Common Mistakes When Fitting Models to Data

Dr. Rossi discusses the common errors that are made when fitting statistical models to data. Focuses on the planning, data analysis, and interpretation phases of a statistical analysis, and highlights the errors that are commonly made by researchers of these phases. The implications of these commonly made errors are discussed along with a discussion of the methods that can be used to prevent these errors from occurring. A prescription for carrying out a correct statistical analysis will be discussed.

Cluster randomized clinical trials in orthodontics: design, analysis and reporting issues

Cluster randomized trials (CRTs) use as the unit of randomization clusters, which are usually defined as a collection of individuals sharing some common characteristics. Common examples of clusters include entire dental practices, hospitals, schools, school classes, villages, and towns. Additionally, several measurements (repeated measurements) taken on the same individual at different time points are also considered to be clusters. In dentistry, CRTs are applicable as patients may be treated as clusters containing several individual teeth. CRTs require certain methodological procedures during sample calculation, randomization, data analysis, and reporting, which are often ignored in dental research publications. In general, due to similarity of the observations within clusters, each individual within a cluster provides less information compared with an individual in a non-clustered trial. Therefore, clustered designs require larger sample sizes compared with non-clustered randomized designs, and special statistical analyses that account for the fact that observations within clusters are correlated. It is the purpose of this article to highlight with relevant examples the important methodological characteristics of cluster randomized designs as they may be applied in orthodontics and to explain the problems that may arise if clustered observations are erroneously treated and analysed as independent (non-clustered).

From big smartphone data to worldwide research: The Mobile Data Challenge

This paper presents an overview of the Mobile Data Challenge (MDC), a large-scale research initiative aimed at generating innovations around smartphone-based research, as well as community-based evaluation of mobile data analysis methodologies. First, we review the Lausanne Data Collection Campaign (LDCC), an initiative to collect unique longitudinal smartphone dataset for the MDC. Then, we introduce the Open and Dedicated Tracks of the MDC, describe the specific datasets used in each of them, discuss the key design and implementation aspects introduced in order to generate privacy-preserving and scientifically relevant mobile data resources for wider use by the research community, and summarize the main research trends found among the 100+ challenge submissions. We finalize by discussing the main lessons learned from the participation of several hundred researchers worldwide in the MDC Tracks.

SURVIVAL PREDICTION FOR BRAIN TUMOR PATIENTS USING GENE EXPRESSION DATA

Brain tumor is one of the most aggressive types of cancer in humans, with an estimated median survival time of 12 months and only 4% of the patients surviving more than 5 years after disease diagnosis. Until recently, brain tumor prognosis has been based only on clinical information such as tumor grade and patient age, but there are reports indicating that molecular profiling of gliomas can reveal subgroups of patients with distinct survival rates. We hypothesize that coupling molecular profiling of brain tumors with clinical information might improve predictions of patient survival time and, consequently, better guide future treatment decisions. In order to evaluate this hypothesis, the general goal of this research is to build models for survival prediction of glioma patients using DNA molecular profiles (U133 Affymetrix gene expression microarrays) along with clinical information. First, a predictive Random Forest model is built for binary outcomes (i.e. short vs. long-term survival) and a small subset of genes whose expression values can be used to predict survival time is selected. Following, a new statistical methodology is developed for predicting time-to-death outcomes using Bayesian ensemble trees. Due to a large heterogeneity observed within prognostic classes obtained by the Random Forest model, prediction can be improved by relating time-to-death with gene expression profile directly. We propose a Bayesian ensemble model for survival prediction which is appropriate for high-dimensional data such as gene expression data. Our approach is based on the ensemble "sum-of-trees" model which is flexible to incorporate additive and interaction effects between genes. We specify a fully Bayesian hierarchical approach and illustrate our methodology for the CPH, Weibull, and AFT survival models. We overcome the lack of conjugacy using a latent variable formulation to model the covariate effects which decreases computation time for model fitting. Also, our proposed models provides a model-free way to select important predictive prognostic markers based on controlling false discovery rates. We compare the performance of our methods with baseline reference survival methods and apply our methodology to an unpublished data set of brain tumor survival times and gene expression data, selecting genes potentially related to the development of the disease under study. A closing discussion compares results obtained by Random Forest and Bayesian ensemble methods under the biological/clinical perspectives and highlights the statistical advantages and disadvantages of the new methodology in the context of DNA microarray data analysis.

