Validation of the PediBIRN-7 clinical prediction rule for pediatric abusive head trauma.

BACKGROUND: The PediBIRN-7 clinical prediction rule incorporates the (positive or negative) predictive contributions of completed abuse evaluations to estimate abusive head trauma (AHT) probability after abuse evaluation. Applying definitional criteria as proxies for AHT and non-AHT ground truth, it performed with sensitivity 0.73 (95 % CI: 0.66-0.79), specificity 0.87 (95 % CI: 0.82-0.90), and ROC-AUC 0.88 (95 % CI: 0.85-0.92) in its derivation study. OBJECTIVE: To validate the PediBIRN-7's AHT prediction performance in a novel, equivalent, patient population. PARTICIPANTS AND SETTINGS: Consecutive, acutely head-injured children <3 years hospitalized for intensive care across eight sites between 2017 and 2020 with completed skeletal surveys and retinal exams (N = 342). METHODS: Secondary analysis of an existing, cross-sectional, prospective dataset, including assignment of patient-specific estimates of AHT probability, calculation of AHT prediction performance measures (ROC-AUC, sensitivity, specificity, predictive values), and completion of sensitivity analyses to estimate best- and worst-case prediction performances. RESULTS: Applying the same definitional criteria, the PediBIRN-7 performed with sensitivity 0.74 (95 % CI: 0.66-0.81), specificity 0.77 (95 % CI: 0.70-0.83), and ROC-AUC 0.83 (95 % CI: 0.78-0.88). The reduction in ROC-AUC was statistically insignificant (p = .07). Applying physicians' final consensus diagnoses as proxies for AHT and non-AHT ground truth, the PediBIRN-7 performed with sensitivity 0.73 (95 % CI: 0.66-0.79), specificity 0.87 (95 % CI: 0.82-0.90), and ROC-AUC 0.90 (95 % CI: 0.87-0.94). Sensitivity analyses demonstrated minimal changes in rule performance. CONCLUSION: The PediBIRN-7's overall AHT prediction performance has been validated in a novel, equivalent, patient population. Its patient-specific estimates of AHT probability can inform physicians' AHT-related diagnostic reasoning after abuse evaluation.

An overview of log likelihood ratio cost in forensic science - Where is it used and what values can we expect?

There is increasing support for reporting evidential strength as a likelihood ratio (LR) and increasing interest in (semi-)automated LR systems. The log-likelihood ratio cost (Cllr) is a popular metric for such systems, penalizing misleading LRs further from 1 more. Cllr = 0 indicates perfection while Cllr = 1 indicates an uninformative system. However, beyond this, what constitutes a "good" Cllr is unclear. Aiming to provide handles on when a Cllr is "good", we studied 136 publications on (semi-)automated LR systems. Results show Cllr use heavily depends on the field, e.g., being absent in DNA analysis. Despite more publications on automated LR systems over time, the proportion reporting Cllr remains stable. Noticeably, Cllr values lack clear patterns and depend on the area, analysis and dataset. As LR systems become more prevalent, comparing them becomes crucial. This is hampered by different studies using different datasets. We advocate using public benchmark datasets to advance the field.

Measurement of Feline Alpha-1 Acid Glycoprotein in Serum and Effusion Using an ELISA Method: Analytical Validation and Diagnostic Role for Feline Infectious Peritonitis.

BACKGROUND: Alpha-1 acid glycoprotein (AGP) may support a clinical diagnosis of feline infectious peritonitis (FIP). In this study, we assessed the analytical and diagnostic performances of a novel ELISA method to measure feline AGP. METHODS: AGP was measured in sera and effusions from cats with FIP (n = 20) or with other diseases (n = 15). Precision was calculated based on the coefficient of variation (CV) of repeated testing, and accuracy was calculated by linearity under dilution (LUD). RESULTS: The test is precise (intra-assay CVs: <6.0% in individual samples, <15.0% in pooled samples; inter-assay CVs <11.0% and <15.0%) and accurate (serum LUD r2: 0.995; effusion LUD r2: 0.950) in serum and in effusions. AGP is higher in cats with FIP than in other cats in both serum (median: 1968, I-III interquartile range: 1216-3371 µg/mL and 296, 246-1963 µg/mL; p = 0.009) and effusion (1717, 1011-2379 µg/mL and 233, 165-566 µg/mL; p < 0.001). AGP discriminates FIP from other diseases (area under the receiver operating characteristic curve: serum, 0.760; effusion, 0.877), and its likelihood ratio is high (serum: 8.50 if AGP > 1590 µg/mL; effusion: 3.75 if AGP > 3780 µg/mL). CONCLUSION: This ELISA method is precise and accurate. AGP in serum and in effusions is a useful diagnostic marker for FIP.

