High Outliers Research Articles

Annual Time Series of Global VIIRS Nighttime Lights Derived from Monthly Averages: 2012 to 2019

A consistently processed annual global nighttime lights time series (2012–2019) was produced using monthly cloud-free radiance averages made from low light imaging day/night band (DNB) data collected by the NASA/NOAA Visible Infrared Imaging Radiometer Suite (VIIRS). The processing steps are modified from the original methods developed to produce annual nighttime lights products from nightly data. Only two years of VIIRS nighttime lights (VNL) were produced with the V.1 methods: 2015 and 2016. Here we report on methods used to produce a V.2 VNL time series from the monthly averages with filtering to remove extraneous features such as biomass burning, aurora, and background. In this case, outlier removal is achieved with a twelve-month median, which discards high and low radiance outliers, thus isolating the background to a narrow range of radiances under 1 nW/cm2/sr. Background areas with no detectable lighting are further isolated using a statistical measure of texture, 3 × 3 data range (DR). The DR threshold for zeroing out background rises as the number of cloud-free observations falls. The V.2 method extends the temporal leverage in the noise filtering by developing the DR threshold from a multiyear maximum DR and a multiyear percent cloud-free grid. Additional noise filtering is achieved by zeroing out grid cells that have low average radiances (<0.6 nW/cm2/sr) and detection in only one or two years out of eight. The spatial extent and average radiance levels are compared for the V.1 and V.2 2015 VNL. For the vast majority of grid cells, the average radiances are nearly the same in the two products. However, the V.2 product has more areas of dim lighting detected. The key advantages of the V.2 time series include consistent processing and threshold levels across all years, thus optimizing the set for change detection analyses.

Practical globally optimal consensus maximization by Branch-and-bound based on interval arithmetic

Consensus maximization is widely used in robust model fitting, and it is usually solved by RANSAC-type methods in practice. However, these methods cannot guarantee global optimality and sometimes return the wrong solutions. A series of Branch-and-bound (BnB) based globally optimal methods have been proposed, most of which involve deriving a complex bound. Interval arithmetic was utilized to derive simple bounds for BnB in solving geometric matching problems in 2003. However, this idea was somewhat forgotten in the community because it seems natural that the simple interval arithmetic based bounds might be worse than those elaborate bounds. Recently, some new globally optimal algorithms without using BnB were developed for consensus maximization, but they can only work with a small number of data points and low outlier ratios. In this work, we draw the idea of simple bounds by interval arithmetic back on the map and demonstrate its practicability by making substantial extensions. Concretely, we give detailed derivation of solving robust model fitting problems with both linear and quasi-convex residuals and propose practical methods to use them under Unit-Norm constraint and in a high-dimensional problem. Extensive experiments show that the proposed method can handle practical problems with large number of data points and high outlier ratios. It outperforms state-of-the-art global, RANSAC-type, and deterministic methods in terms of both accuracy and efficiency in low-dimensional problems. The source code is publicly available.22https://github.com/YiruWangYuri/Demo-for-GoIA.