The transformation of a health care system: An ecology perspective

Beginning in the early 1980s, the health care system experienced momentous realignments. Fundamental changes in structures of traditional health care organizations, shifts in authority and relationships of professionals and institutions, and the increasing influence of managed care contributed to a relatively stable industry entering into a state of turbulence. The dynamics of these changes are recurring themes in the health services literature. The purpose of this dissertation was to examine the content of this literature over a defined time period and within the perspective of a theory of organizational change. ^ Using a theoretical framework based upon the organizational theory known as Organizational Ecology, secondary data from the period between 1983 and 1994 was reviewed. Analysis of the literature identified through a defined search methodology was focused upon determining the manner in which the literature characterized changes that were described. Using a model constructed from fundamentals of Organizational Ecology with which to structure an assessment of content, literature was summarized for the manner and extent of change in specific organizational forms and for the changes in emphasis by the environmental dynamics directing changes in the population of organizations. Although it was not the intent of the analysis to substantiate causal relationships between environmental resources selected as the determinants of organizational change and the observed changes in organizational forms, the structured review of content of the literature established a strong basis for inferring such a relationship. ^ The results of the integrative review of the literature and the power of the appraisal achieved through the theoretical framework constructed for the analysis indicate that there is considerable value in such an approach. An historical perspective on changes which have transformed the health care system developed within a defined organizational theory provide a unique insight into these changes and indicate the need for further development of such an analytical model. ^

Injury severity and mortality of adult zebra crosswalk and non-zebra crosswalk road crossing accidents: a cross-sectional analysis

PRINCIPALS Over a million people worldwide die each year from road traffic injuries and more than 10 million sustain permanent disabilities. Many of these victims are pedestrians. The present retrospective study analyzes the severity and mortality of injuries suffered by adult pedestrians, depending on whether they used a zebra crosswalk. METHODS Our retrospective data analysis covered adult patients admitted to our emergency department (ED) between 1 January 2000 and 31 December 2012 after being hit by a vehicle while crossing the road as a pedestrian. Patients were identified by using a string term. Medical, police and ambulance records were reviewed for data extraction. RESULTS A total of 347 patients were eligible for study inclusion. Two hundred and three (203; 58.5%) patients were on a zebra crosswalk and 144 (41.5%) were not. The mean ISS (injury Severity Score) was 12.1 (SD 14.7, range 1-75). The vehicles were faster in non-zebra crosswalk accidents (47.7 km/n, versus 41.4 km/h, p<0.027). The mean ISS score was higher in patients with non-zebra crosswalk accidents; 14.4 (SD 16.5, range 1-75) versus 10.5 (SD13.14, range 1-75) (p<0.019). Zebra crosswalk accidents were associated with less risk of severe injury (OR 0.61, 95% CI 0.38-0.98, p<0.042). Accidents involving a truck were associated with increased risk of severe injury (OR 3.53, 95%CI 1.21-10.26, p<0.02). CONCLUSION Accidents on zebra crosswalks are more common than those not on zebra crosswalks. The injury severity of non-zebra crosswalk accidents is significantly higher than in patients with zebra crosswalk accidents. Accidents involving large vehicles are associated with increased risk of severe injury. Further prospective studies are needed, with detailed assessment of motor vehicle types and speed.

