I have a Master in Health Science, focusing on quantitative data analysis, chronic disease prevention, epidemiology, genetics and clinical trial. My working experience as a biostatistician includes academic organizations, hospitals, contract research organizations (CRO) and communities. I have a strong background in medical genetics, molecular biology, and lab experience enriched with courses and practical activities in environmental health sciences and health informatics. I am also actively involved in government-advocated, community-based and cooperation-related activities in health sciences and statistical methods. I am an expert in using statistical software and epidemiological methodology.

Current Role at Stanford

- Implement protocol for quality control
- Create analytic files with detailed documentation
- Implement data analysis through statistical programming
- Present results for investigators by graphs and tables
- Summarize finding orally and in written forms
- Participate in preparations of papers for publication
- Assist with grant preparation

Honors & Awards

  • Member at Delta Omega Public Health Honor Society Chapter Alpha, Johns Hopkins University, Bloomberg School of Public Health (May 2014)
  • Nancy-Fink Scholarship, Johns Hopkins University, Bloomberg School of Public Health (May 2013)

Education & Certifications

  • SAS Programmer, SAS, Statistical Programming (2013)
  • Certificate, Johns Hopkins University, Bloomberg School of Public Health, Public Health Informatics (2014)
  • Certficate, Johns Hopkins University, Bloomberg School of Public Health, Environmental and Occupational Health (2013)
  • BS, Fudan University, Biology (2012)
  • MHS, Johns Hopkins University, Bloomberg School of Public Health, Epidemiology (2014)


All Publications

  • Can Machine Learning Methods Produce Accurate and Easy-to-use Prediction Models of 30-day Complications and Mortality After Knee or Hip Arthroplasty? Clinical orthopaedics and related research Harris, A. H., Kuo, A. C., Weng, Y., Trickey, A. W., Bowe, T., Giori, N. J. 2019


    BACKGROUND: Existing universal and procedure-specific surgical risk prediction models of death and major complications after elective total joint arthroplasty (TJA) have limitations including poor transparency, poor to modest accuracy, and insufficient validation to establish performance across diverse settings. Thus, the need remains for accurate and validated prediction models for use in preoperative management, informed consent, shared decision-making, and risk adjustment for reimbursement.QUESTIONS/PURPOSES: The purpose of this study was to use machine learning methods and large national databases to develop and validate (both internally and externally) parsimonious risk-prediction models for mortality and complications after TJA.METHODS: Preoperative demographic and clinical variables from all 107,792 nonemergent primary THAs and TKAs in the 2013 to 2014 American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP) were evaluated as predictors of 30-day death and major complications. The NSQIP database was chosen for its high-quality data on important outcomes and rich characterization of preoperative demographic and clinical predictors for demographically and geographically diverse patients. Least absolute shrinkage and selection operator (LASSO) regression, a type of machine learning that optimizes accuracy and parsimony, was used for model development. Tenfold validation was used to produce C-statistics, a measure of how well models discriminate patients who experience an outcome from those who do not. External validation, which evaluates the generalizability of the models to new data sources and patient groups, was accomplished using data from the Veterans Affairs Surgical Quality Improvement Program (VASQIP). Models previously developed from VASQIP data were also externally validated using NSQIP data to examine the generalizability of their performance with a different group of patients outside the VASQIP context.RESULTS: The models, developed using LASSO regression with diverse clinical (for example, American Society of Anesthesiologists classification, comorbidities) and demographic (for example, age, gender) inputs, had good accuracy in terms of discriminating the likelihood a patient would experience, within 30 days of arthroplasty, a renal complication (C-statistic, 0.78; 95% confidence interval [CI], 0.76-0.80), death (0.73; 95% CI, 0.70-0.76), or a cardiac complication (0.73; 95% CI, 0.71-0.75) from one who would not. By contrast, the models demonstrated poor accuracy for venous thromboembolism (C-statistic, 0.61; 95% CI, 0.60-0.62) and any complication (C-statistic, 0.64; 95% CI, 0.63-0.65). External validation of the NSQIP- derived models using VASQIP data found them to be robust in terms of predictions about mortality and cardiac complications, but not for predicting renal complications. Models previously developed with VASQIP data had poor accuracy when externally validated with NSQIP data, suggesting they should not be used outside the context of the Veterans Health Administration.CONCLUSIONS: Moderately accurate predictive models of 30-day mortality and cardiac complications after elective primary TJA were developed as well as internally and externally validated. To our knowledge, these are the most accurate and rigorously validated TJA-specific prediction models currently available ( Methods to improve these models, including the addition of nonstandard inputs such as natural language processing of preoperative clinical progress notes or radiographs, should be pursued as should the development and validation of models to predict longer term improvements in pain and function.LEVEL OF EVIDENCE: Level III, diagnostic study.

