with good statisticalproperties, like the MCD or the S-estimators (see, e.g. Maronna et al., 2006), relyon a full-rank data matrix Y which is not the case for clr transformed data (see,e.g., Aitchison, 1986).The isometric logratio (ilr) transformation (Egozcue et al., 2003) solves thissingularity problem. The data matrix Y is transformed to a matrix Z by usingan orthonormal basis of lower dimension. Using the ilr transformed data, a robustcovariance matrix C(Z) can be estimated. The result can be back-transformed tothe clr space byC(Y ) = V C(Z)V Twhere the matrix V with orthonormal columns comes from the relation betweenthe clr and the ilr transformation. Now the parameters in the model (2) can beestimated (Basilevsky, 1994) and the results have a direct interpretation since thelinks to the original variables are still preserved.The above procedure will be applied to data from geochemistry. Our specialinterest is on comparing the results with those of Reimann et al. (2002) for the Kolaproject data

Kriging coordinates: what does that mean?

Kriging is an interpolation technique whose optimality criteria are based on normality assumptions either for observed or for transformed data. This is the case of normal, lognormal and multigaussian kriging.When kriging is applied to transformed scores, optimality of obtained estimators becomes a cumbersome concept: back-transformed optimal interpolations in transformed scores are not optimal in the original sample space, and vice-versa. This lack of compatible criteria of optimality induces a variety of problems in both point and block estimates. For instance, lognormal kriging, widely used to interpolate positivevariables, has no straightforward way to build consistent and optimal confidence intervals for estimates.These problems are ultimately linked to the assumed space structure of the data support: for instance, positive values, when modelled with lognormal distributions, are assumed to be embedded in the whole real space, with the usual real space structure and Lebesgue measure

Error and complexity of random walk Monte Carlo radiosity

The author studies the error and complexity of the discrete random walk Monte Carlo technique for radiosity, using both the shooting and gathering methods. The author shows that the shooting method exhibits a lower complexity than the gathering one, and under some constraints, it has a linear complexity. This is an improvement over a previous result that pointed to an O(n log n) complexity. The author gives and compares three unbiased estimators for each method, and obtains closed forms and bounds for their variances. The author also bounds the expected value of the mean square error (MSE). Some of the results obtained are also shown

Hospital volume and patient outcomes in pulmonary embolism.

BACKGROUND: In numerous high-risk medical and surgical conditions, a greater volume of patients undergoing treatment in a given setting or facility is associated with better survival. For patients with pulmonary embolism, the relation between the number of patients treated in a hospital (volume) and patient outcome is unknown. METHODS: We studied discharge records from 186 acute care hospitals in Pennsylvania for a total of 15 531 patients for whom the primary diagnosis was pulmonary embolism. The study outcomes were all-cause mortality in hospital and within 30 days after presentation for pulmonary embolism and the length of hospital stay. We used logistic models to study the association between hospital volume and 30-day mortality and discrete survival models to study the association between in-hospital mortality and time to hospital discharge. RESULTS: The median annual hospital volume for pulmonary embolism was 20 patients (interquartile range 10-42). Overall in-hospital mortality was 6.0%, whereas 30-day mortality was 9.3%. In multivariable analysis, very-high-volume hospitals (> or = 42 cases per year) had a significantly lower odds of in-hospital death (odds ratio [OR] 0.71, 95% confidence interval [CI] 0.51-0.99) and of 30-day death (OR 0.71, 95% CI 0.54-0.92) than very-low-volume hospitals (< 10 cases per year). Although patients in the very-high-volume hospitals had a slightly longer length of stay than those in the very-low-volume hospitals (mean difference 0.7 days), there was no association between volume and length of stay. INTERPRETATION: In hospitals with a high volume of cases, pulmonary embolism was associated with lower short-term mortality. Further research is required to determine the causes of the relation between volume and outcome for patients with pulmonary embolism.

Validation of a new tool to assess health-related quality of life in psoriasis: the PSO-LIFE questionnaire.

