Zhaohui Cui
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Zhaohui Cui has a total of 45 co-authors. Here are the top co-authors by number of shared publications:
Amanda Conrad
Division of Foodborne, Waterborne and Environmental Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, USA
1 shared publication- Anne M. Hotop
Division of Epidemiology & Community Health, School of Public Health, University of Minnesota, Minneapolis, MN, USA
1 shared publication Beau B. Bruce
Division of Foodborne, Waterborne and Environmental Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, USA
1 shared publication- Bonnie Bruce
Department of Ophthalmology, Emory University School of Medicine, Atlanta, Georgia
1 shared publication - Charlotte A. Pratt
National Heart, Lung, and Blood Institute, National Institutes of Health, 6705 Rockledge Drive, Bethesda, MD 20892, USA
1 shared publication - Christy Bennett
Division of Foodborne, Waterborne, and Environmental Diseases, Centers for Disease Control and Prevention , Atlanta, Georgia , USA
1 shared publication
Latest Publications
- 29 November 2023
- Published by SAGE Publications in Foodborne Pathogens and Disease
- Vol. 20 (12) , 579-586
Predicting Food Sources of Listeria monocytogenes Based on Genomic Profiling Using Random Forest Model
Listeria monocytogenes can cause severe foodborne illness, including miscarriage during pregnancy or death in newborn infants. When outbreaks of L. monocytogenes illness occur, it may be possible to determine the food source of the outbreak. However, most reported L. monocytogenes illnesses do not occur as part of a recognized outbreak and most of the time the food source of sporadic L. monocytogenes illness in people cannot be determined. In the United States, L. monocytogenes isolates from patients, foods, and environments are routinely sequenced and analyzed by whole genome multilocus sequence typing (wgMLST) for outbreak detection by PulseNet, the national molecular surveillance system for foodborne illnesses. We investigated whether machine learning approaches applied to wgMLST allele call data could assist in attribution analysis of food source of L. monocytogenes isolates. We compiled isolates with a known source from five food categories (dairy, fruit, meat, seafood, and vegetable) using the metadata of L. monocytogenes isolates in PulseNet, deduplicated closely genetically related isolates, and developed random forest models to predict the food sources of isolates. Prediction accuracy of the final model varied across the food categories; it was highest for meat (65%), followed by fruit (45%), vegetable (45%), dairy (44%), and seafood (37%); overall accuracy was 49%, compared with the naive prediction accuracy of 28%. Our results show that random forest can be used to capture genetically complex features of high-resolution wgMLST for attribution of isolates to their sources.
- 29 November 2023
- Published by SAGE Publications in Foodborne Pathogens and Disease
- Vol. 20 (12) , 579-586
Predicting Food Sources of Listeria monocytogenes Based on Genomic Profiling Using Random Forest Model
Listeria monocytogenes can cause severe foodborne illness, including miscarriage during pregnancy or death in newborn infants. When outbreaks of L. monocytogenes illness occur, it may be possible to determine the food source of the outbreak. However, most reported L. monocytogenes illnesses do not occur as part of a recognized outbreak and most of the time the food source of sporadic L. monocytogenes illness in people cannot be determined. In the United States, L. monocytogenes isolates from patients, foods, and environments are routinely sequenced and analyzed by whole genome multilocus sequence typing (wgMLST) for outbreak detection by PulseNet, the national molecular surveillance system for foodborne illnesses. We investigated whether machine learning approaches applied to wgMLST allele call data could assist in attribution analysis of food source of L. monocytogenes isolates. We compiled isolates with a known source from five food categories (dairy, fruit, meat, seafood, and vegetable) using the metadata of L. monocytogenes isolates in PulseNet, deduplicated closely genetically related isolates, and developed random forest models to predict the food sources of isolates. Prediction accuracy of the final model varied across the food categories; it was highest for meat (65%), followed by fruit (45%), vegetable (45%), dairy (44%), and seafood (37%); overall accuracy was 49%, compared with the naive prediction accuracy of 28%. Our results show that random forest can be used to capture genetically complex features of high-resolution wgMLST for attribution of isolates to their sources.
