Category: Statistics

JCO Oncol Pract. 2022 May 24:OP2100748. doi: 10.1200/OP.21.00748. Online ahead of print.

ABSTRACT

PURPOSE: The Cannabis Consultation Service (CCS) is an innovative pharmacist-led resource at the Sunnybrook Odette Cancer Centre. Its mandate is to provide education and guide patients through access and appropriate use of high-quality plant-derived cannabinoids (PDCs). Our objective was to describe the CCS, explain its processes, and characterize patient disposition with respect to use of PDCs.

METHODS: We retrospectively reviewed the charts of patients referred to the CCS from July 13, 2020, to March 05, 2021. We used descriptive statistics to report on the patient population and service metrics.

RESULTS: During the 34-week period, 96 patients accessed the CCS (median age, 61 years). The top reasons for CCS consultation were management of cancer pain, insomnia, and general interest. Medical cannabis was supported as an option in 44/96 patients. Reasons for not supporting PDC use included lack of indication, potential drug interaction/contraindication, or requiring treatment with first-line therapy. Of the 40 patients requiring a medical document, 22 initiated therapy. The most common product used was a 2:50 THC:CBD (Tetrahydrocannabinol:Cannabidiol) cannabis oil. At the date of last contact, few patients remained on therapy because of lack of benefit, patient choice, and/or hesitancy.

CONCLUSION: Despite patients with cancer having interest in seeking PDCs for symptom management, only a few initiated and continued therapy. Pharmacists have an opportunity to advise patients and the oncology team on the risks and benefits of PDCs. These results can be used to support the development of medical cannabis programs by oncology centers and focus future research priorities.

PMID:35609230 | DOI:10.1200/OP.21.00748

Proc Natl Acad Sci U S A. 2022 May 31;119(22):e2118636119. doi: 10.1073/pnas.2118636119. Epub 2022 May 24.

ABSTRACT

SignificanceRandom Forests (RFs) are among the most successful machine-learning algorithms in terms of prediction accuracy. In many domain problems, however, the primary goal is not prediction, but to understand the data-generation process-in particular, finding important features and feature interactions. There exists strong empirical evidence that RF-based methods-in particular, iterative RF (iRF)-are very successful in terms of detecting feature interactions. In this work, we propose a biologically motivated, Boolean interaction model. Using this model, we complement the existing empirical evidence with theoretical evidence for the ability of iRF-type methods to select desirable interactions. Our theoretical analysis also yields deeper insights into the general interaction selection mechanism of decision-tree algorithms and the importance of feature subsampling.

PMID:35609192 | DOI:10.1073/pnas.2118636119

ASAIO J. 2022 May 24. doi: 10.1097/MAT.0000000000001775. Online ahead of print.

ABSTRACT

A retrospective study was performed examining the trend of inflammatory markers, including D-dimers, in 29 COVID-19 patients requiring veno-venous (VV) extracorporeal membrane oxygenation (ECMO) support. We observed that COVID-19 patients with pre-cannulation D-dimer levels >3,000 ng/mL had a significantly shorter time from admission to cannulation (4.78 vs. 8.44 days, p = 0.049) compared to those with D-dimer <3,000 ng/mL. Furthermore, patients with D-dimer >3,000 ng/mL had a trend of lower pH (7.24 vs. 7.33), higher pCO2 (61.33 vs. 50.69), and higher vasoactive inotropic score (7.23 vs. 3.97) at time of cannulation, however, these were not statistically significant. This cohort of patients also required a longer duration of ECMO support (51.44 vs. 31.25 days). However, 13 patients required at least one ECMO-circuit exchange and 16 patients did not require any exchanges. There was a consistent drop in D-dimer values after every circuit exchange, which was not observed in any of the other examined inflammatory markers, including ferritin, lactate dehydrogenase, or C-reactive protein. We propose that elevated D-dimer levels (>3,000 ng/mL) reflect increased disease severity in COVID-19, and predict a longer ECMO course. Once on ECMO, however, the D-Dimer level consistently decreased with every circuit exchange, which may reflect thrombus within the oxygenator rather than just disease severity.

PMID:35609187 | DOI:10.1097/MAT.0000000000001775

Perm J. 2021 Oct 25;26(1):11-20. doi: 10.7812/TPP/21.024.

ABSTRACT

INTRODUCTION: Effective, equity-promoting interventions implemented by health care systems are needed to address health care disparities and population-level health disparities. We evaluated the impact of a clinical decision support tool to improve evidence-based thiazide diuretic prescribing among Black patients to address racial disparities in hypertension control.

METHODS: We employed an interrupted time series design and qualitative interviews to evaluate the implementation of the tool. Our primary outcome measure was the monthly rate of thiazide use among eligible patients before and after implementation of the tool (January 2013-December 2016). We modeled month-to-month changes in thiazide use for Black and White patients, overall, and by sex and medical center racial composition. We conducted key informant interviews to identify modifiable facilitators and barriers to implementation of the tool across medical centers.