BACKGROUND Several questionnaires have been used to measure health related quality of life (HRQoL) in patients with psoriasis, few have been adapted for use in Spain; none of them was developed specifically for the Spanish population. The purpose of the study was to validate and assess the sensitivity to change of a new questionnaire to measure HRQOL in patients with psoriasis (PSO-LIFE). METHODS Observational, prospective, multicenter study performed in centers around Spain. Patients with active or inactive psoriasis completed the PSO-LIFE together with other Dermatology Quality of Life Index (DLQI) and Psoriasis Disability Index (PDI). A control group of patients with urticaria or atopic dermatitis was also included. Internal consistency and test-retest reliability of the PSO-LIFE were assessed by calculating Cronbach's alpha and Intraclass Correlation Coefficient (ICC). Validity was assessed by examining factorial structure, the capacity to discriminate between groups, and correlations with other measures. Sensitivity to change was measured using effect sizes. RESULTS The final sample included for analysis consisted of 304 patients and 56 controls. Mean (SD) age of psoriasis patients was 45.3 (14.5) years compared to 38.8 (14) years for controls (p < 0.01). Cronbach's alpha for the PSO-LIFE was 0.95 and test-retest reliability using the ICC was 0.98. Factor analysis showed the questionnaire to be unidimensional. Mean (SD) PSO-LIFE scores differed between patients with psoriasis and controls (64.9 [22.5] vs 69.4 [17.3]; p < 0.05), between those with active and inactive disease (57.4 [20.4] vs 76.4 [20.6]; p < 0.01), and between those with visible and non-visible lesions (63.0 [21.9] vs. 74.8 [23.9]; p < 0.01). The correlation between PSO-LIFE and PASI scores was moderate (r = -0.43) while correlations with DLQI and PDI dimensions ranged from moderate to high (between 0.4 and 0.8). Effect size on the PSO-LIFE in patients reporting 'much improved' health status at study completion was 1.01 (large effect size). CONCLUSIONS The present results provide substantial support for the reliability, validity, and responsiveness of the PSO-LIFE questionnaire in the population for which it was designed.

Amniotic fluid iodine concentrations do not vary in pregnant women with varying iodine intake.

Iodine deficiency is an important clinical and public health problem. Its prevention begins with an adequate intake of iodine during pregnancy. International agencies recommend at least 200 microg iodine per d for pregnant women. We assessed whether iodine concentrations in the amniotic fluid of healthy pregnant women are independent of iodine intake. This cross-sectional, non-interventional study included 365 consecutive women who underwent amniocentesis to determine the fetal karyotype. The amniocentesis was performed with abdominal antisepsis using chlorhexidine. The iodine concentration was measured in urine and amniotic fluid. The study variables were the intake of iodized salt and multivitamin supplements or the prescription of a KI supplement. The mean level of urinary iodine was 139.0 (SD 94.5) microg/l and of amniotic fluid 15.81 (SD 7.09) microg/l. The women who consumed iodized salt and those who took a KI supplement had significantly higher levels of urinary iodine than those who did not (P = 0.01 and P = 0.004, respectively). The urinary iodine levels were not significantly different in the women who took a multivitamin supplement compared with those who did not take this supplement, independently of iodine concentration or multivitamin supplement. The concentrations of iodine in the amniotic fluid were similar, independent of the dietary iodine intake. Urine and amniotic fluid iodine concentrations were weakly correlated, although the amniotic fluid values were no higher in those women taking a KI supplement. KI prescription at recommended doses increases the iodine levels in the mother without influencing the iodine levels in the amniotic fluid.

Estudio de los exitus con diagnóstico secundario de desnutrición en un hospital de tercer nivel.

BACKGROUND: Malnutrition is a major public health problems, according to WHO, is the leading cause of death, when it affects the group of hospitalized patients, making denominating separate entity "hospital malnutrition". OBJECTIVES: The overall objective is to quantify the main diagnoses frequently high, causing exitus, with secondary diagnosis of malnutrition. METHODS: This is a descriptive study, which included all hospital discharges in 2011 and first half of 2012, which have been exitus and whose secondary diagnosis of malnutrition, with the total of 33. We performed a descriptive analysis, effected the Mann-Whitney nonparametric test (p < 0.05). RESULTS: The most frequent main diagnoses among 33 analyzed are high sepsis (12.1%), liver metastases (9.1%), pneumonia (6.1%), acute respiratory failure (6.1%) and renal acute renal (6.1%). CONCLUSIONS: Although the most frequent primary diagnosis of sepsis, by grouping the diagnoses, the most frequent DRG is respiratory disease, so it has to make comprehensive and quality coding to adjust the relative weight of the same reality. It is essential to specify the source of clinical information used for coding, the degree of malnutrition, for greater specificity in the data.

Evaluation of factors affecting plate waste of inpatients in different healthcare settings.