Variations in Persistent Use of Low-Value Breast Cancer Surgery

Through the Choosing Wisely campaign, surgical specialties identified 4 low-value breast cancer operations. Preliminary data suggest varying rates of deimplementation and have identified patient-level and clinician-level determinants of continued overuse. However, little information exists about facility-level variation or determinants of differential deimplementation. To identify variation and determinants of persistent use of low-value breast cancer surgical care. Retrospective cohort study in which reliability-adjusted facility rates of each procedure were calculated using random-intercept hierarchical logistic regression before and after evidence demonstrated that each procedure was unnecessary. The National Cancer Database is a prospective cancer registry of patients encompassing approximately 70% of all new cancer diagnoses from more than 1500 facilities in the United States. Data were analyzed from November 2019 to August 2020. The registry included women 18 years and older diagnosed as having breast cancer between 2004 and 2016 and meeting inclusion criteria for each Choosing Wisely recommendation. Rate of each low-value breast cancer procedure based on facility type and breast cancer volume categories before and after the release of data supporting each procedure's omission. The total cohort included 920 256 women with a median age of 63 years. Overall, 86% self-identified as White, 10% as Black, 3% as Asian, and 4.5% as Hispanic. Most women in this cohort were insured (51% private and 47% public), were living in a metropolitan or urban area (88% and 11%, respectively), and originated from the top half of income-earning households (65.5%). While there was significant deimplementation of axillary lymph node dissection and lumpectomy reoperation in response to guidelines supporting omission of these procedures, rates of contralateral prophylactic mastectomy and sentinel lymph node biopsy in older women increased during the study period. Academic research programs and high-volume facilities overall demonstrated the greatest reduction in use of these low-value procedures. There was significant interfacility variation for each low-value procedure. Facility-level axillary lymph node dissection rates ranged from 7% to 47%, lumpectomy reoperation rates ranged from 3% to 62%, contralateral prophylactic mastectomy rates ranged from 9% to 67%, and sentinel lymph node biopsy rates ranged from 25% to 97%. Pearson correlation coefficient for each combination of 2 of the 4 procedures was less than 0.11, suggesting that hospitals were not consistent in their deimplementation performance across all 4 procedures. Many were high outliers in one procedure but low outliers in another. Interfacility variation demonstrates a performance gap and an opportunity for formal deimplementation efforts targeting each procedure. Several facility-level characteristics were associated with differential deimplementation and performance.

Spatial pattern of land surface temperatures and its relation to underground coal fires in the Khanh Hoa Coal Field, North-East of Vietnam

Change of land surface temperatures (LSTs) is indicative of underground coal fires (UCFs). The limitation of commonly used methods for UCF detection is that spatial pattern of LSTs is not taken into account. This study aims to identify spatial pattern of LSTs retrieved from remotely sensed data and its relation to UCFs. LSTs were firstly retrieved from the Landsat-8 TIRS data using the radiative transfer equation (RTE). The local Moran’s I statistics was then used to identify the spatial pattern of these LSTs. Different degrees of spatial pattern (low, medium, high, extreme high, lower, and upper outliers) were then identified by setting the first, second, third, and fourth quartiles and the MEDIAN + 1.5 × IQR (IQR is the interquartile range) and MEDIAN + 3 × IQR formulas as thresholds. The relation of LST spatial pattern to UCFs was finally identified by overlapping spatial pattern layers on known active UCF sites and compared with those reported previously. Results obtained from 2nd December 2013 Landsat-8 TIRS data in the Khanh Hoa coal field (north-east of Vietnam) showed that LSTs followed a pattern of spatial clustering of high values in the coal field. The areas with the degrees of spatial correlation in the quartiles of 0–25%, 25–50%, 50–75%, and above 75% were 66.5, 66.9, 34.1, and 71.1 ha, respectively. Lower and upper outliers were detected at positions of 12.38 and 21.31 corresponding to the areas of 23.4 and 4.6 ha, respectively. These outliers were mainly concentrated around eight active UCF sites and were highly consistent with those obtained from previous studies. The results of this investigation show that the closer the UCF sites, the higher the spatial autocorrelation level. There exists a strong degree of positive correlation between the distribution of LST spatial pattern with active UCFs. These findings suggest that high degree of autocorrelation of LSTs can be used to effectively detect UCFs.

Identifying Outlier Hospitals in Gastric Cancer Lymph Node Yield Using the National Cancer Database

BackgroundLymph node (LN) yield is a key quality indicator that is associated with improved staging in surgically resected gastric cancer. The National Comprehensive Cancer Network recommends a yield of ≥15 LNs for proper staging, yet most facilities in the United States fail to achieve this number. The present study aimed to identify factors that could affect LN yield on a facility level and identify outlier hospitals. MethodsThis was a retrospective review of adults (aged ≥18 y) with gastric cancer (Tumor-Node-Metastasis Stages I-III) who underwent gastrectomy. Data were analyzed from the National Cancer Database (2004-2016). Multivariate analysis identified patient and tumor characteristics, whereas an observed-to-expected ratio of identified outlier hospitals. Facility factors were compared between high and low outliers. ResultsA total of 26,590 patients were included in this study. Of these patients, only 50.3% had an LN yield ≥15. The multivariate model of patient and tumor characteristics demonstrated a concordance index was 0.684. A total of 1245 facilities were included. There were 198 low outlier LN yield hospitals and 135 high outlier LN yield hospitals (observed-to-expected ratio of 0.42 ± 0.24 versus 1.38 ± 0.19, P < 0.0001). There was a difference in facility type between low and high outliers (P < 0.0001). High LN yield hospitals had a larger surgical volume than low LN yield hospitals (median 8.4 [4.9, 13.5] versus 3.5 [2.4, 5.2]; P < 0.0001). ConclusionsNearly half of the population exhibited low compliance to National Comprehensive Cancer Network recommendations. Facility-level disparities exist as high yearly surgical volume and academic facility status distinguished high-performing outlier hospitals.