A new Time-Of-Flight Mass Spectrometer (TOF-MS) for noble gas analysis

Noble gas analysis in early solar system materials, which can provide valuable information about early solar system processes and timescales, are very challenging because of extremely low noble gas concentrations (ppt). We therefore developed a new compact sized (33 cm length, 7.2cm diameter, 1.3 L internal volume) Time-of-Flight (TOF) noble gas mass spectrometer for high sensitivity. We call it as Edel Gas Time-of-flight (EGT) mass spectrometer. The instrument uses electron impact ionization coupled to an ion trap, which allows us to ionize and measure all noble gas isotopes. Using a reflectron set-up improves the mass resolution. In addition, the reflectron set-up also enables some extra focusing. The detection is via MCPs and the signals are processed either via ADC or TDC systems. The objective of this work is to understand the newly developed Time-Of-Flight (TOF) mass spectrometer for noble gas analysis in presolar grains of the meteorites. Chapter 1 briefly introduces the basic idea and importance of the instrument. The physics relevant to time-of-flight mass spectrometry technique is discussed in the Chapter 2 and Chapter 3 will present the oxidation technique of nanodiamonds of the presolar grains by using copper oxide. Chapter 4 will present the details about EGT data analysis software. Chapter 5 and Chapter 6 will explain the details about EGT design and operation. Finally, the performance results will be presented and discussed in the Chapter 7, and whole work is summarized in Chapter 8 and also outlook of the future work is given.

Passive wireless sensor systems can recognize activites of daily living

The ability to determine what activity of daily living a person performs is of interest in many application domains. It is possible to determine the physical and cognitive capabilities of the elderly by inferring what activities they perform in their houses. Our primary aim was to establish a proof of concept that a wireless sensor system can monitor and record physical activity and these data can be modeled to predict activities of daily living. The secondary aim was to determine the optimal placement of the sensor boxes for detecting activities in a room. A wireless sensor system was set up in a laboratory kitchen. The ten healthy participants were requested to make tea following a defined sequence of tasks. Data were collected from the eight wireless sensor boxes placed in specific places in the test kitchen and analyzed to detect the sequences of tasks performed by the participants. These sequence of tasks were trained and tested using the Markov Model. Data analysis focused on the reliability of the system and the integrity of the collected data. The sequence of tasks were successfully recognized for all subjects and the averaged data pattern of tasks sequences between the subjects had a high correlation. Analysis of the data collected indicates that sensors placed in different locations are capable of recognizing activities, with the movement detection sensor contributing the most to detection of tasks. The central top of the room with no obstruction of view was considered to be the best location to record data for activity detection. Wireless sensor systems show much promise as easily deployable to monitor and recognize activities of daily living.

Single crowns versus conventional fillings for the restoration of root-filled teeth.

BACKGROUND Endodontic treatment involves removal of the dental pulp and its replacement by a root canal filling. Restoration of root filled teeth can be challenging due to structural differences between vital and non-vital root-filled teeth. Direct restoration involves placement of a restorative material e.g. amalgam or composite, directly into the tooth. Indirect restorations consist of cast metal or ceramic (porcelain) crowns. The choice of restoration depends on the amount of remaining tooth, and may influence durability and cost. The decision to use a post and core in addition to the crown is clinician driven. The comparative clinical performance of crowns or conventional fillings used to restore root-filled teeth is unknown. This review updates the original, which was published in 2012. OBJECTIVES To assess the effects of restoration of endodontically treated teeth (with or without post and core) by crowns versus conventional filling materials. SEARCH METHODS We searched the following databases: the Cochrane Oral Health Group's Trials Register, CENTRAL, MEDLINE via OVID, EMBASE via OVID, CINAHL via EBSCO, LILACS via BIREME. We also searched the reference lists of articles and ongoing trials registries.There were no restrictions regarding language or date of publication. The search is up-to-date as of 26 March 2015. SELECTION CRITERIA Randomised controlled trials (RCTs) or quasi-randomised controlled trials in participants with permanent teeth that have undergone endodontic treatment. Single full coverage crowns compared with any type of filling materials for direct restoration or indirect partial restorations (e.g. inlays and onlays). Comparisons considered the type of post and core used (cast or prefabricated post), if any. DATA COLLECTION AND ANALYSIS Two review authors independently extracted data from the included trial and assessed its risk of bias. We carried out data analysis using the 'treatment as allocated' patient population, expressing estimates of intervention effect for dichotomous data as risk ratios, with 95% confidence intervals (CI). MAIN RESULTS We included one trial, which was judged to be at high risk of performance, detection and attrition bias. The 117 participants with a root-filled, premolar tooth restored with a carbon fibre post, were randomised to either a full coverage metal-ceramic crown or direct adhesive composite restoration. None experienced a catastrophic failure (i.e. when the restoration cannot be repaired), although only 104 teeth were included in the final, three-year assessment. There was no clear difference between the crown and composite group and the composite only group for non-catastrophic failures of the restoration (1/54 versus 3/53; RR 0.33; 95% CI 0.04 to 3.05) or failures of the post (2/54 versus 1/53; RR 1.96; 95% CI 0.18 to 21.01) at three years. The quality of the evidence for these outcomes is very low. There was no evidence available for any of our secondary outcomes: patient satisfaction and quality of life, incidence or recurrence of caries, periodontal health status, and costs. AUTHORS' CONCLUSIONS There is insufficient evidence to assess the effects of crowns compared to conventional fillings for the restoration of root-filled teeth. Until more evidence becomes available, clinicians should continue to base decisions about how to restore root-filled teeth on their own clinical experience, whilst taking into consideration the individual circumstances and preferences of their patients.