INTRODUCTION AND AIMS: Satisfaction of inpatients with served food within a hospital care system still constitutes one of the main attempts to modernize food services. The impact of type of menu, food category, hospital centre and timetable on the meals wastage produced in different Spanish healthcare settings, was evaluated. METHODS: Meal wastage was measured through a semiquantitative 5-point scale ("nothing on plate"; "¼ on plate"; "half on plate"; "¾ on plate" and "all on plate"). The study was carried out in two periods of three months each in 2010 and 2011. A trained person took charge of measuring plate waste classified into 726 servings belonging to 11 menus. In total 31,392 plates were served to 7,868 inpatients. A Kruskal-Wallis non-parametric test (p < 0.05) was applied to evaluate significant differences among the variables studied. RESULTS: The menus were satisfactorily consumed because more than 50% of the plates were classified as "nothing on plate". Regarding food categories, 26.78% of the plates corresponded to soups and purées, while pasta and rice, and prepared foods were only distributed in 4-5% of the servings. Desserts were mostly consumed, while cooked vegetables were less accepted by the inpatients evaluated. Other factors such as hospital centre influenced plate waste (p < 0.05) but timetable did not (p > 0.05). CONCLUSION: Visual inspections of plate waste might be useful to optimize type and quality of menus served. The type of menu served and the food category could have a great influence on food acceptability by the inpatientsstudied.

El proyecto EMECAS: protocolo del estudio multicéntrico en España de los efectos a corto plazo de la contaminación atmosférica sobre la salud.

The EMECAM Project demonstrated the short-term effect of air pollution on the death rate in 14 cities in Spain throughout the 1990-1995 period. The Spanish Multicentre Study on Health Effects of Air Pollution (EMECAS) is broadening these objectives by incorporating more recent data, information on hospital disease admissions and totaling 16 Spanish cities. This is an ecological time series study in which the response variables are the daily deaths and the emergency hospitalizations due to circulatory system diseases and respiratory diseases among the residents in each city. Pollutants analyses: suspended particles, SO2, NO2, CO and O3. Control variables: meteorological, calendar, seasonality and influenza trend and incidence. Statistical analysis: estimate of the association in each city by means of the construction of generalized additive Poisson regression models and metanalysis for obtaining combined estimators. The EMECAS Project began with the creation of three working groups (Exposure, Epidemiology and Analysis Methodology) which defined the protocol. The average levels of pollutants were below those established under the current regulations for sulfur dioxide, carbon monoxide and ozone. The NO2 and PM10 values were around those established under the regulations (40 mg/m3). This is the first study of the relationship between air pollution and disease rate among one group of Spanish cities. The pollution levels studied are moderate for some pollutants, although for others, especially NO2 and particles, these levels could entail a problem with regard to complying with the regulations in force.

Eating habits of preschool children with high migrant status in Switzerland according to a new food frequency questionnaire.

Assessment of eating habits in young children from multicultural backgrounds has seldom been conducted. Our objectives were to study the reproducibility and the results of a food frequency questionnaire (FFQ) developed to assess changes in eating habits of preschool children with a high migrant population, in the context of a multidisciplinary multilevel lifestyle intervention. Three kindergarten classes (53% from migrant backgrounds) in French-speaking Switzerland were randomly selected and included 16 girls and 28 boys (mean age +/- SD, 5.4 +/- 0.7 years). The FFQ was filled out twice within a 4-week interval by the parents. Spearman rank correlations between the first and the second FFQ for the 39 items of the food questions were as follows: low (r < 0.50) for 8 (7 P < .05 and 1 nonsignificant), moderate (0.50 <or= r < 0.70) for 22 (all P < .01), and high (r >or= 0.70) for 9 (all P < .01). In addition, 28 of 39 intraclass correlation coefficients were high (>0.50, all P < .01). Eighty-six percent of the children ate breakfast at home daily, but only 67% had lunch at home. The percentages of children eating at least once a week in front of the TV were as follows: 50% for breakfast, 33% for lunch, 38% for dinner, and 48% for snacks. Forty percent of children asked their parents to buy food previously seen in advertisements and ate fast food between once a week and once a month. Children generally consumed foods with a high-energy content. The FFQ yielded good test-retest reproducibility for most items of the food questions and gave relevant findings about the eating habits of preschool children in areas with a high migrant population.

The use of rapid dengue diagnostic tests in a routine clinical setting in a dengue-endemic area of Colombia

There is insufficient evidence of the usefulness of dengue diagnostic tests under routine conditions. We sought to analyse how physicians are using dengue diagnostics to inform research and development. Subjects attending 14 health institutions in an endemic area of Colombia with either a clinical diagnosis of dengue or for whom a dengue test was ordered were included in the study. Patterns of test-use are described herein. Factors associated with the ordering of dengue diagnostic tests were identified using contingency tables, nonparametric tests and logistic regression. A total of 778 subjects were diagnosed with dengue by the treating physician, of whom 386 (49.5%) were tested for dengue. Another 491 dengue tests were ordered in subjects whose primary diagnosis was not dengue. Severe dengue classification [odds ratio (OR) 2.2; 95% confidence interval (CI) 1.1-4.5], emergency consultation (OR 1.9; 95% CI 1.4-2.5) and month of the year (OR 3.1; 95% CI 1.7-5.5) were independently associated with ordering of dengue tests. Dengue tests were used both to rule in and rule out diagnosis. The latter use is not justified by the sensitivity of current rapid dengue diagnostic tests. Ordering of dengue tests appear to depend on a combination of factors, including physician and institutional preferences, as well as other patient and epidemiological factors.