A new density-based subspace selection method using mutual information for high dimensional outlier detection

Outlier detection in high dimensional data faces the challenge of curse of dimensionality, where irrelevant features may prevent detection of outliers. In this research, we propose a novel efficient unsupervised density-based subspace selection for outlier detection in the projected subspace. First, the Maximum-Relevance-to-Density algorithm(MRD) is proposed to select the relevant subspace based on the mutual information. Then, applying the concept of redundancy among features, we present an efficient relevant subspace selection method called minimum-Redundancy-Maximum-Relevance-to-Density (mRMRD). Finally, the degree of outlierness of data points in the corresponding relevant subspace is computed based on Local Outlier Factor(LOF). Experimental results on both real and synthetic data demonstrate that the proposed algorithms – based on MRD and mRMRD criteria – increase the accuracy of outlier detection while reducing computational complexity and execution time. Moreover, as the dimensionality increases, the accuracy of outlier detection on mRMRD-based relevant subspace is higher than MRD-based relevant subspace. This verifies that the proposed mRMRD-based subspace selection algorithm can efficiently select the subspace by considering the relevance between features.

A Streamlined scRNA-Seq Data Analysis Framework Based on Improved Sparse Subspace Clustering

One advantage of single-cell RNA sequencing is its ability in revealing cell heterogeneity by cell clustering. However, cell clustering based on single-cell RNA sequencing is challenging due to the high transcript amplification noise, sparsity and outlier cell populations. In this study, we propose a novel sparse subspace clustering method called Structured Sparse Subspace Clustering and Completion for single-cell RNA sequencing analysis by assuming the cells related together are in the same subspace, and so the relationships among cells can be described within a subspace instead of between cell pairs. The proposed optimization model is solved by the Linearized Alternating Direction Method of Multipliers, in which data completion and spectral clustering are combined as a whole by mutual constraint. It is worth noting that random walk is used in the model to make the coefficient matrix more diagonal in the optimum iterative procedure, and the effect is significant. Our model is applied and compared with 5 state-of-the-art clustering methods on 6 public single cell datasets and a simulated data set with cell numbers varying from 56 to over 3000. As a result, our model outperforms the other clustering methods in clustering accuracy as evaluated by Adjusted Rand Index, Normalized Mutual Information, Homogeneity and Completeness, especially compared with the other improved sparse subspace clustering methods.

A Practical O(N2) Outlier Removal Method for Correspondence-Based Point Cloud Registration.

Point cloud registration (PCR) is an important and fundamental problem in 3D computer vision, whose goal is to seek an optimal rigid model to register a point cloud pair. Correspondence-based PCR techniques do not require initial guesses and gain more attentions. However, 3D keypoint techniques are much more difficult than their 2D counterparts, which results in extremely high outlier rates. Current robust techniques suffer from very high computational cost. In this paper, we propose a polynomial time ( O(N2), where N is the number of correspondences.) outlier removal method. Its basic idea is to reduce the input set into a smaller one with a lower outlier rate based on bound principle. To seek tight lower and upper bounds, we originally define two concepts, i.e., correspondence matrix (CM) and augmented correspondence matrix (ACM). We propose a cost function to minimize the determinant of CM or ACM, where the cost of CM rises to a tight lower bound and the cost of ACM leads to a tight upper bound. Then, we propose a scale-adaptive Cauchy estimator (SA-Cauchy) for further optimization. Extensive experiments on simulated and real PCR datasets demonstrate that the proposed method is robust at outlier rates above 99 percent and 1 ∼2 orders faster than its competitors. The source code will be made publicly available in https://ljy-rs.github.io/web/.