- Open Access
- 17 July 2023
- Published by Oxford University Press (OUP) in Open Forum Infectious Diseases
- Vol. 10 (8) , ofad378
Epidemiology and Antimicrobial Resistance of Campylobacter Infections in the United States, 2005–2018
Campylobacter is the most common cause of bacterial diarrhea in the United States; resistance to macrolides and fluoroquinolones limits treatment options. We examined the epidemiology of US Campylobacter infections and changes in resistance over time. The Foodborne Diseases Active Surveillance Network receives information on laboratory-confirmed Campylobacter cases from 10 US sites, and the National Antimicrobial Resistance Monitoring System receives a subset of isolates from these cases for antimicrobial susceptibility testing. We estimated trends in incidence of Campylobacter infection, adjusting for sex, age, and surveillance changes attributable to culture-independent diagnostic tests. We compared percentages of isolates resistant to erythromycin or ciprofloxacin during 2005–2016 with 2017–2018 and used multivariable logistic regression to examine the association of international travel with resistance. Adjusted Campylobacter incidence remained stable or decreased for all groups analyzed since 2012. Among 2449 linked records in 2017–2018, the median patient age was 40.2 years (interquartile range, 21.6–57.8 years), 54.8% of patients were male, 17.2% were hospitalized, and 0.2% died. The percentage of resistant infections increased from 24.5% in 2005–2016 to 29.7% in 2017–2018 for ciprofloxacin (P < .001) and from 2.6% to 3.3% for erythromycin (P = .04). Persons with recent international travel had higher odds than nontravelers of having isolates resistant to ciprofloxacin (adjusted odds ratio [aOR] varied from 1.7 to 10.6 by race/ethnicity) and erythromycin (aOR = 1.7; 95% confidence interval, 1.3–2.1). Campylobacter incidence has remained stable or decreased, whereas resistance to antimicrobials recommended for treatment has increased. Recent international travel increased the risk of resistance.
- 17 July 2023
- Published by Oxford University Press (OUP) in Open Forum Infectious Diseases
- Vol. 10 (8) , ofad378
Open AccessEpidemiology and Antimicrobial Resistance of Campylobacter Infections in the United States, 2005–2018
Campylobacter is the most common cause of bacterial diarrhea in the United States; resistance to macrolides and fluoroquinolones limits treatment options. We examined the epidemiology of US Campylobacter infections and changes in resistance over time. The Foodborne Diseases Active Surveillance Network receives information on laboratory-confirmed Campylobacter cases from 10 US sites, and the National Antimicrobial Resistance Monitoring System receives a subset of isolates from these cases for antimicrobial susceptibility testing. We estimated trends in incidence of Campylobacter infection, adjusting for sex, age, and surveillance changes attributable to culture-independent diagnostic tests. We compared percentages of isolates resistant to erythromycin or ciprofloxacin during 2005–2016 with 2017–2018 and used multivariable logistic regression to examine the association of international travel with resistance. Adjusted Campylobacter incidence remained stable or decreased for all groups analyzed since 2012. Among 2449 linked records in 2017–2018, the median patient age was 40.2 years (interquartile range, 21.6–57.8 years), 54.8% of patients were male, 17.2% were hospitalized, and 0.2% died. The percentage of resistant infections increased from 24.5% in 2005–2016 to 29.7% in 2017–2018 for ciprofloxacin (P < .001) and from 2.6% to 3.3% for erythromycin (P = .04). Persons with recent international travel had higher odds than nontravelers of having isolates resistant to ciprofloxacin (adjusted odds ratio [aOR] varied from 1.7 to 10.6 by race/ethnicity) and erythromycin (aOR = 1.7; 95% confidence interval, 1.3–2.1). Campylobacter incidence has remained stable or decreased, whereas resistance to antimicrobials recommended for treatment has increased. Recent international travel increased the risk of resistance.