    View details for PubMedID 30624314

  • Utility of Diffusion Tensor Imaging Tractography in Evaluating Motor Examination and Functional Outcomes in Patients with Surgically Resected Deep Intracranial Cavernous Malformations: A Preliminary Model Journal of Neurological Surgery Kumar, A., Nielsen, T., Weng, Y., Han, S., Iv, M. S., Steinberg, G. K. 2019

    View details for DOI 10.1055/s-0039-1679505

  • Trajectory analysis for postoperative pain using electronic health records: A nonparametric method with robust linear regression and K-medians cluster analysis Health Informatics Journal Weng, Y. 2019
  • NOS3 895G>T and CBR3 730G>A Are Associated with Recurrence Risk in Non-Muscle-Invasive Bladder Cancer with Intravesical Instillations of THP. Chemotherapy Zhou, Q., Ding, W., Weng, Y., Ding, G., Xia, G., Xu, J., Xu, K., Ding, Q. 2018; 63 (4): 191?97


    To analyze the correlation between pharmacogenomic biomarkers and the efficacy of pirarubicin (THP, also named 4'-O-tetrahydropyranyl-adriamycin) and to explore potential associations of individual genetic backgrounds with the clinical outcomes of non-muscle-invasive bladder cancer (NMIBC) patients.Between July 2003 and June 2011, a total of 91 patients were treated with transurethral resection (TUR) of the bladder tumor and were histopathologically confirmed to have NMIBC. Patients received an immediate instillation and maintenance therapy with THP. All patients underwent follow-up for recurrence. We genotyped 13 single nucleotide polymorphisms (SNPs) from blood and saliva DNA samples of all patients.The associations of patients' genotypes with tumor recurrence risks were analyzed by survival analysis. A total of 16 (17.6%) of the 91 patients with NMIBC had tumor recurrences with a median follow-up of 17 months (range, 2-83 months). We confirmed the effect of the European Organization for Research and Treatment of Cancer (EORTC) risk score for predicting tumor recurrence (p = 0.002, log-rank test). We adjusted for the EORTC score and found that 2 SNPs, NOS3 895G>T (rs1799983) (p = 0.02, HR = 4.32, 95% CI, 1.30-14.39, GT+TT vs. GG) and CBR3 730G>A (rs1056892) (p = 0.04, HR = 2.57, 95% CI, 1.07-6.18, GA+AA vs. GG), were significantly associated with a higher recurrence risk after TUR and instillations of THP in NMIBC patients.Our results suggest that NOS3 895G>T and CBR3 730G>A are genetic markers that can be used to predict tumor recurrence in NMIBC patients receiving intravesical instillations of THP. The effects of those 2 SNPs are independent of the EORTC scores. Further studies with larger sample sizes and longer follow-ups are needed to confirm our results.

