A pinboard by
this curator

Student, University of California, Los Angeles


Innovatively visualize data sets and harmonize measures to make data sources mutually compatible.

I have three domains for my research. In statistics and data science, I focus data harmonization, aimed to adjust differences (including measurements, methods, procedures, schedules, specifications, or systems) and make data sets from different sources mutually compatible. In clinical application, I focus on improving medication adherence for HIV patients and innovative techniques to improve the quality of measurements, such as, electronic device system (Chip card in the bottle cap), real-time measuring techniques based on SIM (subscriber identification module in the phone) card, and ingestible sensor techniques (sensor with ingestible with the pills). I work on statistical models to predicting the viral load outcomes in order to better health outcomes. In survey research area, I focus on improving the self-reported outcomes regarding measuring oral health status of children and adolescents. I work on using item response models to select the best set of survey questions. During the field testing, I work with dental professionals to collect the oral health measurements, both survey items and clinical exams, in the dental clinics in Los Angeles County. Besides my own research, one of my daily job is to consult medical doctors and dental residence with their statistical needs to solve the real world problems in clinical settings. These consulting services provide the best motivations for new statistical techniques.


Predictive power of the severity measure of attachment loss for periodontal care need.

Abstract: The prevalence of periodontal diseases is high, and >15% of adults have severe gum disease. Clinical attachment loss (AL) is one of the most important measures for periodontal disease severity. With AL, one could measure the worst scenario, the average, or the cumulative sum of AL among all teeth. The objective of this study is to evaluate which of the 15 measures of periodontal problems (e.g., maximum, mean, and cumulative AL) best predict the need for periodontal treatment.Using detailed periodontal data obtained through clinical examination from the National Health and Nutrition Examination Survey 1999 to 2002, weighted logistic regression was used to model the periodontal treatment need of 15 different periodontal disease measures. The outcome measure is the clinically determined periodontal need.After adjustment for the covariates of age, sex, ethnicity, education, smoking status, and diabetes, the three most predictive measures were identified as: 1) the sum of the maximum mid-buccal (B) and mesio-buccal (MB) measures, which reflects the worst case of both B and MB measures; 2) the sum of the maximum MB measure or the worst case of the MB measure; and 3) the sum of all B and MB measures, or the cumulative AL measures.Cumulative periodontal morbidity, particularly the worst case of B and MB measures, has the strongest impact on the need for periodontal care. All the demographic variables and covariates follow the classic pattern of association with periodontal disease.

Pub.: 24 Nov '12, Pinned: 28 Jun '17

Patient-Reported oral health outcome measurement for children and adolescents.

Abstract: Oral health is an important component of daily functioning and well-being. A comprehensive patient-reported oral health measure is needed to gauge the impact of oral health status on children and adolescents. This study aims to develop oral health item banks and associated short-form surveys for children and adolescents 2-17 year olds.Using children and adolescents, ages 2-17 years, selected from diverse dental sites in Greater Los Angeles Area, we propose to develop state-of-the-science methods to create oral health item banks to effectively measure oral health outcomes for children and adolescents. Methods include a literature review of existing measures, focus groups, cognitive interviews, drafting and field testing of survey items, and evaluation of the psychometric properties of the measures.Based on the systematic literature search and focus groups, we identified core (physical health, mental health, and social function domains) and peripheral (e.g., need and access) oral health domains. We then drafted survey items and revised them based on 66 cognitive interviews (27 children/adolescents and 39 parents) with 39 families. The revised items will be administered in a field test of 500 children and adolescents ages 2-17, and their parents.The qualitative methods used in the initial phases of the project (focus group and cognitive interviews) are the initial steps in the development of oral health item banks and associated short-form surveys for children and adolescents. The oral health items can potentially be used to create effective computerized adaptive test and/or create ad hoc short forms targeting specific areas of oral health to survey large populations of children with much less cost compared with traditional clinical oral health examination.

Pub.: 17 Sep '16, Pinned: 28 Jun '17

Socio-behavioral predictors of self-reported oral health-related quality of life.