Subspace-Based Preprocessing Module for Fast Hyperspectral Endmember Selection

Endmember extraction algorithms (EEAs) play a crucial role in hyperspectral image (HSI) perception, and yet they normally suffer from three flaws: 1) High computational burden, 2) weak noise robustness, and 3) high outlier sensitivity. To solve these problems, this article proposes a fast subspace-based preprocessing module, called fast subspace-based preprocessing module (FSPM), to select a high-quality data subset for subsequent endmember extraction. Specifically, FSPM first transforms an HSI into a low-dimensional data subspace using singular value decomposition. For each component pair, FSPM then detects its convex hull vertices and proposes a local outlier score measurement to remove potential outliers. FSPM finally transforms determined data points into noise-reduced data space for the sake of identifying endmembers. The proposed FSPM sheds new light on the current preprocessing field, which can fast reduce noises and outliers as well as remove redundant data points. Based on various validation metrics, experiments conducted on both synthetic and real HSIs indicate that the proposed FSPM is superior to current state-of-the-art preprocessing techniques.

1348. Indirect Standardization to Improve Comparison of Children’s Hospitals’ Antimicrobial Use

BackgroundAntimicrobial use (AU) measured by days of therapy per 1000 patient-days (DOT/1000pd), the most established metric, varies widely between children’s hospitals despite robust adoption of antimicrobial stewardship. Differences in diagnoses and procedures (case mix) between hospitals are a source of AU variation not included in adjustment methods such as the Standardized Antimicrobial Administration Ratio. In this study, we evaluated an indirect standardization method to adjust children’s hospital AU for case mix.MethodsThis multicenter retrospective cohort study included 51 children’s hospitals participating in the Pediatric Health Information System database from 2016-2018. All inpatient, observation, and neonatal admissions were included, with a total of 2,558,948 discharges. Hospitalizations were grouped into 83 strata defined based on All Patients Refined Diagnosis Related Groups (APR-DRGs). Observed to expected (O:E) ratios were calculated by indirect standardization of mean antibiotic DOT per case, with expected values from 2016-2018 and observed values from 2018, and compared to DOT/1000pd. Outlier hospitals were defined by O:E z-scores corresponding to below 10th percentile (low outlier) and above 90th percentile (high outlier).ResultsAntibacterial DOT/1000pd ranged from 345 to 776 (2.2-fold variation from lowest to highest), whereas O:E ratios ranged from 0.8 to 1.14 (1.4-fold variation from lowest to highest) (Figure 1). O:E ratios were moderately correlated with DOT/1000pd (correlation estimate 0.45; 95% CI 0.19-0.64; p=0.0008). Three high outlier hospitals and 6 low outlier hospitals were identified. Examining hospitals with comparably high DOT/1000pd but discordant O:E ratios, differences could be explained by variation in both case mix and condition-specific AU within strata defined by APR-DRGs.Figure 1. Individual hospitals labeled on the X-axis, ordered by level of antibacterial DOT/1000pd (left axis), represented by bars. Diamonds represent O:E ratios derived by indirect standardization (right axis). Outlier hospitals (low and high) are highlighted in yellow. Dashed horizontal lines represent 10th percentile (lower) and 90th percentile (upper) limits of the O:E ratio distribution. ConclusionThe observed variation in DOT/1000pd between hospitals is reduced when indirect standardization is applied to account for case mix differences. This approach can be adapted for more specific uses including clinical conditions, patient populations, or antimicrobial agents. Indirect standardization may enhance stewardship efforts by providing adjusted comparisons that incorporate case mix differences between hospitals.Disclosures All Authors: No reported disclosures

Differences Between Center-level Outcomes in Emergency and Elective General Surgery