From autoantibody test results to decision making: Incorporating likelihood ratios within medical practice.

Autoantibodies are important laboratory markers to support diagnosis of autoimmune diseases. Interpretation of autoantibodies is classically done in a dichotomous way (positive versus negative). Yet, interpretation of autoantibody test results can be improved by reporting likelihood ratios. Likelihood ratios convey information on how much more/less likely a test result is in individuals with the disease compared to individuals without the disease. It incorporates information on the antibody level (the higher the antibody level, the higher the association with the disease), which is helpful for (differential) diagnosis. Likelihood ratios are unit-independent and allow users to harmonize test result interpretation. When the likelihood ratio is combined with information on the pre-test probability, post-test probability can be appraised. In this review, the applicability of likelihood ratio in autoimmune diagnostics will be reviewed from the perspective of the clinician, the laboratory professional and the in vitro diagnostic industry.

Regional consistency assessment in multiregional clinical trials.

Multiregional clinical trials (MRCTs) have become a favored strategy for new drug development. The accurate evaluation of treatment effects across different regions is crucial for interpreting the results of MRCTs. Consistency between regional and overall results ensures the extrapolability of the overall conclusions to individual regions. While numerous statistical methods have been proposed for consistency assessment, a notable proportion necessitate a substantial escalation in sample size, particularly in scenarios involving more than four regions within MRCTs. This, paradoxically, undermines the fundamental intent of MRCTs. In addition, standardized statistical criteria for concluding consistency are yet to be established. In this paper, we develop further consistency assessment approaches in the framework of two multivariate likelihood ratio test-based methods, namely mLRTa and mLRTb, wherein consistency is cast as the alternative and null hypotheses. Notably, our exploration unveils that qualitative methods such as the funnel approach and PMDA methods are special instances of mLRTa. Furthermore, our work underscores that these three qualitative methodologies roughly share the same level of assurance probability (AP). Intriguingly, when the number of regions in an MRCT surpasses five, even when the overall sample size guarantees a power of 90% or more and the true treatment effects remain uniform across regions, the AP remains below the 70% mark. Drawing from our meticulous examination of operational attributes, we recommend mLRTa with positive treatment effects in all regions in the alternative hypothesis with significance level 0.5 or mLRTb with all regional treatment effects being equal in the null and significance level of 0.2.

The clinical utility of cystatin C based eGFR in assessing renal function among HIV/AIDs patients on ART at Mildmay Uganda.

BACKGROUND: In clinical practice, Measurement of estimated glomerular filtration rates (eGFR) is the gold standard assessing renal function the glomerular filtration rate often estimated from plasma creatinine. Several studies have shown Cystatin C based eGFR (Cys C) to be a better parameter for the diagnosis of impaired renal function. Cystatin C based eGFR has been proposed as a potential renal function marker but its use in HIV&AIDS patients has not been well evaluated. METHODS: A cross sectional study was carried out on 914 HIV&AIDS patients on antiretroviral therapy (ART) attending Mildmay Uganda for care and treatment between January to March 2015. Serum Cystatin C based eGFR was measured using the particle enhanced immunoturbidimetric assay. Creatinine was analyzed using enzymatic Creatinine PAP method and creatinine clearance was calculated according to C&G. RESULTS: The sensitivity of Cystatin C based eGFR was 15.1% (95% CI = 8.4, 24) with specificity 99.3% (95% CI = 98- 99.7). The positive and negative predictive values were 70.0% (95% CI 45.7-88.1) and 91.2% (95% CI 98.11-92.94) respectively. The positive likelihood ratio was 18.81 and negative likelihood ratio was 0.85. Cystatin C based eGFR had diagnostic accuracy of 90.7 and area under curve was 0.768. CONCLUSION: Cystatin C based eGFR exhibited a high specificity and a high positive likelihood ratio in diagnosis of kidney disease among HIV&AIDS patients. Cystatin C based eGFR can be used as a confirmatory test.