    View details for DOI 10.1159/000489402

    View details for PubMedID 30125887

  • Effect of Medicare's Nonpayment Policy on Surgical Site Infections Following Orthopedic Procedures. Infection control and hospital epidemiology Kwong, J. Z., Weng, Y., Finnegan, M., Schaffer, R., Remington, A., Curtin, C., McDonald, K. M., Bhattacharya, J., Hernandez-Boussard, T. 2017: 1-6


    OBJECTIVE Orthopedic procedures are an important focus in efforts to reduce surgical site infections (SSIs). In 2008, the Centers for Medicare and Medicaid (CMS) stopped reimbursements for additional charges associated with serious hospital-acquired conditions, including SSI following certain orthopedic procedures. We aimed to evaluate the CMS policy's effect on rates of targeted orthopedic SSIs among the Medicare population. DESIGN We examined SSI rates following orthopedic procedures among the Medicare population before and after policy implementation compared to a similarly aged control group. Using the Nationwide Inpatient Sample database for 2000-2013, we estimated rate ratios (RRs) of orthopedic SSIs among Medicare and non-Medicare patients using a difference-in-differences approach. RESULTS Following policy implementation, SSIs significantly decreased among both the Medicare and non-Medicare populations (RR, 0.7; 95% confidence interval [CI], 0.6-0.8) and RR, 0.8l; 95% CI, 0.7-0.9), respectively. However, the estimated decrease among the Medicare population was not significantly greater than the decrease among the control population (RR, 0.9; 95% CI, 0.8-1.1). CONCLUSIONS While SSI rates decreased significantly following the implementation of the CMS nonpayment policy, this trend was not associated with policy intervention but rather larger secular trends that likely contributed to decreasing SSI rates over time. Infect Control Hosp Epidemiol 2017;1-6.

    View details for DOI 10.1017/ice.2017.86

    View details for PubMedID 28487001

  • Improving Pain Management and Long-Term Outcomes Following High-Energy Orthopaedic Trauma (Pain Study) JOURNAL OF ORTHOPAEDIC TRAUMA Castillo, R. C., Raja, S. N., Frey, K. P., Vallier, H. A., Tornetta, P., Jaeblon, T., Goff, B. J., Gottschalk, A., Scharfstein, D. O., O'Toole, R. V., METRC 2017; 31: S71?S77


    Poor pain control after orthopaedic trauma is a predictor of physical disability and numerous negative long-term outcomes. Despite increased awareness of the negative consequences of poorly controlled pain, analgesic therapy among hospitalized patients after orthopaedic trauma remains inconsistent and often inadequate. The Pain study is a 3 armed, prospective, double-blind, multicenter randomized trial designed to evaluate the effect of standard pain management versus standard pain management plus perioperative nonsteroidal anti-inflammatory drugs or pregabalin in patients of ages 18-85 with extremity fractures. The primary outcomes are chronic pain, opioid utilization during the 48 hours after definitive fixation and surgery for nonunion in the year after fixation. Secondary outcomes include preoperative and postoperative pain intensity, adverse events and complications, physical function, depression, and post-traumatic stress disorder. One year treatment costs are also compared between the groups.

    View details for DOI 10.1097/BOT.0000000000000793

    View details for Web of Science ID 000402085500014

    View details for PubMedID 28323806

  • National Readmission Patterns of Isolated Splenic Injuries Based on Initial Management Strategy. JAMA surgery Rosenberg, G. M., Knowlton, L., Rajasingh, C., Weng, Y., Maggio, P. M., Spain, D. A., Staudenmayer, K. L. 2017; 152 (12): 1119?25