BackgroundCenter-level outcome metrics have long been tracked in elective surgery (ELS). Despite recent interest in measuring emergency general surgery (EGS) quality, centers are often compared based on elective or combined outcomes. Therefore, quality of care for emergency surgery specifically is unknown. MethodsWe extracted data on EGS and ELS patients from the 2016 State Inpatient Databases of Florida, New York, and Kentucky. Centers that performed >100 ELS and EGS operations were included. Risk-adjusted mortality, complication, and failure to rescue (FTR, death after complication) rates were calculated and observed-to-expected ratios were calculated by center for ELS and EGS patients. Centers were determined to be high or low outliers if the 90% CI for the observed: expected ratio excluded 1. We calculated the frequency with which centers demonstrated a different performance status between EGS and ELS. Kendall's tau values were calculated to assess for correlation between EGS and ELS status. ResultsA total of 204 centers with 45,500 EGS cases and 49,380 ELS cases met inclusion criteria. Overall mortality, complication, and FTR rates were 1.7%, 8.0%, and 14.5% respectively. There was no significant correlation between mortality performance in EGS and ELS, with 36 centers in a different performance category (high outlier, low outlier, as expected) in EGS than in ELS. The correlation for complication rates was 0.20, with 60 centers in different categories for EGS and ELS. For FTR rates, there was no correlation, with 16 centers changing category. ConclusionsThere was minimal correlation between outcomes for ELS and EGS. High performers in one category were rarely high performers in the other. There may be important differences between the processes of care that are important for EGS and ELS outcomes that may yield meaningful opportunities for quality improvement.

Multiple Outlier Detection Tests for Parametric Models

We propose a simple multiple outlier identification method for parametric location-scale and shape-scale models when the number of possible outliers is not specified. The method is based on a result giving asymptotic properties of extreme z-scores. Robust estimators of model parameters are used defining z-scores. An extensive simulation study was done for comparing of the proposed method with existing methods. For the normal family, the method is compared with the well known Davies-Gather, Rosner’s, Hawking’s and Bolshev’s multiple outlier identification methods. The choice of an upper limit for the number of possible outliers in case of Rosner’s test application is discussed. For other families, the proposed method is compared with a method generalizing Gather-Davies method. In most situations, the new method has the highest outlier identification power in terms of masking and swamping values. We also created R package outliersTests for proposed test.

Accuracy Assessment of GEDI Terrain Elevation and Canopy Height Estimates in European Temperate Forests: Influence of Environmental and Acquisition Parameters

Lidar remote sensing has proven to be a powerful tool for estimating ground elevation, canopy height, and additional vegetation parameters, which in turn are valuable information for the investigation of ecosystems. Spaceborne lidar systems, like the Global Ecosystem Dynamics Investigation (GEDI), can deliver these height estimates on a near global scale. This paper analyzes the accuracy of the first version of GEDI ground elevation and canopy height estimates in two study areas with temperate forests in the Free State of Thuringia, central Germany. Digital terrain and canopy height models derived from airborne laser scanning data are used as reference heights. The influence of various environmental and acquisition parameters (e.g., canopy cover, terrain slope, beam type) on GEDI height metrics is assessed. The results show a consistently high accuracy of GEDI ground elevation estimates under most conditions, except for areas with steep slopes. GEDI canopy height estimates are less accurate and show a bigger influence of some of the included parameters, specifically slope, vegetation height, and beam sensitivity. A number of relatively high outliers (around 9–13% of the measurements) is present in both ground elevation and canopy height estimates, reducing the estimation precision. Still, it can be concluded that GEDI height metrics show promising results and have potential to be used as a basis for further investigations.

SubTSBR to tackle high noise and outliers for data-driven discovery of differential equations

Data-driven discovery of differential equations has been an emerging research topic. We propose a novel algorithm subsampling-based threshold sparse Bayesian regression (SubTSBR) to tackle high noise and outliers. The subsampling technique is used for improving the accuracy of the Bayesian learning algorithm. It has two parameters: subsampling size and the number of subsamples. When the subsampling size increases with fixed total sample size, the accuracy of our algorithm goes up and then down. When the number of subsamples increases, the accuracy of our algorithm keeps going up. We demonstrate how to use our algorithm step by step and compare our algorithm with threshold sparse Bayesian regression (TSBR) for the discovery of differential equations. We show that our algorithm produces better results. We also discuss the merits of discovering differential equations from data and demonstrate how to discover models with random initial and boundary condition as well as models with bifurcations. The numerical examples are: (1) predator-prey model with noise, (2) shallow water equations with outliers, (3) heat diffusion with random initial and boundary condition, and (4) fish-harvesting problem with bifurcations.