Bootstrap tests for simultaneous monotone ordering of effects in a two-way ANOVA.

In a two-way additive analysis of variance (ANOVA) model, we consider the problem of testing for homogeneity of both row and column effects against their simultaneous ordering. The error variances are assumed to be heterogeneous with unbalanced samples in each cell. Two simultaneous test procedures are developed-the first one using the likelihood ratio test (LRT) statistics of two independent hypotheses and another based on the consecutive pairwise differences of estimators of effects. The parametric bootstrap (PB) approach is used to find critical points of both the tests and the asymptotic accuracy of the bootstrap is established. An extensive simulation study shows that the proposed tests achieve the nominal size and have very good power performance. The robustness of the tests is also analyzed under deviation from normality. An "R" package is developed and shared on "GitHub" for ease of implementation of users. The proposed tests are illustrated using a real data set on the mortality due to alcoholic liver disease and it is shown that age and gender have a significant impact on the increasing incidence of mortality.

A Characterization of Most(More) Powerful Test Statistics with Simple Nonparametric Applications.

Data-driven most powerful tests are statistical hypothesis decision-making tools that deliver the greatest power against a fixed null hypothesis among all corresponding data-based tests of a given size. When the underlying data distributions are known, the likelihood ratio principle can be applied to conduct most powerful tests. Reversing this notion, we consider the following questions. (a) Assuming a test statistic, say T, is given, how can we transform T to improve the power of the test? (b) Can T be used to generate the most powerful test? (c) How does one compare test statistics with respect to an attribute of the desired most powerful decision-making procedure? To examine these questions, we propose one-to-one mapping of the term "most powerful" to the distribution properties of a given test statistic via matching characterization. This form of characterization has practical applicability and aligns well with the general principle of sufficiency. Findings indicate that to improve a given test, we can employ relevant ancillary statistics that do not have changes in their distributions with respect to tested hypotheses. As an example, the present method is illustrated by modifying the usual t-test under nonparametric settings. Numerical studies based on generated data and a real-data set confirm that the proposed approach can be useful in practice.

Research on likelihood ratio evaluation method of fingerprint evidence based on parameter estimation method.

Fingerprints with similar morphological characteristics but from different individuals can lead to errors in individual identification, especially when dealing with large databases containing millions of fingerprints. To address this issue and enhance the accuracy of similar fingerprint identification, the use of the likelihood ratio (LR) model for quantitative evaluation of fingerprint evidence has emerged as an effective research method. In this study, the LR fingerprint evidence evaluation model was established by using mathematical statistical methods, such as parameter estimation and hypothesis testing. This involved various steps, including database construction, scoring, fitting, calculation, and visual evaluation. Under the same-source conditions, the optimal parameter methods selected by different number of minutiae are gamma and Weibull distribution, while normal, Weibull, and lognormal distributions were the fitting parameters selected for minutiae configurations. The fitting parameters selected by different number of minutiae under different-source conditions are lognormal distribution, and the parameter methods selected for different minutiae configurations include Weibull, gamma, and lognormal distributions. The results of the LR model showed increased accuracy as the number of minutiae increased, indicating strong discriminative and corrective power. However, the accuracy of the LR evaluation based on different configurations was comparatively lower. In addition, the LR models with different numbers of minutiae outperformed those with different minutiae configurations. Our study shows that the use of LR models based on parametric methods is favoured in reducing the risk of fingerprint evidence misidentification, improving the quantitative assessment methods of fingerprint evidence, and promoting fingerprint identification from experience to science. Key points: Likelihood ratio (LR) method based on parameter estimation was applied to scientific evaluation of fingerprint evidence with excellent discriminatory and calibration capabilities.Both the number of minutiae and configuration of minutiae have significant effects on the score-based LR method.Fingerprints from the same source contain many different patterns of deformation.Databases containing 10 million fingerprints from different sources have been used for building the LR model.

Nonlinear Regression Modelling: A Primer with Applications and Caveats.