- Open Access
- 28 May 2019
- Published by Springer Nature in Trials
- Vol. 20 (1) , 1-10
Recruitment strategies for predominantly low-income, multi-racial/ethnic children and parents to 3-year community-based intervention trials: Childhood Obesity Prevention and Treatment Research (COPTR) Consortium
The recruitment of participants into community-based randomized controlled trials studying childhood obesity is often challenging, especially from low-income racial/ethnical minorities and when long-term participant commitments are required. This paper describes strategies used to recruit and enroll predominately low-income racial/ethnic minority parents and children into the Childhood Obesity Prevention and Treatment Research (COPTR) consortium. The COPTR consortium has run four independent 3-year, multi-level (individual, family, school, clinic, and community) community-based randomized controlled trials. Two were prevention trials in preschool children and the other two were treatment trials in pre-adolescents and adolescent youth. All trials reported monthly participant recruitment numbers using a standardized method over the projected 18–24 months of recruitment. After randomization of participants was completed, recruitment staff and investigators from each trial retrospectively completed a survey of recruitment strategies and their perceived top three recruitment strategies and barriers. Recruitment was completed in 15–21 months across trials, enrolling a total of 1745 parent-child dyads- out of 6314 screened. The number of children screened per randomized child was 4.6 and 3.5 in the two prevention trials, and 3.1 and 2.5 in the two treatment trials. Recruitment strategies reported included: (1) careful planning, (2) working with trusting community partners, (3) hiring recruitment staff who were culturally sensitive, personality appropriate, and willing to work flexible hours, (4) contacting potential participants actively and repeatedly, (5) recruiting at times and locations convenient for participants, (6) providing incentives to participants to complete baseline measures, (7) using a tracking database, (8) evaluating whether participants understand the activities and expectations of the study, and (9) assessing participants’ motivation for participating. Working with community partners, hiring culturally sensitive staff, and contacting potential participants repeatedly were cited by two trials among their top three strategies. The requirement of a 3-year commitment to the trial was cited by two trials to be among the top three recruitment barriers. Comprehensive strategies that include community partnership support, culturally sensitive recruitment staff, and repeated contacts with potential participants can result in successful recruitment of low-income racial/ethnic minority families into obesity prevention and treatment trials. NET-Works trial: ClinicalTrials.gov, NCT01606891 . Registered on 28 May 2012. GROW trial: ClinicalTrials.gov, NCT01316653 . Registered on 16 March 2011. GOALS trial: ClinicalTrials.gov, NCT01642836 . Registered on 17 July 2012. IMPACT trial: ClinicalTrials.gov, NCT01514279 . Registered on 23 January 2012.
- 28 May 2019
- Published by Springer Nature in Trials
- Vol. 20 (1) , 1-10
Open AccessRecruitment strategies for predominantly low-income, multi-racial/ethnic children and parents to 3-year community-based intervention trials: Childhood Obesity Prevention and Treatment Research (COPTR) Consortium
The recruitment of participants into community-based randomized controlled trials studying childhood obesity is often challenging, especially from low-income racial/ethnical minorities and when long-term participant commitments are required. This paper describes strategies used to recruit and enroll predominately low-income racial/ethnic minority parents and children into the Childhood Obesity Prevention and Treatment Research (COPTR) consortium. The COPTR consortium has run four independent 3-year, multi-level (individual, family, school, clinic, and community) community-based randomized controlled trials. Two were prevention trials in preschool children and the other two were treatment trials in pre-adolescents and adolescent youth. All trials reported monthly participant recruitment numbers using a standardized method over the projected 18–24 months of recruitment. After randomization of participants was completed, recruitment staff and investigators from each trial retrospectively completed a survey of recruitment strategies and their perceived top three recruitment strategies and barriers. Recruitment was completed in 15–21 months across trials, enrolling a total of 1745 parent-child dyads- out of 6314 screened. The number of children screened per randomized child was 4.6 and 3.5 in the two prevention trials, and 3.1 and 2.5 in the two treatment trials. Recruitment strategies reported included: (1) careful planning, (2) working with trusting community partners, (3) hiring recruitment staff who were culturally sensitive, personality appropriate, and willing to work flexible hours, (4) contacting potential participants actively and repeatedly, (5) recruiting at times and locations convenient for participants, (6) providing incentives to participants to complete baseline measures, (7) using a tracking database, (8) evaluating whether participants understand the activities and expectations of the study, and (9) assessing participants’ motivation for participating. Working with community partners, hiring culturally sensitive staff, and contacting potential participants repeatedly were cited by two trials among their top three strategies. The requirement of a 3-year commitment to the trial was cited by two trials to be among the top three recruitment barriers. Comprehensive strategies that include community partnership support, culturally sensitive recruitment staff, and repeated contacts with potential participants can result in successful recruitment of low-income racial/ethnic minority families into obesity prevention and treatment trials. NET-Works trial: ClinicalTrials.gov, NCT01606891 . Registered on 28 May 2012. GROW trial: ClinicalTrials.gov, NCT01316653 . Registered on 16 March 2011. GOALS trial: ClinicalTrials.gov, NCT01642836 . Registered on 17 July 2012. IMPACT trial: ClinicalTrials.gov, NCT01514279 . Registered on 23 January 2012.
- 1 April 2019
- Published by SAGE Publications in Metabolic Syndrome and Related Disorders
- Vol. 17 (3) , 143-148
Changes in Cardiovascular Disease Risk Factors with Unintentional Versus Intentional Weight Loss: The Coronary Artery Risk Development in Young Adults Study
Background: It is well known that total and intentional weight loss (IWL) are associated with reductions in cardiovascular disease (CVD) risk factors, however, associations with unintentional weight loss (UWL) have not been studied. Methods: We examined annual changes in blood pressure and lipids associated with UWL versus IWL in 576 participants from the Coronary Artery Risk Development in Young Adults (CARDIA) study who had lost weight over one of two time intervals (i.e., study years 5–7 or 7–10) and reported whether the weight loss was intentional at the end of the interval. Results: After adjusting for age, sex, race, and field center, an annual weight loss smaller than 6% resulted in small improvements in lipids that were similar regardless of intentionality. When annual weight loss was ≥6%, UWL was associated with less reduction in total cholesterol (−0.2 vs. −10.4 mg/dL), low-density lipoprotein cholesterol (−3.2 vs. −9.4 mg/dL), and triglycerides (−0.5 vs. −19.0 mg/dL) compared with IWL. Intentionality was not associated with the size of changes in blood pressure (systolic blood pressure: −2.7 vs. −2.0 mmHg; diastolic blood pressure: −2.2 vs. −1.3 mmHg) and high-density lipoprotein cholesterol (3.2 vs. 2.7 mg/dL). Conclusions: Substantial UWL may be accompanied or preceded by other biological changes in this young adult population that alter changes in CVD risk factors that are associated with IWL.
- 1 April 2019
- Published by SAGE Publications in Metabolic Syndrome and Related Disorders
- Vol. 17 (3) , 143-148
Changes in Cardiovascular Disease Risk Factors with Unintentional Versus Intentional Weight Loss: The Coronary Artery Risk Development in Young Adults Study
Background: It is well known that total and intentional weight loss (IWL) are associated with reductions in cardiovascular disease (CVD) risk factors, however, associations with unintentional weight loss (UWL) have not been studied. Methods: We examined annual changes in blood pressure and lipids associated with UWL versus IWL in 576 participants from the Coronary Artery Risk Development in Young Adults (CARDIA) study who had lost weight over one of two time intervals (i.e., study years 5–7 or 7–10) and reported whether the weight loss was intentional at the end of the interval. Results: After adjusting for age, sex, race, and field center, an annual weight loss smaller than 6% resulted in small improvements in lipids that were similar regardless of intentionality. When annual weight loss was ≥6%, UWL was associated with less reduction in total cholesterol (−0.2 vs. −10.4 mg/dL), low-density lipoprotein cholesterol (−3.2 vs. −9.4 mg/dL), and triglycerides (−0.5 vs. −19.0 mg/dL) compared with IWL. Intentionality was not associated with the size of changes in blood pressure (systolic blood pressure: −2.7 vs. −2.0 mmHg; diastolic blood pressure: −2.2 vs. −1.3 mmHg) and high-density lipoprotein cholesterol (3.2 vs. 2.7 mg/dL). Conclusions: Substantial UWL may be accompanied or preceded by other biological changes in this young adult population that alter changes in CVD risk factors that are associated with IWL.
