ZDZheng Dai

Zheng Dai

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Zheng Dai has a total of 53 co-authors. Here are the top co-authors by number of shared publications:

  • Ruchi Bhandari

    School of Public Health West Virginia University Morgantown West Virginia USA

    7 shared publications
  • Gordon S. Smith

    School of Public Health, West Virginia University, Morgantown, West Virginia

    6 shared publications
  • James C. Kraner

    West Virginia Office of the Chief Medical Examiner, West Virginia University, Charleston, West Virginia

    6 shared publications
  • Sijin Wen

    Department of Biostatistics, School of Public Health, West Virginia University, One Medical Center Drive, Morgantown, WV, 26506, USA

    6 shared publications
  • Yilin Cai

    Department of Biostatistics, School of Public Health, West Virginia University, One Medical Center Drive, Morgantown, WV, 26506, USA

    6 shared publications
  • Allen R. Mock

    West Virginia Office of the Chief Medical Examiner, West Virginia University, Charleston, West Virginia

    5 shared publications

Latest Publications

  • Open Access

    Evaluating tramadol utilization and patterns by county-level social determinants of health characteristics from 2015 to 2022

    by Chisom Eze, Marie Abate, Gordon Smith, Zheng Dai, Nathan Wood, Mohammad Al-Mamun

    Background: Tramadol, an opioid analgesic, is liable to abuse and implicated in drug overdose deaths in the U.S. However, tramadol use and its driving factors have not been intricately examined. This study aims to evaluate the tramadol utilization trends by dosage and county, and social determinants of health (SDoH)-related factors associated with the utilization rates. Methods: The retrospective study utilized 2015-2022 WV controlled substances monitoring program, CDC opioid dispensing rate, and County Health Ranking and Roadmaps data. Average tramadol daily dose and morphine milligram equivalent were calculated. Annual tramadol dispensing and use rate per 100 population were calculated for WV and each county. Pooled Poisson regression model was used to analyze the relationship between tramadol dispensing rate, opioid dispensing rate, and SDoH variables. Results: Tramadol dispensing rate declined by 35% (2015-2022), but varied within counties with Grant (34.10), McDowell (28.91), and Wyoming (26.67) average annual rates exceeding the overall WV rate (17.90). High tramadol dispensing rate was associated with a high percentage of the population with poor/fair health ( β=0.07, p=0.01) , physically inactive ( β=0.10, p=0.0003 ), uninsured ( β=0.09, p=0.001) , and elevated primary care provider (PCP) rate ( β=0.10, p=0.0003 ) and opioid dispensing rate ( β=0.14, p ). Conclusion: Our study found heterogenous trends of tramadol dispensing rate within WV and was associated with county-wise health status, physical inactivity, insurance, PCP, and opioid dispensing rates. Considering these factors in local surveillance might improve health, and reduce disease burden, drug, and health resource utilization.

  • Open Access

    Male Predominance in West Virginia Unintentional Overdose Deaths Is Influenced by Alcohol and Co-Intoxicants

    by Zheng Dai, Marie A. Abate, Mohammad A. Al-Mamun, James C. Kraner, Allen R. Mock, Gordon S. Smith

    To examine sex differences in overdose (OD) mortality based upon substances involved. A retrospective database analysis of West Virginia OD decedents (12,666 unintentional OD deaths, 2005-early 2023). Exposures were substances judged to contribute to death. The main outcome measure was determination of male to female death ratios with varying co-intoxicant involvement, particularly related to alcohol and fentanyl. Secondary outcomes included associations of fentanyl concentrations with alcohol concentrations and male sex, including fentanyl (F) and inactive metabolite norfentanyl (N) concentration variability between sexes. Alcohol co-intoxication in OD deaths was associated with higher male:female death ratios, from 2.0 (alcohol absent) to 3.3 (alcohol present). There was a greater increase over time in alcohol involvement in recent deaths involving females compared to males (relative increases of 52% vs. 6%, respectively). Male:female ratios with alcohol and fentanyl co-involvement ranged from 5.9:1 (only two drugs involved) to 2.4:1 (= 5 substances), with females significantly more likely to have multiple substances contributing to death. Overall, males had statistically significantly larger fentanyl (F) to norfentanyl (N) median concentration ratios compared to females (8.8 vs. 6.9, respectively). Multivariable analyses found alcohol presence was associated with a statistically significant 22% reduction in predicted fentanyl concentrations. Male:female ratios in unintentional OD deaths were higher with greater alcohol involvement and lower with fewer co-intoxicants. Fentanyl and norfentanyl concentration differences by sex were observed. It is important to determine possible contributors to sex differences in OD death rates to better target prevention and treatment initiatives.

  • Demographic, co-intoxicants and other characteristics of citalopram-involved overdose deaths

    by José E. Muñoz, Marie A. Abate, Zheng Dai, Gordon S. Smith, James C. Kraner, Allen R. Mock

    Selective serotonin reuptake inhibitors (SSRIs) including citalopram are commonly used antidepressants that can be involved in drug-related deaths along with opioids and other substances. This study characterized citalopram involvement in West Virginia (WV) drug-related deaths compared to other SSRI and non-SSRI-related deaths. All 2005–2021 WV drug-related deaths were analyzed in this retrospective study. Demographics, other substances involved, and comorbidities in cases in which citalopram was listed on the death certificate were compared to other SSRI-related and total non-SSRI deaths. Citalopram concentrations and the association between citalopram presence with predicted fentanyl concentrations were determined. Citalopram was the most common antidepressant present in the deaths (4.5% of 14,363 total), with most (81%) unintentional. Male: female ratios in citalopram cases (0.9:1) were significantly lower than in non-SSRI deaths (2.3:1). Almost two-thirds of citalopram deaths had ≥ 4 substances involved compared to 26% of non-SSRI deaths. Overall, oxycodone was most frequently identified in citalopram deaths (fentanyl more commonly in recent years), followed by alprazolam and diazepam. Cardiovascular comorbidity was significantly more common in citalopram than non-SSRI deaths. No association was found between citalopram presence and predicted fentanyl concentrations. Most citalopram-related deaths were unintentional and involved proportionately more females, with larger numbers of concurrent substances present and more cardiovascular comorbidity compared to non-SSRI deaths. Citalopram is widely used and less toxic than many antidepressants. The extent to which it contributed to overdose deaths can be difficult to ascertain given the multiple substances usually present.


  • Open Access

    Machine Learning-Based Prediction of Binge Drinking among Adults in the United State: Analysis of the 2022 Health Information National Trends Survey

    by Xinya Huang, Zheng Dai, Kesheng Wang, Xingguang Luo

    Little is known about the association of social media and belief in alcohol and cancer with binge drinking. This study aimed to perform feature selection and develop machine learning (ML) tools to predict occurrence of binge drinking among adults in the United State. A total of 5,886 adults including 1,252 who ever experienced with binge drinking were selected from the 2022 Health Information National Trends Survey (HINTS 6). Feature selection of 69 variables was conducted using Boruta and the Least Absolute Shrinkage and Selection Operator (LASSO). Seven machine learning (ML) tools including the Support Vector Machines (SVMs) algorithms, Logistic Regression, Naïve Bayes, Random Forest, K-Nearest Neighbor, Gradient Boosting Machine, and XGBoost were applied to develop ML models to predict binge drinking. The overall prevalence of binge drinking among U.S. adults is 21.3%. Both Boruta and LASSO selected 28 identical variables. SVM with Radial Basis Function revealed the best model with the highest AUC of 0.949 and sensitivity of 0.958. The top risk factors of binge drinking were tobacco use (e-cigarette use and smoking status), belief in alcohol (alcohol decreases the risk of future health), belief in cancer (prevention is not possible, worry about getting cancer), and social media (social media visits and sharing health information). These findings underscore the need for multiple health behavior interventions to enhance education related to alcohol use and cancer and how to effectively employ social media to improve health outcomes.


  • Quantifying a potential protective effect of buprenorphine on fatality risk during acute fentanyl exposures

    by Zheng Dai, Marie A. Abate, Erin Winstanley, James C. Kraner, Eric Lundstrom, Allen R. Mock, Gordon S. Smith


  • Open Access

    Comorbidities of Rural Children and Adolescents with Migraine and without Migraine

    by Suzy Mascaro Walter, Zheng Dai, Kesheng Wang

    (1) Background: Migraine is associated with comorbidities that are common in the general rural pediatric population. The purpose of this study is to evaluate the differences in the occurrence of comorbidities between rural children and adolescents with and without migraine. (2) Methods: A cross-sectional, secondary data analysis using electronic medical records of 1296 patients (53.8% females, aged 12.4 ± 3.2) was completed. Mann–Whitney U test was used to detect the difference in the number of comorbidities between the two groups. Chi-square test was used to identify the differences in the number of comorbidities, which were classified as low (0–1 comorbidities), medium (2–3 comorbidities), and high (4 or plus comorbidities) degree of comorbidities. (3) Results: Significant differences were found between those children and adolescents with migraine vs. those without for depression (p < 0.0001), anxiety (p < 0.0001), and Ehlers–Danlos Syndrome (EDS; p = 0.0309). A marginally significant difference was found between those children and adolescents with migraine (47.2%; n = 306) vs. those without (42.1%; n = 273) for unhealthy weight (p = 0.0652). Approximately 40% of the migraineurs had 2–3 comorbidities, whereas 32% of the non-migraineurs had 2–3 comorbidities (p = 0.0003). (4) Conclusions: Findings demonstrate the importance of identifying comorbidities associated with rural pediatric migraine in order to develop effective treatment strategies that optimize patient outcomes.


  • Open Access

    The association between early onset of alcohol, smokeless tobacco and marijuana use with adult binge drinking in United States

    by Zheng Dai, Kesheng Wang

    Binge drinking is a deadly pattern of excessive alcohol use that is associated with multiple diseases in the United States. To date, little is known about the associations between the early onset of substance use and other factors with the severity of adult binge drinking. The 2018 National Survey on Drug Use and Health data was used to identify binge drinking (binary and in number of days in the past month). Age at onset was categorized into four groups as 1–12, 13–14, 15–17, or beyond 18. Weighted multivariate logistic regression and Poisson regression analyses were performed to examine the associations between early onset of alcohol, smokeless tobacco, and marijuana use with binge drinking. The severity of binge drinking was statistically significantly associated with substance use (4.15 days in a month), early onset of alcohol, smokeless tobacco, and marijuana use (2.15–4.93 days, all p-values < 0.0001), after accounting for the covariates. Past year substance use disorder is strongly associated with binge drinking. The severity of adult binge drinking is significantly associated with early onset of substance use including alcohol, smokeless tobacco, and marijuana. Continued efforts are warranted to improve substance use prevention and treatment tailored for adolescents and youths to prevent development of adult binge drinking.


  • Characterization of Unintentional Deaths Among Buprenorphine Users

    by Zheng Dai, George N. Limen, Marie A. Abate, James C. Kraner, Allen R. Mock, Gordon S. Smith

    Objective: Medications used to treat opioid use disorder (OUD) reduce drug overdose risk. Buprenorphine is often the preferred treatment for OUD because of its high safety profile. Given expanding buprenorphine use, this study sought to examine buprenorphine-involved deaths (BIDs) and compare them with other drug-related deaths. Method: West Virginia drug-related deaths from 2005 to early 2020 were identified. Study data included decedent demographics, toxicology, autopsy findings, and medical and prescription histories. Characteristics of BIDs compared with other drug-related deaths were statistically analyzed. Results: Among 11,764 drug-related deaths, only 564 (4.8%) involved buprenorphine. Buprenorphine alone was present in 32 deaths, of which 20 were considered the direct cause of death (0.2% of all drug-related deaths). Significantly more BIDs involved five or more drugs (23%) compared with other opioid deaths (14.9%). Co-intoxicants found most frequently in BIDs were benzodiazepines (47.3%), methamphetamine (27.1%), and fentanyl (22.9%). Cardiovascular and pulmonary comorbidities were identified in 43% and 21% of BIDs, respectively. Of the 564 BIDs, a current buprenorphine prescription was present in 132 deaths (23.4%). Conclusions: Despite increasing buprenorphine use, BIDs comprised less than 5% of overall West Virginia drug-related deaths. Seldom was it the only drug found, and most decedents did not have current prescriptions for buprenorphine. Although buprenorphine is effective, with a wide safety margin, clinicians and patients should be aware that buprenorphine can be involved in overdose deaths, especially when buprenorphine is taken in combination with drugs such as benzodiazepines, methamphetamine, or fentanyl, and in persons with underlying cardiovascular or pulmonary comorbidities.


  • Characteristics of xylazine‐related deaths in West Virginia—Xylazine‐related deaths

    by Julie Sibbesen, Marie A. Abate, Zheng Dai, Gordon S. Smith, Eric Lundstrom, James C. Kraner, Allen R. Mock

    Background and Objectives: The involvement of xylazine, a veterinary drug, in West Virginia (WV) human drug‐related deaths was examined.Methods: WV drug deaths from 2019 (when xylazine was first identified) to mid‐2021. Characteristics including toxicology findings were compared between xylazine and nonxylazine deaths.Results: Of 3292 drug deaths, 117 involved xylazine, and the proportions of deaths with it have increased (1% [2019] to 5% [mid‐2021)]. Xylazine decedents had more cointoxicants, with fentanyl (98%) predominant followed by methamphetamine. Xylazine decedents had a significantly greater history of drug or alcohol misuse and hepatic disease.Conclusions and Scientific Significance: In one of the largest analyses of xylazine‐involved deaths in a predominantly rural state, identification of xylazine was increasing with multiple cointoxicants (especially fentanyl), and was present in a few deaths with only one other substance involved. Health professionals should be aware of possible enhanced toxicity from xylazine ingestion especially since naloxone does not reverse xylazine's adverse effects.


  • Open Access

    The association between early onset of alcohol, smokeless tobacco and marijuana use with adult binge drinking in United States

    by Zheng Dai, Kesheng Wang

    1) BackgroundBinge drinking remains significant public health problems in United States. To date, less is known how the early onset of substance use and other factors impacted the severity of adult binge drinking. 2) MethodThe 2018 National Survey on Drug Use and Health data was used to identify binge drinking (binary and in number of days in the past month). Age at onset was categorized four groups as 1–12, 13–14, 15–17, or beyond 18. Weighted multivariate logistic regression and Poisson regression analyses were performed to examine the association between early onset of alcohol, smokeless tobacco, and marijuana use with binge drinking. 3) ResultsThe severity of binge drinking was statistically significantly associated with substance use (2.66 days in a month), early (before age 15) onset of alcohol, smokeless tobacco, and marijuana use (3.20–4.93 days, all p-values < 0.0001), after accounting for the covariates. Past year substance use disorder is strongly associated with binge drinking. 4) ConclusionsThe severity of adult binge drinking is significantly associated with early onset of substance use including alcohol, smokeless tobacco, and marijuana. Continued efforts are warranted to improve substance use prevention and treatment tailored for adolescents and youth to prevent development of adult binge drinking.