RESULTS: Of the 318,720 patients, 15.5% were Black. We observed no change in thiazide use or blood pressure control following the implementation of the tool in either racial subgroup. There was a slight but statistically significant reduction (2.32 percentage points, p < 0.01) in thiazide use among Black patients following the removal the tool that was not observed among White patients. Factors affecting the tool’s implementation included physician and pharmacist resistance to thiazide use and a lack of ongoing promotion of the tool.

DISCUSSION: The clinical decision support tool was insufficient to change prescribing practices and improve blood pressure control among Black patients.

CONCLUSIONS: Future interventions should consider physician attitudes about thiazide prescribing and the importance of multilevel approaches to address hypertension disparities.

PMID:35609161 | DOI:10.7812/TPP/21.024

Transplantation. 2022 May 9. doi: 10.1097/TP.0000000000004171. Online ahead of print.

ABSTRACT

BACKGROUND: Much of our understanding regarding geographic issues in transplantation is based on statistical techniques that do not formally account for geography and is based on obsolete boundaries such as donation service area.

METHODS: We applied spatial epidemiological techniques to analyze liver-related mortality and access to liver transplant services at the county level using data from the Centers for Disease Control and Prevention and Scientific Registry of Transplant Recipients from 2010 to 2018.

RESULTS: There was a significant negative spatial correlation between transplant rates and liver-related mortality at the county level (Moran’s I, -0.319; P = 0.001). Significant clusters were identified with high transplant rates and low liver-related mortality. Counties in geographic clusters with high ratios of liver transplants to liver-related deaths had more liver transplant centers within 150 nautical miles (6.7 versus 3.6 centers; P < 0.001) compared with all other counties, as did counties in geographic clusters with high ratios of waitlist additions to liver-related deaths (8.5 versus 2.5 centers; P < 0.001). The spatial correlation between waitlist mortality and overall liver-related mortality was positive (Moran’s I, 0.060; P = 0.001) but weaker. Several areas with high waitlist mortality had some of the lowest overall liver-related mortality in the country.

CONCLUSIONS: These data suggest that high waitlist mortality and allocation model for end-stage liver disease do not necessarily correlate with decreased access to transplant, whereas local transplant center density is associated with better access to waitlisting and transplant.

PMID:35609185 | DOI:10.1097/TP.0000000000004171

Int J Dermatol. 2022 May 24. doi: 10.1111/ijd.16278. Online ahead of print.

ABSTRACT

BACKGROUND: Erysipeloid cutaneous leishmaniasis (ECL) is known as the chronic form of cutaneous leishmaniasis (CL). However, keeping its clinical presentation in view, there is a need to revisit this form of the disease.

AIMS: To describe ECL in view of clinical features and treatment modalities.

METHODS: We include a case series seen in Sfax (Southern Tunisia) from January 2017 to January 2021. All patients clinically suggestive and laboratory confirmed with a diagnosis of CL were registered. Patients of all age groups and of either gender having cutaneous lesions resembling erysipela on the face were included in the study. Different demographic features of the patients and clinical aspects were identified. Descriptive statistics were used for analysis.

RESULTS: Of 1300 registered patients with CL, 40 (3%) were diagnosed as ECL. Ages ranged from 15 to 65 years, and duration of lesions varied from 15 to 180 days. All patients had lesions over the face. Clinically, a painful infiltrated inflammatory placard of the central facial area with a butterfly shape was observed in 14 cases, as well as zones of the cheekbone (11 cases), cheekbone and nose (5 cases), cheekbone and eyelid (8 cases), and cheekbone with ear (2 cases). Several therapeutic methods were prescribed with a sufficient result with no recurrence.

CONCLUSION: ECL is a rare presentation that typically occurs on the face, looking like erysipelas, in patients who are native from an endemic region of CL.

PMID:35609142 | DOI:10.1111/ijd.16278

IEEE Trans Biomed Eng. 2022 May 24;PP. doi: 10.1109/TBME.2022.3176224. Online ahead of print.

ABSTRACT

Fluorescence lifetime imaging is a valuable technique for probing characteristics of wide ranging samples and sensing of the molecular environment. However, the desire to measure faster and reduce effects such as photo bleaching in optical photon-count measurements for lifetime estimation lead to inevitable effects of convolution with the instrument response functions and noise, causing a degradation of the lifetime accuracy and precision. To tackle the problem, this paper presents a robust and computationally efficient framework for recovering fluorophore sample decay from the histogram of photon-count arrivals modelled as a decaying single-exponential function. In the proposed approach, the temporal histogram data is first decomposed into multiple bins via an adaptive multi-bin signal representation. Then, at each level of the multi-resolution temporal space, decay information including both the amplitude and the lifetime of a single-exponential function is rapidly decoded based on a novel statistical estimator. Ultimately, a game-theoretic model consisting of two players in an “amplitude-lifetime” game is constructed to be able to robustly recover optimal fluorescence decay signal from a set of fused multi-bin estimates. In addition to theoretical demonstrations, the efficiency of the proposed framework is experimentally shown on both synthesised and real data in different imaging circumstances. On a challenging low photon-count regime, our approach achieves about 28% improvement in bias than the best competing method. On real images, the proposed method processes data on average around 63 times faster than the gold standard least squares fit. Implementation codes are available to researchers.