Use of nonlinear statistical methods and models are ubiquitous in scientific research. However, these methods may not be fully understood, and as demonstrated here, commonly-reported parameter p-values and confidence intervals may be inaccurate. The gentle introduction to nonlinear regression modelling and comprehensive illustrations given here provides applied researchers with the needed overview and tools to appreciate the nuances and breadth of these important methods. Since these methods build upon topics covered in first and second courses in applied statistics and predictive modelling, the target audience includes practitioners and students alike. To guide practitioners, we summarize, illustrate, develop, and extend nonlinear modelling methods, and underscore caveats of Wald statistics using basic illustrations and give key reasons for preferring likelihood methods. Parameter profiling in multiparameter models and exact or near-exact versus approximate likelihood methods are discussed and curvature measures are connected with the failure of the Wald approximations regularly used in statistical software. The discussion in the main paper has been kept at an introductory level and it can be covered on a first reading; additional details given in the Appendices can be worked through upon further study. The associated online Supplementary Information also provides the data and R computer code which can be easily adapted to aid researchers to fit nonlinear models to their data.

Inferring a directed acyclic graph of phenotypes from GWAS summary statistics.

Estimating phenotype networks is a growing field in computational biology. It deepens the understanding of disease etiology and is useful in many applications. In this study, we present a method that constructs a phenotype network by assuming a Gaussian linear structure model embedding a directed acyclic graph (DAG). We utilize genetic variants as instrumental variables and show how our method only requires access to summary statistics from a genome-wide association study (GWAS) and a reference panel of genotype data. Besides estimation, a distinct feature of the method is its summary statistics-based likelihood ratio test on directed edges. We applied our method to estimate a causal network of 29 cardiovascular-related proteins and linked the estimated network to Alzheimer's disease (AD). A simulation study was conducted to demonstrate the effectiveness of this method. An R package sumdag implementing the proposed method, all relevant code, and a Shiny application are available.

Improved likelihood ratios for face recognition in surveillance video by multimodal feature pairing.

In forensic and security scenarios, accurate facial recognition in surveillance videos, often challenged by variations in pose, illumination, and expression, is essential. Traditional manual comparison methods lack standardization, revealing a critical gap in evidence reliability. We propose an enhanced images-to-video recognition approach, pairing facial images with attributes like pose and quality. Utilizing datasets such as ENFSI 2015, SCFace, XQLFW, ChokePoint, and ForenFace, we assess evidence strength using calibration methods for likelihood ratio estimation. Three models-ArcFace, FaceNet, and QMagFace-undergo validation, with the log-likelihood ratio cost (Cllr) as a key metric. Results indicate that prioritizing high-quality frames and aligning attributes with reference images optimizes recognition, yielding similar Cllr values to the top 25% best frames approach. A combined embedding weighted by frame quality emerges as the second-best method. Upon preprocessing facial images with the super resolution CodeFormer, it unexpectedly increased Cllr, undermining evidence reliability, advising against its use in such forensic applications.

Bayesian spatial cluster signal learning with application to adverse event (AE).

There is growing interest in understanding geographic patterns of medical device-related adverse events (AEs). A spatial scan method combined with the likelihood ratio test (LRT) for spatial-cluster signal detection over the geographical region is universally used. The spatial scan method used a moving window to scan the entire study region and collected some candidate sub-regions from which the spatial-cluster signal(s) will be found. However, it has some challenges, especially in computation. First, the computational cost increased when the number of sub-regions increased. Second, the computational cost may increase if a large spatial-cluster pattern is present and a flexible-shaped window is used. To reduce the computational cost, we propose a Bayesian nonparametric method that combines the ideas of Markov random field (MRF) to leverage geographical information to find potential signal clusters. Then, the LRT is applied for the detection of spatial cluster signals. The proposed method provides an ability to capture both locally spatially contiguous clusters and globally discontiguous clusters, and is manifested to be effective and tractable using hypothetical Left Ventricular Assist Device (LVAD) data as an illustration.

Fuzzy information recognition and translation processing in English interpretation based on a generalized maximum likelihood ratio algorithm.

English interpretation plays a vital role as a critical link in cross-language communication. However, there are various types of ambiguous information in many interpreting scenarios, such as ambiguity, ambiguous vocabulary, and syntactic structures, which may lead to inaccuracies and fluency issues in translation. This article proposes a method based on the generalized maximum likelihood ratio algorithm (GLR) to identify and process fuzzy information in English interpretation to improve the quality and efficiency of performance. Firstly, we systematically analyzed the common types of fuzzy information in interpretation and delved into the basic principles and applications of the generalized maximum likelihood ratio algorithm. This algorithm is widely used in natural language processing to solve uncertainty problems and has robust modeling and inference capabilities, making it suitable for handling fuzzy information in interpretation. Then, we propose a fuzzy information recognition model based on the generalized maximum likelihood ratio algorithm. This model utilizes a large-scale interpretation corpus for training and identifies potential fuzzy information in the interpretation process through statistical analysis and pattern recognition. Once fuzzy information is detected, we adopt a series of effective translation processing strategies, including contextual inference and adaptation, to ensure the accuracy and naturalness of interpretation. Finally, we conducted a series of experiments to evaluate the performance of the proposed method. The experimental results show that the fuzzy information recognition and translation processing method based on the generalized maximum likelihood ratio algorithm performs well in different interpretation scenarios, significantly improving the quality and fluency of interpretation and reducing ambiguity caused by fuzzy information.

From data to a validated score-based LR system: A practitioner's guide.

Likelihood ratios (LRs) are a useful measure of evidential strength. In forensic casework consisting of a flow of cases with essentially the same question and the same analysis method, it is feasible to construct an 'LR system', that is, an automated procedure that has the observations as input and an LR as output. This paper is aimed at practitioners interested in building their own LR systems. It gives an overview of the different steps needed to get to a validated LR system from data. The paper is accompanied by a notebook that illustrates each step with an example using glass data. The notebook introduces open-source software in Python constructed by NFI (Netherlands Forensic Institute) data scientists and statisticians.

Similarity quantification of bare footprint based on linear measurement and shape context contour combined score-based likelihood ratio evaluation.

Bare footprints are a type of physical evidence that can be crucial for solving difficult cases. To assist with identification, we explored a similarity quantification method using 3062 bare footprints from 1391 volunteers. We used linear similarity to cover seven linear lengths and then calculated the contour similarity of the heel region, anterior margin, and toe region using the shape context. Discriminant analysis was applied to determine the weights of the linear and contour similarities. Linear similarity had a weight of 0.56 whereas contour similarity had a weight of 0.44. The similarity between the same source and non-matches bare footprints was significantly different, with a leave-one-out cross-validation accuracy of 98.8%. Using the constructed similarity model, we developed a score-based likelihood ratio model based on similarity scores. We applied this model to five representative test samples including different volunteers, six months apart bare footprints, dynamic walking and static bare footprints. This method eliminated the interference of motion states and allowed for accurate determination of the same source and non-matches test samples. Overall, we quantified the shape contour and established a similarity assessment system for bare footprints that can assist in evaluation and identification.

An approach to unified formulae for likelihood ratio calculation in pairwise kinship analysis.

Introduction: The likelihood ratio (LR) can be an efficient means of distinguishing various relationships in forensic fields. However, traditional list-based methods for derivation and presentation of LRs in distant or complex relationships hinder code editing and software programming. This paper proposes an approach for a unified formula for LRs, in which differences in participants' genotype combinations can be ignored for specific identification. This formula could reduce the difficulty of by-hand coding, as well as running time of large-sample-size simulation. Methods: The approach is first applied to a problem of kinship identification in which at least one of the participants is alleged to be inbred. This can be divided into two parts: i) the probability of different identical by descent (IBD) states according to the alleged kinship; and ii) the ratio of the probability that specific genotype combination can be detected assuming the alleged kinship exists between the two participants to the similar probability assuming that they are unrelated, for each state. For the probability, there are usually recognized results for common identification purposes. For the ratio, subscript letters representing IBD alleles of individual A's alleles are used to eliminate differences in genotype combinations between the two individuals and to obtain a unified formula for the ratio in each state. The unification is further simplified for identification cases in which it is alleged that both of the participants are outbred. Verification is performed to show that the results obtained with the unified and list-form formulae are equivalent. Results: A series of unified formulae are derived for different identification purposes, based on which an R package named KINSIMU has been developed and evaluated for use in large-size simulations for kinship analysis. Comparison between the package with two existing tools indicated that the unified approach presented here is more convenient and time-saving with respect to the coding process for computer applications compared with the list-based approach, despite appearing more complicated. Moreover, the method of derivation could be extended to other identification problems, such as those with different hypothesis sets or those involving multiple individuals. Conclusion: The unified approach of LR calculation can be beneficial in kinship identification field.

Assessment of the diagnostic performance of intradermal tuberculin test and post-mortem inspection for the diagnosis of bovine tuberculosis according to WOAH guidelines.

Bovine tuberculosis (bTB) constitutes a global challenge for public and animal health with still some deficiencies regarding its diagnosis. This study aimed to estimate the accuracy of the single intradermal tuberculin test (SIT) and post-mortem inspection for different diagnostic objectives following WOAH guidelines. Tissue samples from 59 microbiological culture/PCR-positive and 58 microbiological culture/PCR-negative cattle were evaluated. The diagnostic sensitivity and specificity, the positive and negative probability indices as well as the positive and negative predictive values (PPV and NPV) of each technique were estimated for different pretest probabilities. The SIT with strict interpretation demonstrated moderate precision in confirming the absence of infection in populations historically free of bTB, with a 12.1% rate of false positives, but also detecting positive animals in the early stage of the eradication programs, with a 13.6% rate of false negatives. The diagnostic performance for ruling out bTB was notably high (NPV > 90%) in animals with a pre-test probability (PTP) below 42%. Post-mortem inspection constituted an interesting alternative tool to confirm suspected and positive cases for SIT, particularly in areas with bTB prevalence exceeding 19%, where implementing SIT and eradication measures may be impractical. In these areas, the likelihood that animals with tuberculosis-like lesions are affected by the disease surpasses 90%. Similarly, in herds with a PTP below 25%, the absence of bTB could be confidently ruled out with over 90% certainty. These findings highlight the effectiveness of SIT and post-mortem inspection as valuable techniques for current eradication programs and controlling bTB in high-prevalence areas where molecular techniques may not be feasible.

Likelihood ratio combination of multiple biomarkers via smoothing spline estimated densities.

The diagnostic accuracy of multiple biomarkers in medical research is crucial for detecting diseases and predicting patient outcomes. An optimal method for combining these biomarkers is essential to maximize the Area Under the Receiver Operating Characteristic (ROC) Curve (AUC). Although the optimality of the likelihood ratio has been proven by Neyman and Pearson, challenges persist in estimating the likelihood ratio, primarily due to the estimation of multivariate density functions. In this study, we propose a non-parametric approach for estimating multivariate density functions by utilizing Smoothing Spline density estimation to approximate the full likelihood function for both diseased and non-diseased groups, which compose the likelihood ratio. Simulation results demonstrate the efficiency of our method compared to other biomarker combination techniques under various settings for generated biomarker values. Additionally, we apply the proposed method to a real-world study aimed at detecting childhood autism spectrum disorder (ASD), showcasing its practical relevance and potential for future applications in medical research.

Estimation of additive and maternal covariance of production traits in Murrah buffalo.

The study was done to determine additive, maternal and common permanent environmental effects and best-suited model for some production traits using six univariate animal models that differed in the (co)variance components fitted to assess the importance of maternal effect using likelihood ratio test in Murrah buffaloes. Data from 614 Murrah buffaloes related to production traits were collected from history pedigree sheets maintained at the buffalo farm, Department of Livestock Production and Management (LPM), LUVAS, Hisar. The production traits under this study were 305 days milk yield (305DMY), peak yield (PY), lactation length (LL), dry period (DP), lactation milk yield (LMY) and wet average (WA). The heritability estimates were in the range of 0.33-0.44 for 305DMY, 0.25-0.51 for PY, 0.05-0.13 for LL, 0.03-0.23 for DP, 0.17-0.40 for LMY and 0.37-0.66 for WA. Model 1 was considered best for most of the traits, viz., 305DMY, PY, LL, LMY and WA followed by model 2 for DP. Covariance and correlated values within the traits caused inflation of heritability in model 3 and model 6. The maximum covariance between the additive and maternal effect was found in trait LMY, which was 14,183.90 in model 3 and the minimum value was also reported in the same trait for model 6, valued at -3522.37. Multivariate analysis showed that all production traits were moderate to high and positively correlated with each other except for DP, which was low and negative genetic and phenotypic correlated. Spearman's rank correlation coefficients of breeding value among all six models were high and significant, ranged from 0.78 to 1.00 for all the traits except DP, therefore any of the models could be taken into account depending upon the availability of data.