Estimating parameters in Markov models for longitudinal studies with missing data or surrogate outcomes

The discrete-time Markov chain is commonly used in describing changes of health states for chronic diseases in a longitudinal study. Statistical inferences on comparing treatment effects or on finding determinants of disease progression usually require estimation of transition probabilities. In many situations when the outcome data have some missing observations or the variable of interest (called a latent variable) can not be measured directly, the estimation of transition probabilities becomes more complicated. In the latter case, a surrogate variable that is easier to access and can gauge the characteristics of the latent one is usually used for data analysis. ^ This dissertation research proposes methods to analyze longitudinal data (1) that have categorical outcome with missing observations or (2) that use complete or incomplete surrogate observations to analyze the categorical latent outcome. For (1), different missing mechanisms were considered for empirical studies using methods that include EM algorithm, Monte Carlo EM and a procedure that is not a data augmentation method. For (2), the hidden Markov model with the forward-backward procedure was applied for parameter estimation. This method was also extended to cover the computation of standard errors. The proposed methods were demonstrated by the Schizophrenia example. The relevance of public health, the strength and limitations, and possible future research were also discussed. ^

Manifestation of depression symptoms among Mexican American adolescents

Background. Previous findings reported more depression among Mexican American adolescents than among adolescents from other groups. There has been minimal research published on distribution of depression symptoms among Mexican American adolescents and practically no data has been published from community samples. ^ Objective. To examine the phenomenology of DSM-IV major depression symptoms across groups defined by ethnic status, by gender and language use focusing on the Mexican American group. ^ Methods. Secondary data from 2624 adolescents (ages 10-17) among three ethnic subgroups, Mexican (26.7%), African (45%) and Anglo Americans (28.3%), was analyzed. Data come from the Teen Life Changes (TLC) Survey conducted in 1994 by Roberts et al. (1997). A self-report questionnaire, which includes the DSD scale to measure depression, was used. ^ Results. Analysis of data showed significant differences among youth in the phenomenology of depression symptoms by ethnicity, by gender and by language use at home. ^ Conclusion. This study adds knowledge to the psychopathology and mental health literature from the identification of depression symptoms profile as well as permits the design of more appropriate policy for prevention and intervention programs among culturally diverse youth. ^

Prevalence and risk factors of methicillin-resistant Staphylococcus aureus in critically-ill hospitalized patients in a tertiary care center in Houston, Texas: An active surveillance pilot project

Objective. The objective of this study is to determine the prevalence of MRSA colonization in adult patients admitted to intensive care units at an urban tertiary care hospital in Houston, Texas and to evaluate the risk factors associated with colonization during a three month active-screening pilot project. Design. This study used secondary data from a small cross-sectional pilot project. Methods. All patients admitted to the seven specialty ICUs were screened for MRSA by nasal culture. Results were obtained utilizing the BD GeneOhm™ IDI-MRSA assay in vitro diagnostic test, for rapid MRSA detection. Statistical analysis was performed using the STATA 10, Epi Info, and JavaStat. Results . 1283/1531 (83.4%) adult ICU admissions were screened for nasal MRSA colonization. Of those screened, demographic and risk factor data was available for 1260/1283 (98.2%). Unresolved results were obtained for 73 patients. Therefore, a total of 1187/1531 (77.5%) of all ICU admissions during the three month study period are described in this analysis. Risk factors associated with colonization included the following: hospitalization within the last six months (odds ratio 2.48 [95% CI, 1.70-3.63], p=0.000), hospitalization within the last 12 months, (odds ratio 2.27 [95% CI, 1.57-3.80], p=0.000), and having diabetes mellitus (odds ratio 1.63 [95% CI, 1.14-2.32], p=0.007). Conclusion. Based on the literature, the prevalence of MRSA for this population is typical of other prevalence studies conducted in the United States and coincides with the continual increasing trend of MRSA colonization. Significant risk factors were similar to those found in previous studies. Overall, the active surveillance screening pilot project has provided valuable information on a population not widely addressed. These findings can aid in future interventions for the education, control, prevention, and treatment of MRSA. ^

