Try a new search

Format these results:

Searched for:

person:fangy03

in-biosketch:yes

Total Results:

93


Sparse Convex Clustering

Wang, Binhuan; Zhang, Yilong; Sun, Will Wei; Fang, Yixin
u Convex clustering, a convex relaxation of k-means clustering and hierarchical clustering, has drawn recent attentions since it nicely addresses the instability issue of traditional nonconvex clustering methods. Although its computational and statistical properties have been recently studied, the performance of convex clustering has not yet been investigated in the high-dimensional clustering scenario, where the data contains a large number of features and many of them carry no information about the clustering structure. In this article, we demonstrate that the performance of convex clustering could be distorted when the uninformative features are included in the clustering. To overcome it, we introduce a new clustering method, referred to as Sparse Convex Clustering, to simultaneously cluster observations and conduct feature selection. The key idea is to formulate convex clustering in a form of regularization, with an adaptive group-lasso penalty term on cluster centers. To optimally balance the trade-off between the cluster fitting and sparsity, a tuning criterion based on clustering stability is developed. Theoretically, we obtain a finite sample error bound for our estimator and further establish its variable selection consistency. The effectiveness of the proposed method is examined through a variety of numerical experiments and a real data application. Supplementary material for this article is available online.
ISI:000435688200013
ISSN: 1061-8600
CID: 3182282

Online Bootstrap Confidence Intervals for the Stochastic Gradient Descent Estimator

Fang, Yixin; Xu, Jinfeng; Yang, Lei
In many applications involving large dataset or online learning, stochastic gradient descent (SGD) is a scalable algorithm to compute parameter estimates and has gained increasing popularity due to its numerical convenience and memory efficiency. While the asymptotic properties of SGD-based estimators have been well established, statistical inference such as interval estimation remains much unexplored. The classical bootstrap is not directly applicable if the data are not stored in memory. The plug-in method is not applicable when there is no explicit formula for the covariance matrix of the estimator. In this paper, we propose an online bootstrap procedure for the estimation of confidence intervals, which, upon the arrival of each observation, updates the SGD estimate as well as a number of randomly perturbed SGD estimates. The proposed method is easy to implement in practice. We establish its theoretical properties for a general class of models that includes linear regressions, generalized linear models, M-estimators and quantile regressions as special cases. The finite-sample performance and numerical utility is evaluated by simulation studies and real data applications.
ISI:000454479400001
ISSN: 1532-4435
CID: 3575252

Changes in Peak Airflow Measurement During Maximal Cough After Vocal Fold Augmentation in Patients With Glottic Insufficiency

Dion, Gregory R; Achlatis, Efstratios; Teng, Stephanie; Fang, Yixin; Persky, Michael; Branski, Ryan C; Amin, Milan R
Importance/UNASSIGNED:Compromised cough effectiveness is correlated with dysphagia and aspiration. Glottic insufficiency likely yields decreased cough strength and effectiveness. Although vocal fold augmentation favorably affects voice and likely improves cough strength, few data exist to support this hypothesis. Objective/UNASSIGNED:To assess whether vocal fold augmentation improves peak airflow measurements during maximal-effort cough following augmentation. Design, Setting, and Participants/UNASSIGNED:This case series study was conducted in a tertiary, academic laryngology clinic. Participants included 14 consecutive individuals with glottic insufficiency due to vocal fold paralysis, which was diagnosed via videostrobolaryngoscopy as a component of routine clinical examination. All participants who chose to proceed with augmentation were considered for the study whether office-based or operative augmentation was planned. Postaugmentation data were collected only at the first follow-up visit, which was targeted for 14 days after augmentation but varied on the basis of participant availability. Data were collected from June 5, 2014, to October 1, 2015. Data analysis took place between October 2, 2015, and March 3, 2017. Main Outcomes and Measures/UNASSIGNED:Peak airflow during maximal volitional cough was quantified before and after vocal fold augmentation. Participants performed maximal coughs, and peak expiratory flow during the maximal cough was captured according to American Thoracic Society guidelines. Results/UNASSIGNED:Among the 14 participants (7 men and 7 women), the mean (SD) age was 62 (18) years. Three types of injectable material were used for vocal fold augmentation: carboxymethylcellulose in 5 patients, hyaluronic acid in 5, and calcium hydroxylapatite in 4. Following augmentation, cough strength increased in 11 participants and decreased cough strength was observed in 3. Peak airflow measurements during maximal cough varied from a decrease of 40 L/min to an increase of 150 L/min following augmentation. When preaugmentation and postaugmentation peak airflow measurements were compared, the median improvement was 50 L/min (95% CI, 10-75 L/min; P = .01). Immediate peak airflow measurements during cough collected within 30 minutes of augmentation varied when compared with measurements collected at follow-up (103-380 vs 160-390 L/min). Conclusions and Relevance/UNASSIGNED:Peak airflow during maximal cough may improve with vocal fold augmentation. Additional assessment and measurements are needed to further delineate which patients will benefit most regarding their cough from vocal fold augmentation.
PMCID:5710351
PMID: 28715529
ISSN: 2168-619x
CID: 2919422

Impact of diabetes on heart failure incidence in adults with ischemic heart disease

Patel, Nirav; Chen, Olivia; Donahue, Caroline; Wang, Binhuan; Fang, Yixin; Donnino, Robert; Natarajan, Sundar
BACKGROUND: Ischemic heart disease (IHD) is the most potent risk factor for heart failure (HF). Our study aims to evaluate the incremental impact of diabetes on the incidence of HF in individuals with IHD. METHODS: Data from the NHANES Epidemiologic Follow-Up Study (Baseline: 1971 to 1974) were linked to the facility and mortality files up to 1992. Our analyses were restricted to patients with IHD without prevalent HF at baseline. The cumulative incidence of HF in patients with diabetes and IHD versus those with IHD alone was assessed using failure curves. Cox proportional hazards models were used to control for important covariates. All analyses incorporated the complex sample design by including the weights and clustering variables. RESULTS: Out of the 14,407 participants, 497 had IHD without prevalent HF and had information about diabetes status. Among these participants, the cumulative incidence of HF was 38.1% for those with diabetes (n=63) and 26.5% in those without diabetes (n=434) (log-rank p-value<0.005). The multivariate hazard ratio (adjusted for age, BMI, alcohol consumption, hypertension, high cholesterol, and smoking) for incident HF for people who had myocardial infarction (MI) and diabetes compared to people who had MI alone was 2.98 (95% CI 1.51, 5.88). CONCLUSION: Among participants with MI, those with diabetes had a substantially higher incidence of HF than those without diabetes. Based on these findings, practitioners should focus greater attention on patients with diabetes and previous MI in order to potentially prevent incident HF.
PMID: 28947278
ISSN: 1873-460x
CID: 2717702

The value of intraoperative EABRs in auditory brainstem implantation

Anwar, Abbas; Singleton, Alison; Fang, Yixin; Wang, Binhuan; Shapiro, William; Roland, J Thomas Jr; Waltzman, Susan B
OBJECTIVE: To compare the intraoperative electrically evoked auditory brainstem response (EABR) morphologies between neurofibromatosis II (NF2) adult auditory brainstem implant (ABI) recipients who had auditory percepts post-operatively and those who did not and between NF2 adult ABI recipients and non-NF2 pediatric ABI recipients. METHODS: This was a retrospective case series at a single tertiary academic referral center examining all ABI recipients from 1994 to 2016, which included 34 NF2 adults and 11 non-NF2 children. The morphologies of intraoperative EABRs were evaluated for the number of waveforms showing a response, the number of positive peaks in those responses, and the latencies of each of these peaks. RESULTS: 27/34 adult NF2 patients and 9/10 children had EABR waveforms. 20/27 (74.0%) of the adult patients and all of the children had ABI devices that stimulated post-operatively. When comparing the waveforms between adults who stimulated and those who did not stimulate, the proportion of total number of intraoperative EABR peaks to total possible peaks was significantly higher for the adults who stimulated than for those who did not (p < 0.05). Children had a significantly higher proportion of total number of peaks to total possible peaks when compared to adults who stimulated (p < 0.02). Additionally, there were more likely to be EABR responses at the initial stimulation than intraoperatively in the pediatric ABI population (p = 0.065). CONCLUSIONS: The value of intraoperative EABR tracing may lie in its ability to predict post-operative auditory percepts based on the placement of the array providing the highest number of total peaks.
PMID: 28964288
ISSN: 1872-8464
CID: 2719782

Impact of Diabetes on Heart Failure Incidence in Adults with Ischemic Heart Disease [Meeting Abstract]

Patel, Nirav; Chen, Olivia; Donahue, Caroline; Wang, Binhuan; Fang, Yixin; Donnio, Robert; Natarajan, Sundar
ISI:000408403100196
ISSN: 1532-8414
CID: 2696062

What are the Patterns Between Depression, Smoking, Unhealthy Alcohol Use, and Other Substance Use Among Individuals Receiving Medical Care? A Longitudinal Study of 5479 Participants

Ruggles, Kelly V; Fang, Yixin; Tate, Janet; Mentor, Sherry M; Bryant, Kendall J; Fiellin, David A; Justice, Amy C; Braithwaite, R Scott
To evaluate and characterize the structure of temporal patterns of depression, smoking, unhealthy alcohol use, and other substance use among individuals receiving medical care, and to inform discussion about whether integrated screening and treatment strategies for these conditions are warranted. Using the Veterans Aging Cohort Study (VACS) we measured depression, smoking, unhealthy alcohol use and other substance use (stimulants, marijuana, heroin, opioids) and evaluated which conditions tended to co-occur within individuals, and how this co-occurrence was temporally structured (i.e. concurrently, sequentially, or discordantly). Current depression was associated with current use of every substance examined with the exception of unhealthy alcohol use. Current unhealthy alcohol use and marijuana use were also consistently associated. Current status was strongly predicted by prior status (p < 0.0001; OR = 2.99-22.34) however, there were few other sequential relationships. Associations in the HIV infected and uninfected subgroups were largely the same with the following exceptions. Smoking preceded unhealthy alcohol use and current smoking was associated with current depression in the HIV infected subgroup only (p < 0.001; OR = 1.33-1.41 and p < 0.001; OR = 1.25-1.43). Opioid use and current unhealthy alcohol use were negatively associated only in the HIV negative subgroup (p = 0.01; OR = 0.75). Patterns of depression, smoking, unhealthy alcohol use, and other substance use were temporally concordant, particularly with regard to depression and substance use. These patterns may inform future development of more integrated screening and treatment strategies.
PMCID:5542002
PMID: 27475945
ISSN: 1573-3254
CID: 2199312

Treatment of Neurogenic Cough with Tramadol: A Pilot Study

Dion, Gregory R; Teng, Stephanie E; Achlatis, Efstratios; Fang, Yixin; Amin, Milan R
This study employs validated cough assessment tools to prospectively determine the impact of tramadol on cough severity and quality of life in subjects with neurogenic cough. The study was a prospective case series with planned data collection at a tertiary care academic medical center laryngology practice. Sixteen consecutive collected subjects with neurogenic cough prospectively completed pre- and posttreatment validated cough assessment tools, the cough severity index (CSI) and Leicester Cough Questionnaire (LCQ). All subjects in the study reported at least some improvement in their cough symptoms. In a Wilcoxon signed rank test that compared paired results, CSI scores improved from 23 to 14 and LCQ scores improved from 74 to 103 ( P = .003 and P = .005, respectively). This small preliminary assessment suggests that tramadol warrants additional evaluation as a treatment for neurogenic cough.
PMID: 28463539
ISSN: 1097-6817
CID: 2546472

The Effect of Time to Endoscopy on Patient and Procedural Outcomes Among Foreign Body Swallowers: A Prospective Study [Meeting Abstract]

Ali, Rabia; Sartori, Daniel; Chhabra, Natasha; Minhas, Hadi J; Fang, Yixin; Williams, Renee; Goodman, Adam J
ISI:000403087401190
ISSN: 1097-6779
CID: 2611342

Factors Associated With the Use of Postoperative Analgesics in Patients Undergoing Direct Microlaryngoscopy

Taliercio, Salvatore; Sanders, Brian; Achlatis, Stratos; Fang, Yixin; Branski, Ryan; Amin, Milan
OBJECTIVE: Morbidity associated with suspension laryngoscopy has been well documented. However, standard of care with regard to postoperative analgesia has not been described, and anecdotal evidence suggests wide variability with regard to postoperative narcotic and non-narcotic recommendations. We sought to quantify the postoperative course following suspension microlaryngoscopy by relating patient-based and intraoperative measures with analgesic use. METHODS: Body mass index (BMI), Friedman tongue position (FTP), and Mallampati scores as well as laryngoscope type, number of attempts required for optimal visualization, and suspension time were documented in 50 consecutive patients undergoing routine suspension microlaryngoscopy. Postoperative symptoms and analgesic use was queried on postoperative days 1, 3, and 10. RESULTS: In this cohort, 62.5% employed postoperative analgesia. However, only 20% required narcotics. No difference in suspension time was identified in those taking analgesics (33.0 vs 37.3 minutes, P = .44). In addition, no relationship between procedure type and the need for analgesia was noted. The majority of patients (76%) described sore throat persisting for 3 postoperative days; 36% reported sore throat persisting beyond postoperative day 3. CONCLUSIONS: The majority of patients undergoing microlaryngoscopy reported discomfort, but symptoms were largely ameliorated with over-the-counter analgesics. Routine prescription of narcotics following routine suspension laryngoscopy may be unnecessary.
PMID: 28397564
ISSN: 1943-572x
CID: 2527762