Adaptive divergence, neutral panmixia, and algal symbiont population structure in the temperate coral Astrangia poculata along the Mid-Atlantic United States.

Astrangia poculata is a temperate scleractinian coral that exists in facultative symbiosis with the dinoflagellate alga Breviolum psygmophilum across a range spanning the Gulf of Mexico to Cape Cod, Massachusetts. Our previous work on metabolic thermal performance of Virginia (VA) and Rhode Island (RI) populations of A. poculata revealed physiological signatures of cold (RI) and warm (VA) adaptation of these populations to their respective local thermal environments. Here, we used whole-transcriptome sequencing (mRNA-Seq) to evaluate genetic differences and identify potential loci involved in the adaptive signature of VA and RI populations. Sequencing data from 40 A. poculata individuals, including 10 colonies from each population and symbiotic state (VA-white, VA-brown, RI-white, and RI-brown), yielded a total of 1,808 host-associated and 59 algal symbiont-associated single nucleotide polymorphisms (SNPs) post filtration. Fst outlier analysis identified 66 putative high outlier SNPs in the coral host and 4 in the algal symbiont. Differentiation of VA and RI populations in the coral host was driven by putatively adaptive loci, not neutral divergence (Fst = 0.16, p = 0.001 and Fst = 0.002, p = 0.269 for outlier and neutral SNPs respectively). In contrast, we found evidence of neutral population differentiation in B. psygmophilum (Fst = 0.093, p = 0.001). Several putatively adaptive host loci occur on genes previously associated with the coral stress response. In the symbiont, three of four putatively adaptive loci are associated with photosystem proteins. The opposing pattern of neutral differentiation in B. psygmophilum, but not the A. poculata host, reflects the contrasting dynamics of coral host and algal symbiont population connectivity, dispersal, and gene by environment interactions.

Outlier percentage estimation for shape‐ and parameter‐independent outlier detection

Accurate and robust three-dimensional reconstruction of objects allows for applications in many aspects of modern life. Yet, it typically suffers from outliers and noise which often need to be post-processed. Although many algorithms are able to effectively remove the outliers, most require a certain amount of manual tuning of the parameter(s) or to have a parameter(s) set based on the rule of thumb. New machine learning and artificial intelligence-based methods have also been introduced but may require vast parallel computing resources as well as training data. In the present study, a novel combinatory-distance-based method capable of high accuracy outlier detection named as the sorted distance divergence point (SDDP) is introduced. Results show that SDDP is able to achieve an average accuracy of 98% in outlier detection. Moreover, the introduced distance function and outlier percentage allow clear labelling of inliers and outliers cloud points. Therefore, SDDP presents an attractive enhancement to existing methods; namely, the manual parameter(s) tuning may not be necessary. The adaptability and utility of SDDP is further demonstrated by incorporating SDDP with current methods, to produce a high accuracy outlier detector. When tested with 17 objects with 20–50% outliers, attain F1 and F2 scores averaging 0.960 and 0.968, respectively.

Robust Fuzzy Clustering Algorithms for Change-Point Regression Models

This article presents a robust fuzzy procedure for estimating change-point regression models. We propose incorporating the fuzzy change-point algorithm with the M-estimation technique for robust estimations. The fuzzy c partitions concept is embedded into the change-point regression model so the fuzzy c-regressions and fuzzy c-means clustering can be employed to obtain the estimates of change-points and regression parameters. The M estimation with a robust criterion is used to make the estimators robust to the presence of outliers and heavy-tailed distributions. We create two robust algorithms named FCH and FCT by using Huber’s and Tukey’s functions as the robust criterion respectively. Extensive experiments with numerical and real examples are provided for demonstrating the effectiveness and the superiority of the proposed algorithms. The experimental results show the proposed algorithms are resistant to atypical observations and outperform the existing methods. The proposed FCH and FCT are generally comparable but FCT performs better in the presence of extremely high leverage outliers and heavy-tailed distributions. Real data applications show the practical usefulness of the proposed method.

Establishing a National Rate of Rectal Biopsy Positivity in the Diagnosis of Hirschsprung Disease.