Abstract: To examine the relationship between social and financial support, behavioral and sociodemographic variables, and oral health-related quality of life (OHRQoL) in a national probability sample.The National Health and Nutrition Examination Survey (NHANES) 2003-2004 data system was used; there were 12,761 persons selected for the sample, 10,122 of those were interviewed (79.3 %). Oral health-related quality of life, the outcome measure, was evaluated using seven items derived from the 14-item NHANES Oral Health Impact Profile (OHIP) included in the home interview. The aggregated OHRQoL scores ranged from 7 to 28. We included only adults, aged 20 and older, who self-reported their alcohol use during home interview (n = 5,014). Independent variables were social and financial support, and behavioral variables (smoking and alcohol use), with sociodemographic variables as covariates. Multiple linear regression analysis used weighted data representing 124 million persons.Lack of financial support reduced OHRQoL, but not social support. Smoking reduced OHRQoL, but not alcohol use. Compared to ages 20-24, persons aged 24-44 and aged 45-64 had significantly lower OHRQoL scores, but persons aged 65+ did not. Latinos' OHRQoL scores were lower than those of whites; there were no differences between whites and other ethnic groups.The model provides insights into the perception of OHRQoL in that oral health related to the ability to pay for care. Those in the middle years (24-64) rate their OHRQoL lower than do their younger cohorts; there is no difference in OHRQoL between the young and the old.

Pub.: 25 Apr '12, Pinned: 28 Jun '17

Incidence of skin and soft tissue infections in ambulatory and inpatient settings, 2005-2010.

Abstract: The emergence of community-associated methicillin-resistant S. aureus was associated with dramatically increased skin and soft tissue infection (SSTI) incidence in the first few years of the 21(st) century in the U.S. However, subsequent trends are poorly understood.We examined ambulatory and inpatient data of over 48 million persons years aged 0-64 years from the HealthCore Integrated Research Database (HIRD) between 2005 and 2010. Data were extracted from medical, pharmacy, and eligibility databases. We quantified SSTI incidence, type, and complications and comparative incidence trends for urinary tract infections (UTIs) and pneumonia.A total of 2,301,803 SSTIs were identified. Most SSTIs (95 %) were treated in the ambulatory setting and most (60 %) were categorized as abscesses or cellulitis. During the study period, SSTI incidence remained relatively stable from 47.9 (95 % CI: 47.8-48.1) cases/1,000 PY in 2005 to 48.5 cases/1,000 PY (95 % CI: 48.3-48.6) in 2010). Persons aged 45-64 years had the highest incidence of both ambulatory-treated and inpatient-treated SSTIs (51.2 (95 % CI: 51.1-51.3) and 3.87 (95 % CI: 3.84-3.90) cases/1,000 PY, respectively). SSTI complications such as myositis, gangrene, and sepsis occurred in 0.93 % (95 % CI: 0.92-0.94 %) and 16.92 % (95 % CI: 16.87-16.97 %) of ambulatory-treated and inpatient-treated patients, respectively. SSTI incidence was approximately twice that of UTIs and tenfold of that of pneumonia.Among our large, diverse population of persons less than 65 years, SSTI incidence 2005 through 2010 has remained relatively constant at approximately 4.8 SSTIs per 100 person years, suggesting that previously observed increases in SSTI incidence remain sustained.

Pub.: 22 Aug '15, Pinned: 28 Jun '17

Factors Influencing Resident Choice of Prosthodontic Residency Program

Abstract: The decision by prosthodontic residency program directors to employ the Match process highlights the need to understand applicant priorities that influence their choice of which programs to rank highly. The purpose of this study is to determine the factors that were most important to residents when choosing from among nonmilitary based prosthodontics dental residency programs in the United States.Following completion of a pilot study, all currently enrolled prosthodontic residents at nonmilitary residency programs were invited to participate via the internet. The study consisted of a survey instrument asking residents to rank 26 possible factors that might impact an applicant's choice of residency program. In addition, the instrument collected other possible influencing variables including gender and debt load. Mean rank scores were compared to determine the most and least important factors. Kruskal-Wallis test was used to compare specific factors between the possible influencing variables.Two hundred and thirty residents completed the survey instrument, representing a 54.1% response rate of possible participants. With regard to factors influencing program choice, reputation of the residency program was the factor ranked the highest by participants, followed in descending order by the program director's personality, curriculum content, access to use of the latest digital technology, and opportunities for dental implant placement. Quality of schools for children, community outreach opportunities, and the ability to moonlight were ranked as the least important factors. Male and female residents ranked factors such as tuition/stipend, curriculum content, and community outreach opportunities significantly differently. Depending on debt load, residents ranked the factors tuition/stipend, ability to moonlight, curriculum content, and safety of the area where the program is differently.Current prosthodontic residents valued the reputation of the program as the most important factor when applying to residency. Participant gender and debt load influence the factors chosen by residents as more or less important. These data will assist prosthodontic educators position their programs in the best possible light to attract applicants to their programs.

Pub.: 11 Mar '17, Pinned: 28 Jun '17