- Open Access
- 7 January 2019
- Published by Springer Nature in Journal of Translational Medicine
- Vol. 17 (1) , 11
Establishment of clinical diagnosis model of Graves’ disease and Hashimoto’s thyroiditis
A clinical diagnosis model include thyroid functions, thyroid antibodies and radioactive iodine uptake (RAIU) of patients with hyperthyroidism were established and as new evaluation indicators for the differentiation of the Graves’ disease (GD) and Hashimoto’s thyroiditis (HT). Clinical data of patients with newly diagnosed hyperthyroidism including gender, age, thyroid function, thyroid antibodies (FT3, FT4, TSH, TPOAb, TGAb, TRAb), RAIU (2 h, 6 h, 24 h) were collected. A stepwise regression analysis was performed to establish a model based on these variables. Model 1 was subjected to stepwise regression analysis. After screening, the variables that entered the model included FT3, TGAb, TPOAb, TRAb, 2-h RAIU, 24-h RAIU and gender, in which the variables FT3, TGAb, TRAb, 2-h RAIU, 24-h RAIU, and gender were significantly different. Model 2 without RAIU was also subjected to stepwise regression analysis. After screening, the variables that entered the model included FT4, TGAb, TPOAb, TRAb and gender were statistical significant. The larger value of each variable in the two models indicated the higher probability to diagnose GD. The area under the receiver operating characteristic (ROC) curve of model 1 was 0.843 (95% CI 0.779–0.894), and the area under the ROC curve of model 2 was 0.806 (95% CI 0.685–0.824), which showed good differential diagnostic value. GD and HT diagnosis model was established according to the variables including gender, FT3, TGAb, TRAb, the 2-h RAIU, the 24-h RAIU in the model 1, and the variables FT4, TGAb, TPOAb, TRAb and gender in the model 2 that did not include RAIU. These models had high value to differentiate GD and HT for patients with early hyperthyroidism.
- 7 January 2019
- Published by Springer Nature in Journal of Translational Medicine
- Vol. 17 (1) , 11
Open AccessEstablishment of clinical diagnosis model of Graves’ disease and Hashimoto’s thyroiditis
A clinical diagnosis model include thyroid functions, thyroid antibodies and radioactive iodine uptake (RAIU) of patients with hyperthyroidism were established and as new evaluation indicators for the differentiation of the Graves’ disease (GD) and Hashimoto’s thyroiditis (HT). Clinical data of patients with newly diagnosed hyperthyroidism including gender, age, thyroid function, thyroid antibodies (FT3, FT4, TSH, TPOAb, TGAb, TRAb), RAIU (2 h, 6 h, 24 h) were collected. A stepwise regression analysis was performed to establish a model based on these variables. Model 1 was subjected to stepwise regression analysis. After screening, the variables that entered the model included FT3, TGAb, TPOAb, TRAb, 2-h RAIU, 24-h RAIU and gender, in which the variables FT3, TGAb, TRAb, 2-h RAIU, 24-h RAIU, and gender were significantly different. Model 2 without RAIU was also subjected to stepwise regression analysis. After screening, the variables that entered the model included FT4, TGAb, TPOAb, TRAb and gender were statistical significant. The larger value of each variable in the two models indicated the higher probability to diagnose GD. The area under the receiver operating characteristic (ROC) curve of model 1 was 0.843 (95% CI 0.779–0.894), and the area under the ROC curve of model 2 was 0.806 (95% CI 0.685–0.824), which showed good differential diagnostic value. GD and HT diagnosis model was established according to the variables including gender, FT3, TGAb, TRAb, the 2-h RAIU, the 24-h RAIU in the model 1, and the variables FT4, TGAb, TPOAb, TRAb and gender in the model 2 that did not include RAIU. These models had high value to differentiate GD and HT for patients with early hyperthyroidism.