    Options for managing splenic injuries have evolved with a focus on nonoperative management. Long-term outcomes, such as readmissions and delayed splenectomy rate, are not well understood.To describe the natural history of isolated splenic injuries in the United States and determine whether patterns of readmission were influenced by management strategy.The Healthcare Cost and Utilization Project's Nationwide Readmission Database is an all-payer, all-ages, longitudinal administrative database that provides data on more than 35 million weighted US discharges yearly. The database was used to identify patients with isolated splenic injuries and the procedures that they received. Adult patients with isolated splenic injuries admitted from January 1 through June 30, 2013, and from January 1 through June 30, 2014, were included. Those who died during the index hospitalization or who had an additional nonsplenic injury with an Abbreviated Injury Score of 2 or greater were excluded. Univariate and mixed-effects logistic regression analysis controlling for center effect were used. Weighted numbers are reported.Initial management strategy at the time of index hospitalization, including nonprocedural management, angioembolization, and splenectomy.All-cause 6-month readmission rate. Secondary outcome was delayed splenectomy rate.A weighted sample of 3792 patients (2146 men [56.6%] and 1646 women [43.4%]; mean [SE] age, 48.5 [0.7] years) with 5155 admission events was included. During the index hospitalization, 825 (21.8%) underwent splenectomy, 293 (7.7%) underwent angioembolization, and 2673 (70.5%) had no procedure. The overall readmission rate was 21.1% (799 patients). Readmission rates did not differ based on initial management strategy (195 patients undergoing splenectomy [23.6%], 70 undergoing angioembolism [23.9%], and 534 undergoing no procedure [20%]; P?=?.33). Splenectomy was performed in 36 of 799 readmitted patients (4.5%) who did not have a splenectomy at their index hospitalization, leading to an overall delayed splenectomy rate of 1.2% (36 of 2967 patients). In mixed-effects logistic regression analysis controlling for patient, injury, clinical, and hospital characteristics, the choice of splenectomy (odds ratio, 0.93; 95% CI, 0.66-1.31) vs angioembolization (odds ratio, 1.19; 95% CI, 0.72-1.97) as initial management strategy was not associated with readmission.This national evaluation of the natural history of isolated splenic injuries from index admission through 6 months found that approximately 1 in 5 patients are readmitted within 6 months of discharge after an isolated splenic injury. However, the chance of readmission for splenectomy after initial nonoperative management was 1.2%. This finding suggests that the current management strategies used for isolated splenic injuries in the United States are well matched to patient need.

    View details for PubMedID 28768329

  • A Dietary Intervention in Urban African Americans Results of the "Five Plus Nuts and Beans" Randomized Trial AMERICAN JOURNAL OF PREVENTIVE MEDICINE Miller, E. R., Cooper, L. A., Carson, K. A., Wang, N., Appel, L. J., Gayles, D., Charleston, J., White, K., You, N., Weng, Y., Martin-Daniels, M., Bates-Hopkins, B., Robb, I., Franz, W. K., Brown, E. L., Halbert, J. P., Albert, M. C., Dalcin, A. T., Yeh, H. 2016; 50 (1): 87-95


    Unhealthy diets, often low in potassium, likely contribute to racial disparities in blood pressure. We tested the effectiveness of providing weekly dietary advice, assistance with selection of higher potassium grocery items, and a $30 per week food allowance on blood pressure and other outcomes in African American adults with hypertension.We conducted an 8-week RCT with two parallel arms between May 2012 and November 2013.We randomized 123 African Americans with controlled hypertension from an urban primary care clinic in Baltimore, Maryland, and implemented the trial in partnership with a community supermarket and the Baltimore City Health Department. Mean (SD) age was 58.6 (9.5) years; 71% were female; blood pressure was 131.3 (14.7)/77.2 (10.5) mmHg; BMI was 34.5 (8.2); and 28% had diabetes.Participants randomized to the active intervention group (Dietary Approaches to Stop Hypertension [DASH]-Plus) received coach-directed dietary advice and assistance with weekly online ordering and purchasing of high-potassium foods ($30/week) delivered by a community supermarket to a neighborhood library. Participants in the control group received a printed DASH diet brochure along with a debit account of equivalent value to that of the DASH-Plus group.The primary outcome was blood pressure change. Analyses were conducted in January to October 2014.Compared with the control group, the DASH-Plus group increased self-reported consumption of fruits and vegetables (mean=1.4, 95% CI=0.7, 2.1 servings/day); estimated intake of potassium (mean=0.4, 95% CI=0.1, 0.7 grams/day); and urine potassium excretion (mean=19%, 95% CI=1%, 38%). There was no significant effect on blood pressure.A program providing dietary advice, assistance with grocery ordering, and $30/week of high-potassium foods in African American patients with controlled hypertension in a community-based clinic did not reduce BP. However, the intervention increased consumption of fruits, vegetables, and urinary excretion of potassium.

    View details for DOI 10.1016/j.amepre.2015.06.010

    View details for Web of Science ID 000366845300016

    View details for PubMedID 26321012

Footer Links:

Stanford Medicine Resources: