Peihua Qiu
University of Florida, Biostatistics, Faculty Member
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Machine learning methods have been widely used in different applications, including process control and monitoring. For handling statistical process control (SPC) problems, conventional supervised ...
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Background: Body mass index (BMI)-adjusted prostate-specific antigen (PSA) model has been proposed to improve the predictive accuracy of serum PSA in prostate cancer (PCa) screening. However, how BMI change rate may influence PSA levels... more
Background: Body mass index (BMI)-adjusted prostate-specific antigen (PSA) model has been proposed to improve the predictive accuracy of serum PSA in prostate cancer (PCa) screening. However, how BMI change rate may influence PSA levels in PCa-free men has not been well studied. The current study is to examine the relationship between BMI change rate and serum PSA in PCa-free men and whether this relationship is modified by circulating testosterone. Methods: We conducted this study at a tertiary hospital in the Southeastern US using the Electronic Medical Records of PCa-free men with initial PSA less than 4 ng/mL (cutoff for prostate biopsy), at least 1 testosterone measurement and at least 2 BMI measurements during the study period. Time when the first BMI measurement was recorded served as the baseline, and the study period was defined from baseline to the most recent hospital visit. The included medical records ranged from Jun 2001 to Oct 2015. BMI change rate was created in two ways depending on the number of data points. For men with only 2 BMI measurements, it was calculated by firstly subtracting baseline BMI from the second BMI, then dividing the difference by time interval (months) between the two BMI measurements. For men with more than 2 BMI measurements, we firstly regressed BMI to time interval (months) between that measurement and baseline, then took the β regression coefficient (slope) as the BMI change rate for that men. Multivariable linear regression was used to assess the association of BMI change rate with three PSA measures, including peak, the most recent, and mean PSA during the study period. Effect modification by testosterone was assessed through stratified analysis by testosterone level of 280 ng/dL as cutoff. Results: A total of 470 men with a mean study period of 97.6 months were included. Median age at baseline was 62 years. After adjusting for covariates including baseline BMI, no significant association of BMI change rate was observed with peak PSA (β =0.416, P =0.078), the most recent PSA (β =0.360, P =0.139), or mean PSA (β =0.405, P =0.064) in the overall sample. However, testosterone-stratified analyses indicated that BMI change rate was positively associated with peak PSA (β =1.118, P =0.013), the most recent PSA (β =0.932, P =0.044), and mean PSA (β =1.034, P =0.013) in men with testosterone <280 ng/dL, but no significant association was observed in men with testosterone ≥280 ng/dL (for peak PSA, β =0.076, P =0.785; for the most recent PSA, β =0.072, P =0.802; for mean PSA, β =0.099, P =0.700). Conclusion: Accelerated BMI increase in middle-to-late adulthood might correlate with higher PSA level if a low circulating testosterone occurred concurrently. Further studies are needed to confirm this finding. Citation Format: Kai Wang, Mattia Prosperi, Peihua Qiu, Ting-Yuan David Cheng, Victoria Y. Bird, Xinguang Chen, Mingyang Song. Circulating testosterone in modifying the association of BMI change rate with serum PSA in prostate cancer-free men with initial-PSA less than 4 ng/mL [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 591.
Research Interests:
In medical studies, composite indices and/or scores are routinely used for predicting medical conditions of patients. These indices are usually developed from observed data of certain disease risk factors, and it has been demonstrated in... more
In medical studies, composite indices and/or scores are routinely used for predicting medical conditions of patients. These indices are usually developed from observed data of certain disease risk factors, and it has been demonstrated in the literature that single index models can provide a powerful tool for this purpose. In practice, the observed data of disease risk factors are often longitudinal in the sense that they are collected at multiple time points for individual patients, and there are often multiple aspects of a patient's medical condition that are of our concern. However, most existing single‐index models are developed for cases with independent data and a single response variable, which are inappropriate for the problem just described in which within‐subject observations are usually correlated and there are multiple mutually correlated response variables involved. This paper aims to fill this methodological gap by developing a single index model for analyzing longi...
Research Interests:
Disease outbreaks need to be detected in a timely manner for effective disease control. For disease surveillance, conventional statistical process control charts are often included in public health surveillance systems, without taking... more
Disease outbreaks need to be detected in a timely manner for effective disease control. For disease surveillance, conventional statistical process control charts are often included in public health surveillance systems, without taking into account the complicated structure of the disease incidence data and/or additional covariate information. This chapter presents a novel prospective disease surveillance system, named BCEWMA (Biosurveillance via Covariate-Assisted Exponentially Weighted Moving Average Control Chart), which can accommodate seasonality and arbitrary distribution of disease incidence data. Methodologically, BCEWMA is based on the widely used exponentially weighted moving average control chart, incorporating useful information in covariates. This new surveillance system is applied to two real disease incidence datasets: one regarding the hand, foot and mouth disease in Sichuan province of China and the other about the influenza-like-illness in Florida. These real-data e...
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
BackgroundIn contrast to dual-energy x-ray absorptiometry (DXA), the D3-creatine (D3Cr) dilution method provides a direct measure of skeletal muscle mass and in a cohort of older men has been strongly associated with health-related... more
BackgroundIn contrast to dual-energy x-ray absorptiometry (DXA), the D3-creatine (D3Cr) dilution method provides a direct measure of skeletal muscle mass and in a cohort of older men has been strongly associated with health-related outcomes. However, sensitivity to detect changes in D3Cr-derived muscle mass due to an intervention is limited.MethodsTwenty-one older adults (≥70 years) with low-to-moderate physical function were randomized to a 15-week high-intensity strength training (ST) or a health education (HE) group. Full-body progressive intensity ST was performed 3 days per week.ResultsThe mean age was 82.1 years, with 64% females. After 15 weeks, both D3Cr muscle mass (MM; 2.29 kg; 95% CI: 0.22, 4.36) and DXA appendicular lean mass (ALM; 1.04 kg; 95% CI: 0.31, 1.77) were greater in ST group compared to HE. Baseline correlations between D3Cr MM and DXA ALM (r = 0.79; 95% CI: 0.53, 0.92) or total lean body mass (LBM; r = 0.79; 95% CI: 0.52, 0.91) were high. However, longitudinal changes in D3Cr MM were weakly correlated with changes in DXA ALM (r = 0.19; 95% CI: −0.35, 0.64) and LBM (r = 0.40; 95% CI: −0.13, 0.76). More participants showed positive response rates, defined as a >5% increase from baseline, with D3Cr MM (80%) than DXA measures (14%–43%).ConclusionsA progressive ST intervention in low-functioning older adults increased D3Cr MM and DXA ALM. These data suggest that the D3Cr dilution is potentially sensitive to detect changes in muscle mass in response to resistance exercise training. These results are preliminary and could be used for planning larger trials to replicate these results.
Research Interests:
Research Interests:
BackgroundThis study evaluated the association between ratings of perceived exertion (RPE) of walking and major mobility disability (MMD), as well as their transitions in response to a physical activity (PA) compared to a health education... more
BackgroundThis study evaluated the association between ratings of perceived exertion (RPE) of walking and major mobility disability (MMD), as well as their transitions in response to a physical activity (PA) compared to a health education (HE) program.MethodsOlder adults (n = 1633) who were at risk for mobility impairment were randomized to structured PA or HE programs. During a 400 m walk, participants rated exertion as “light” or “hard.” An MMD event was defined as the inability to walk 400 m. MMD events and RPE values were assessed every 6 months for an average of 2.6 years.ResultsParticipants rating their exertion as “hard” had a nearly threefold higher risk of MMD compared with those rating their exertion as “light” (HR: 2.61, 95% CI: 2.19–3.11). The association was held after adjusting for disease conditions, depression, cognitive function, and walking speed (HR: 2.24, 95% CI: 1.87–2.69). The PA group was 25% more likely to transition from “light” to “hard” RPE than the HE group (HR: 1.25, 95% CI: 1.05–1.49). Additionally, the PA group was 27% (HR: 0.73, 95% CI: 0.55 – 0.97) less likely to transition from a “hard” RPE to inability to walk 400 m and was more likely to recover their ability to walk 400 m by transitioning to a “hard” RPE (HR: 2.10, 95% CI: 1.39–3.17) than the HE group.ConclusionsOlder adults rating “hard” effort during a standardized walk test were at increased risk of subsequent MMD. A structured PA program enabled walking recovery, but was more likely to increase transition from “light” to “hard” effort, which may reflect the greater capacity to perform the test.
Research Interests:
We consider the problem of detecting jump location curves of regression surfaces. In the literature, most existing methods detect jumps in regression surfaces based on estimation of either the first-order or the second-order derivatives... more
We consider the problem of detecting jump location curves of regression surfaces. In the literature, most existing methods detect jumps in regression surfaces based on estimation of either the first-order or the second-order derivatives of the regression surface. Methods based on the first-order derivatives are usually better in removing the noise effect, whereas methods based on the second-order derivatives are often superior in localization of the detected jumps. In this thesis research, we suggest a new procedure for jump detection in regression surfaces, which combines the major strengths of the above two types of methods. Our method is based on estimation of both the first-order and second-order derivatives of the true regression surface. Theoretical justifications and numerical studies show that it works well in applications. Jump detection in regression surfaces has many applications. One important application is image segmentation for analyzing gene microarray images. Gene microarray data are widely used in applications, including pharmaceutical and clinical research. By comparing gene expressions in normal and abnormal cells, microarrays can be used for identifying genes involved in particular diseases, and then these genes can be targeted by therapeutic drugs. Many gene expression data are produced from spotted microarray images. A microarray image consists of thousands of spots, with individual DNA sequences first printed at each spot and then equal amount of probe samples from treatment and control cells mixed and hybridized with the printed DNA sequences. To obtain gene expression data, the image needs to be segmented first to separate foregrounds from backgrounds for individual spots, and then averages of foreground pixels are used for computing the gene expression data. So image segmentation of microarray images is related directly to the reliability of gene expression data. Several image segmentation procedures have been suggested and included in some software packages handling gene microarray data. In this thesis research, a new image segmentation methodology is proposed based on local polynomial kernel smoothing. Theoretical arguments and numerical studies show that it has some good statistical properties and would perform well in applications.
Research Interests:
Research Interests:
Research Interests: Neuroscience, Mathematics, Computer Science, Artificial Intelligence, Statistics, and 15 moreMental Health, Nonparametric Statistics, Space Science, Medicine, Denoising, Image Reconstruction, Deconvolution, Image Restoration, Regression, Cluster Analysis, Edges, Smoothing, Deblurring, Point Spread Function, and technometrics
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Vitamin D insufficiency contributes to muscle weakness and a higher risk of falls in older adults. This study explored the impact of vitamin D supplementation on self-reported falls and physical function in older adults with low vitamin D... more
Vitamin D insufficiency contributes to muscle weakness and a higher risk of falls in older adults. This study explored the impact of vitamin D supplementation on self-reported falls and physical function in older adults with low vitamin D levels and a recent fall history. Twenty-five older adults ≥ 70 years with two or more falls during the past year, low vitamin D blood levels (≥10 ng/ml and < 30 ng/mL), and slow gait speed (1.2 m/s) participated in a 6-month vitamin D supplementation (800 IU/day) study. A modified version of the Morse Fall Scale questionnaire was used to assess frequency of falls over one-year prior to study enrollment. Functional outcomes (short physical performance battery, handgrip strength, gait Timed Up and Go, and six-minute walk), and vitamin D levels were assessed at baseline and 6-month follow-up. Based on diaries and pill counts, participants were generally adherent to the intervention (6 of 7 days per week). Supplementation with 800 IU/day of vitamin...
Research Interests:
In general, the change point problem considers inference of a change in distribution for a set of time-ordered observations. This has applications in a large variety of fields, and can also apply to survival data. In survival analysis,... more
In general, the change point problem considers inference of a change in distribution for a set of time-ordered observations. This has applications in a large variety of fields, and can also apply to survival data. In survival analysis, most existing methods compare two treatment groups for the entirety of the study period. Some treatments may take a length of time to show effects in subjects. This has been called the time-lag effect in the literature, and in cases where time-lag effect is considerable, such methods may not be appropriate to detect significant differences between two groups. In this paper, we propose a novel non-parametric approach for estimating the point of treatment time-lag effect by using an empirical divergence measure. Theoretical properties of the estimator are studied. The results from the simulated data and the applications to real data examples support our proposed method.
Research Interests:
Rapid advance of sensor technology is facilitating the collection of high-dimensional data streams (HDS). Apart from real-time detection of potential out-of-control (OC) patterns, post-signal fault diagnosis of HDS is becoming... more
Rapid advance of sensor technology is facilitating the collection of high-dimensional data streams (HDS). Apart from real-time detection of potential out-of-control (OC) patterns, post-signal fault diagnosis of HDS is becoming increasingly important in the filed of statistical process control to isolate abnormal data streams. The major limitations of the existing methods on that topic include i) they cannot achieve reliable diagnostic results in the sense that their performance is highly variable, and ii) the informative correlation among different streams is often neglected by them. This paper elaborates the problem of reliable fault diagnosis for monitoring correlated HDS using the large-scale multiple testing. Under the framework of hidden Markov model dependence, new diagnostic procedures are proposed, which can control the missed discovery exceedance (MDX) at a desired level. Extensive numerical studies along with some theoretical results show that the proposed procedures can control MDX properly, leading to diagnostics with high reliability and efficiency. Also, their diagnostic performance can be improved significantly by exploiting the dependence among different data streams, which is especially appealing in practice for identifying clustered OC streams.
Research Interests:
Background: Disparities exist among patients with pancreatic ductal adenocarcinoma (PDAC). Non-White race is regarded as a negative predictor of expected treatment and overall survival. Data suggest Academic Research Programs (ARP)... more
Background: Disparities exist among patients with pancreatic ductal adenocarcinoma (PDAC). Non-White race is regarded as a negative predictor of expected treatment and overall survival. Data suggest Academic Research Programs (ARP) provide better outcomes for minorities, but ethnic/minority outcomes are underreported. We sought to evaluate outcomes among diverse patients with PDAC, with a focus on Hispanic subgroups. We hypothesize that outcomes among racially/ethnically diverse PDAC patients may be influenced by treatment facility. Methods: The National Cancer Database was used to identify 170,327 patients diagnosed with PDAC between 2004 and 2015. Cox proportional-hazard regression was used to compare survival between race/ethnic groups across facilities. Results: In unadjusted models, compared to Non-Hispanic Whites (NHW), Non-Hispanic Blacks (NHB) had the worst overall survival (HR=1.05, 95%CI:1.03-1.06, p<0.001) and Hispanics had the best overall survival (HR=0.92, 95%CI:0.9...
Research Interests:
Ratings of perceived exertion (RPE) during exercise are linked to several physiological indices and are often elevated in older adults. This study evaluated the association between RPE of walking and incident major mobility disability... more
Ratings of perceived exertion (RPE) during exercise are linked to several physiological indices and are often elevated in older adults. This study evaluated the association between RPE of walking and incident major mobility disability (MMD) as well as response to a physical activity (PA) program. Older adults (n=1633) at-risk for mobility impairment were randomized to a structured PA or health education (HE) program. During a 400m walk, participants rated exertion as “none”, “light”, “somewhat hard” or “hard”. An MMD event was defined as the inability to complete the 400m walk. Transitions between RPE states and an MMD event—when RPE was not collected— were assessed over the follow-up (every 6 months for an average of 2.6 years). Participants rating their exertion as “hard” at baseline 400m walk had nearly 3-fold higher risk of MMD compared with those rating as “light” (HR: 2.61, 95%CI: 2.19-3.11). During follow-up, the PA group was 25% more likely to transition from “light” to “har...
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Applying control charts successfully to batch processes requires the recognition of some distinguishing features of batch processing versus single-unit processing. Traditional Shewhart-type charts can be modified for use in... more
Applying control charts successfully to batch processes requires the recognition of some distinguishing features of batch processing versus single-unit processing. Traditional Shewhart-type charts can be modified for use in batch-processing practitioners of statistical process control (SPC); however, entirely different control chart methods and practices will often be required. This article reviews the common features of batch processing and the data available to control chart practitioners from typical batch processes and suggests alternatives to standard Shewhart charts for effective process monitoring and control. Keywords: batch processes; control charts; SPC ; components of variance; Hotelling's T2 statistic; multivariate SPC; principal components; multivariate projection methods
Research Interests:
ABSTRACT The purpose of this study was to identify reliable predictors of the onset of juvenile myopia. The data from 554 children enrolled in the Orinda Longitudinal Study of Myopia (OLSM) as nonmyopes with baseline data from the third... more
ABSTRACT The purpose of this study was to identify reliable predictors of the onset of juvenile myopia. The data from 554 children enrolled in the Orinda Longitudinal Study of Myopia (OLSM) as nonmyopes with baseline data from the third grade were evaluated to develop a predictive profile for later onset of juvenile myopia. Myopia was defined as at least -0.75 D of myopia in the vertical and horizontal meridians of the right eye as measured by cycloplegic autorefraction (n = 45 children). Chosen predictors were refractive error and the ocular components: corneal power, Gullstrand crystalline lens power, and axial length. Sensitivity and specificity were calculated. Receiver operating characteristic (ROC) curves were generated to evaluate and compare these predictors singly and combined. Refractive error, axial length, Gullstrand lens and pod corneal power were all significant predictive factors for the onset of juvenile myopia. The best single predictor of future myopia onset in the right eye was the right eye&#39;s cycloplegic autorefraction spherical refractive error value (mean sphere across 10 readings) at baseline. For a cut point of less than +0.75 D hyperopia in the third grade, sensitivity was 86.7% and specificity was 73.3%. The area under the ROC curve for this mean sphere was 0.880. Producing a logistic model combining mean sphere, corneal power, Gullstrand lens power, and axial length results in a slight improvement in predictive ability (area under the ROC curve = 0.893). Onset of juvenile myopia can be predicted with moderate accuracy using the mean cycloplegic, spherical refractive error in the third grade. Measurement of other ocular components at this age improves predictive ability, albeit incrementally. Further improvements in the prediction of myopia onset will require the use of longitudinal data in addition to one-time measurement of refractive error and the ocular components.
Research Interests:
Image registration is used widely in applications for mapping one image to another. It is a fundamental task in many imaging applications. Existing image registration methods are either feature-based or intensity-based. Feature-based... more
Image registration is used widely in applications for mapping one image to another. It is a fundamental task in many imaging applications. Existing image registration methods are either feature-based or intensity-based. Feature-based methods first extract relevant image features, and then find a geometrical transformation that best matches the two corresponding sets of features extracted from the two images. The geometrical transformation is estimated directly from the observed image intensities of the two images by an intensity-based image registration method. Most existing methods of both types assume that the mapping transformation has a parametric form or satisfies certain regularity conditions (e.g., it is a smooth function with continuous first or higher order derivatives). They often estimate the mapping transformation globally by solving a global minimization/maximization problem. Such global smoothing methods usually cannot uncover the ill-posed nature of the image registra...
Research Interests:
Some control charts based on machine learning approaches have been developed recently in the statistical process control (SPC) literature. These charts are usually designed for monitoring processes with independent observations at... more
Some control charts based on machine learning approaches have been developed recently in the statistical process control (SPC) literature. These charts are usually designed for monitoring processes with independent observations at different observation times. In practice, however, serial data correlation almost always exists in the observed data of a temporal process. It has been well demonstrated in the SPC literature that control charts designed for monitoring independent data would not be reliable to use in applications with serially correlated data. In this chapter, we suggest using certain existing machine learning control charts together with a recursive data de-correlation procedure. It is shown that the performance of these charts can be substantially improved for monitoring serially correlated processes after data de-correlation. Xiulin Xie Department of Biostatistics, University of Florida, 2004 Mowry Road, Gainesville, FL 32610. e-mail: xiulin.xie@ufl.edu Peihua Qiu Depar...