Although rectal biopsy has long been established as the criterion standard for the diagnosis of Hirschsprung disease, little to no information exists regarding nationwide rates of rectal biopsy positivity or interinstitutional variability. We sought to determine the national rate of rectal biopsy positivity and factors contributing to institutional variability. A retrospective review of the Children's Hospital Association's (CHA) Pediatric Health Information System from 2009 to 2018 identified infants <100 days old with ICD-9/ICD-10 procedural codes for rectal biopsy in addition to codes for pull-through procedures within 500 days of life as a proxy for positive biopsy. A subgroup analysis of only patients biopsied at institutions with positive biopsy rates 1 standard deviation above or below the mean positive biopsy rate (deemed high and low outliers) was performed to better delineate these populations. A total of 7225 children underwent rectal biopsies between 2009 and 2018 at 52 Children's Hospital Association Hospitals. Mean positive biopsy rate for individual institutions was 21.5% (standard deviation ± 6.4%). Linear regression to predict the effect of hospital surgical volume on positive biopsy rate demonstrated no volume outcome relationship (R2 = 0.049). Patients at high outlier hospitals for biopsy positivity were found to travel significantly further to the hospital (232.5 vs 123.1 miles, P < 0.0001) when compared to patients presenting at low outlier hospitals. There appears to be little interinstitutional variability in the rate of surgery following rectal biopsy for presumed Hirschsprung and no significant relationship to surgical volume. About 1 in every 4 infants undergoing biopsy proceeds to surgery.

Reference interval for albumin-adjusted calcium based on a large UK population.

Primary hyperparathyroidism is a common condition and results in hypercalcaemia, especially in older women. Thus, it is critical to obtain a robust estimate for the upper limit of the reference interval for albumin-adjusted serum calcium in the general population. The current reference interval in use in the UK (Pathology Harmony range, 2.20 to 2.60mmol/L) was based on a consensus. To establish a reference interval for albumin-adjusted serum calcium in men and women. Cross-sectional study of men and women who did not have chronic kidney disease or vitamin D deficiency; outliers were identified statistically and then rejected and then a 99% reference interval was calculated. 502524 men and women aged 40 to 69years from the UK Biobank Study. Serum total calcium, albumin, 25-hydroxyvitamin D, estimated glomerular function (eGFR). We developed an equation for albumin-adjusted serum calcium and applied it to 178377 men and women who did not have chronic kidney disease or vitamin D deficiency. We identified 2962 (1.7%) as outliers, and when excluded, we report a 99% reference interval of 2.19 to 2.56mmol/L (8.76 to 10.24mg/dL). We found that for older (55-69years) and younger women (40-55years) the upper limits were 2.59mmol/L and 2.57mmol/L and that for all men, the upper limit was 2.55mmol/L. We have established an upper limit of the reference range for older women that would identify all high outliers (2.60mmol/L and above). The upper limit for young women and for men is lower, at 2.57 and 2.55mmol/L respectively. The current reference interval in use has to be updated and improved based on these findings. These upper limits may prove helpful for identifying hypercalcaemic disorders like primary hyperparathyroidism in clinical practice.

A Modified Robust Support Vector Regression Approach for Data Containing High Leverage Points and Outliers in the Y-direction

The support vector regression (SVR) model is currently a very popular non-parametric method used for estimating linear and non-linear relationships between response and predictor variables. However, there is a possibility of selecting vertical outliers as support vectors that can unduly affect the estimates of regression. Outliers from abnormal data points may result in bad predictions. In addition, when both vertical outliers and high leverage points are present in the data, the problem is further complicated. In this paper, we introduced a modified robust SVR technique in the simultaneous presence of these two problems. Three types of SVR models, i.e. eps-regression (e-SVR), nu-regression (v-SVR) and bound constraint eps-regression (e-BSVR), with eight different kernel functions are integrated into the new proposed algorithm. Based on 10-fold cross-validation and some model performance measures, the best model with a suitable kernel function is selected. To make the selected model robust, we developed a new double SVR (DSVR) technique based on fixed parameters. This can be used to detect and reduce the weight of influential observations or anomalous points in the data set. The effectiveness of the proposed technique is verified by using a simulation study and some well-known contaminated data sets.