Impact of oral contraceptives on weight, body composition, and cardiovascular disease risk factors in women age 18--30 years participating in a 15-week exercise program

Objective. To assess differences in body weight, body composition, total cholesterol, blood pressure, and blood glucose between OC users and non-users age 18-30 y before and after a 15-week cardiovascular exercise program in Houston, TX from 2003 to 2007.^ Study Design. Secondary analysis of prospective data. ^ Study Subjects. 453 Non-Hispanic white (NHW), Hispanic, and African American (AA) women age 18-30 y with no previous live birth, a history of menstruating, no use of other hormonal contraceptives or medications, no menopause or hysterectomy, and no current pregnancies.^ Measurements. Demographic data, medication use, and menstrual history were assessed via self-administered questionnaires at baseline. Anthropometric and laboratory measures were taken at baseline and 15-weeks. ^ Data Analysis. Linear regression assessed the association between OC use and study variables at baseline, and the change in study variables from baseline to 15-weeks. Logistic regression assessed the association between OC use and CVD risk. Each analysis was also stratified by race/ethnicity. ^ Results. At baseline, OC users had higher total cholesterol (p<.0005) and were above cholesterol risk cut points for CVD (OR=4.3, 95% CI=2.4-7.7) compared to non-users. At baseline, OC use was also associated with higher diastolic blood pressure (p=.018) compared to non-users, primarily in non-Hispanic whites (p=.007). OC use was associated with lower blood glucose compared to non-users in Hispanics only (p=.008). OC use was associated with absolute change in diastolic blood pressure (p=.044) and total cholesterol (p=.003). There was evidence that OC use may affect individuals differently based on race/ethnicity for certain obesity and CVD risk factors.^ Conclusions. OC users and non-users responded similarly to a 15-week cardiovascular exercise program. Exceptions included a greater change in diastolic blood pressure and total cholesterol among NHW and Hispanic OC users compared to non-users after exercise intervention. At baseline, OC use was associated with diastolic blood pressure and was most strongly associated with increased levels of total cholesterol. OC users were at greater risk of having total cholesterol above CVD risk cut points than non-users.^

Youth population's health issues in Comal County, Texas

This study assesses adolescent's health issues in Comal County, TX. Adolescents are defined as youth between the ages of 12 to 17 years of age, who resided in Comal County during the time period of 2000 to 2007. The analysis focused on high risk behaviors including use of gateway drugs—tobacco and alcohol; illegal substance use; and reproductive health related indicators, including sexual activity, sexually transmitted diseases, and pregnancy. This study is based on the primary and secondary data collected as part of the 2008 Comal County Community Assessment. It compares findings from the primary data sources to extant data from four secondary data sources including: (1) The Centers for Disease Control & Prevention (national) Healthy People 2010; (2) The Centers for Disease Control & Prevention, Youth Risk Behavior Surveillance Survey, 2007; (3) The Texas Department of State Health Services, 2000 to 2007; and The Pride Survey (Local and Statewide). The methods are drawn from the literature on "rapid epidemiologic appraisal" (Annett H. & Rifkin S. B., 1988). The study focus on corroborating the perceptions, subjective concerns, opinions and beliefs of the Comal County key stakeholders and community participants with qualitative and quantitative indicators of health and well being. The value of this approach is to inform community leaders using a public health perspective and evidence in their decisions about priority setting and resources allocation activities for prevention of high risk behaviors and promotion of adolescent health and well being. ^