PMID:35609109 | DOI:10.1109/TBME.2022.3176224

IEEE Trans Cybern. 2022 May 24;PP. doi: 10.1109/TCYB.2022.3166975. Online ahead of print.

ABSTRACT

Clusters in real data are often restricted to low-dimensional subspaces rather than the entire feature space. Recent approaches to circumvent this difficulty are often computationally inefficient and lack theoretical justification in terms of their large-sample behavior. This article deals with the problem by introducing an entropy incentive term to efficiently learn the feature importance within the framework of center-based clustering. A scalable block-coordinate descent algorithm, with closed-form updates, is incorporated to minimize the proposed objective function. We establish theoretical guarantees on our method by Vapnik-Chervonenkis (VC) theory to establish strong consistency along with uniform concentration bounds. The merits of our method are showcased through detailed experimental analysis on toy examples as well as real data clustering benchmarks.

PMID:35609103 | DOI:10.1109/TCYB.2022.3166975

IEEE Trans Image Process. 2022 May 24;PP. doi: 10.1109/TIP.2022.3176210. Online ahead of print.

ABSTRACT

Denoising videos in real-time is critical in many applications, including robotics and medicine, where varying-light conditions, miniaturized sensors, and optics can substantially compromise image quality. This work proposes the first video denoising method based on a deep neural network that achieves state-of-the-art performance on dynamic scenes while running in real-time on VGA video resolution with no frame latency. The backbone of our method is a novel, remarkably simple, temporal network of cascaded blocks with forward block output propagation. We train our architecture with short, long, and global residual connections by minimizing the restoration loss of pairs of frames, leading to a more effective training across noise levels. It is robust to heavy noise following Poisson-Gaussian noise statistics. The algorithm is evaluated on RAW and RGB data. We propose a denoising algorithm that requires no future frames to denoise a current frame, reducing its latency considerably. The visual and quantitative results show that our algorithm achieves state-of-the-art performance among efficient algorithms, achieving from two-fold to two-orders-of-magnitude speed-ups on standard benchmarks for video denoising.

PMID:35609095 | DOI:10.1109/TIP.2022.3176210

PLoS One. 2022 May 24;17(5):e0268472. doi: 10.1371/journal.pone.0268472. eCollection 2022.

ABSTRACT

BACKGROUND: Economically underdeveloped areas in western China are hotspots of tuberculosis, especially among students. However, the related spatial and temporal patterns and influencing factors are still unclear and there are few studies to analyze the causes of pulmonary tuberculosis in students from the perspective of space.

METHODS: We collected data regarding the reported incidence of pulmonary tuberculosis (PTB) among students at township level in Nanning, from 2012 to 2018. The reported incidence of pulmonary tuberculosis among students in Nanning was analyzed using spatial autocorrelation and spatial scan statistical analysis to depict hotspots of PTB incidence and spatial and temporal clustering. Spatial panel data of the reported incidence rates and influencing factors at district and county levels in Nanning were collected from 2015 to 2018. Then, we analyzed the spatial effects of incidence and influencing factors using the spatial Durbin model to explore the mechanism of each influencing factor in areas with high disease prevalence under spatial effects.

RESULTS: From 2012 to 2018, 1609 cases of PTB were reported among students in Nanning, with an average annual reported incidence rate of 14.84/100,000. Through the Joinpoint regression model, We observed a steady trend in the percentage of cases reported each year (P>0.05). There was spatial autocorrelation between the annual reported incidence and the seven-years average reported incidence from 2012 to 2018. The high-incidence area was distributed in the junction of six urban areas and spread to the periphery, with the junction at the center. The population of college students, per capita financial expenditure on health, per capita gross domestic product, and the number of health technicians per 1,000 population were all influencing factors in the reported incidence of PTB among students.

CONCLUSION: We identified spatial clustering of the reported incidence of PTB among students in Nanning, mainly located in the urban center and its surrounding areas. The clustering gradually decreased from the urban center to the surrounding areas. Spatial effects influenced the reported incidence of PTB. The population density of college students, per capita health financial expenditure, gross domestic product (GDP) per capita, and the number of health technicians per 1,000 were all influencing factors in the reported incidence of PTB among students.

PMID:35609085 | DOI:10.1371/journal.pone.0268472