A comparison of the alr and ilr transformations for kernel density estimation of compositional data

In a seminal paper, Aitchison and Lauder (1985) introduced classical kernel densityestimation techniques in the context of compositional data analysis. Indeed, they gavetwo options for the choice of the kernel to be used in the kernel estimator. One ofthese kernels is based on the use the alr transformation on the simplex SD jointly withthe normal distribution on RD-1. However, these authors themselves recognized thatthis method has some deficiencies. A method for overcoming these dificulties based onrecent developments for compositional data analysis and multivariate kernel estimationtheory, combining the ilr transformation with the use of the normal density with a fullbandwidth matrix, was recently proposed in Martín-Fernández, Chacón and Mateu-Figueras (2006). Here we present an extensive simulation study that compares bothmethods in practice, thus exploring the finite-sample behaviour of both estimators

Free-breathing renal magnetic resonance angiography with steady-state free-precession and slab-selective spin inversion combined with radial k-space sampling and water-selective excitation.

The impact of radial k-space sampling and water-selective excitation on a novel navigator-gated cardiac-triggered slab-selective inversion prepared 3D steady-state free-precession (SSFP) renal MR angiography (MRA) sequence was investigated. Renal MRA was performed on a 1.5-T MR system using three inversion prepared SSFP approaches: Cartesian (TR/TE: 5.7/2.8 ms, FA: 85 degrees), radial (TR/TE: 5.5/2.7 ms, FA: 85 degrees) SSFP, and radial SSFP combined with water-selective excitation (TR/TE: 9.9/4.9 ms, FA: 85 degrees). Radial data acquisition lead to significantly reduced motion artifacts (P < 0.05). SNR and CNR were best using Cartesian SSFP (P < 0.05). Vessel sharpness and vessel length were comparable in all sequences. The addition of a water-selective excitation could not improve image quality. In conclusion, radial k-space sampling reduces motion artifacts significantly in slab-selective inversion prepared renal MRA, while SNR and CNR are decreased. The addition of water-selective excitation could not improve the lower CNR in radial scanning.

Econometrics on GPUs

A graphical processing unit (GPU) is a hardware device normally used to manipulate computer memory for the display of images. GPU computing is the practice of using a GPU device for scientific or general purpose computations that are not necessarily related to the display of images. Many problems in econometrics have a structure that allows for successful use of GPU computing. We explore two examples. The first is simple: repeated evaluation of a likelihood function at different parameter values. The second is a more complicated estimator that involves simulation and nonparametric fitting. We find speedups from 1.5 up to 55.4 times, compared to computations done on a single CPU core. These speedups can be obtained with very little expense, energy consumption, and time dedicated to system maintenance, compared to equivalent performance solutions using CPUs. Code for the examples is provided.

Modulatory effects of 5Hz rTMS over the primary somatosensory cortex in focal dystonia--an fMRI-TMS study.

Dystonia is associated with impaired somatosensory ability. The electrophysiological method of repetitive transcranial magnetic stimulation (rTMS) can be used for noninvasive stimulation of the human cortex and can alter cortical excitability and associated behavior. Among others, rTMS can alter/improve somatosensory discrimation abilities, as shown in healthy controls. We applied 5Hz-rTMS over the left primary somatosensory cortex (S1) in 5 patients with right-sided writer's dystonia and 5 controls. We studied rTMS effects on tactile discrimination accuracy and concomitant rTMS-induced changes in hemodynamic activity measured by functional magnetic resonance imaging (fMRI). Before rTMS, patients performed worse on the discrimination task than controls even though fMRI showed greater task-related activation bilaterally in the basal ganglia (BG). In controls, rTMS led to improved discrimination; fMRI revealed this was associated with increased activity of the stimulated S1, bilateral premotor cortex and BG. In dystonia patients, rTMS had no effect on discrimination; fMRI showed similar cortical effects to controls except for no effects in BG. Improved discrimination after rTMS in controls is linked to enhanced activation of S1 and BG. Failure of rTMS to increase BG activation in dystonia may be associated with the lack of effect on sensory discrimination in this group and may reflect impaired processing in BG-S1 connections. Alternatively, the increased BG activation seen in the baseline state without rTMS may reflect a compensatory strategy that saturates a BG contribution to this task.