RYAN T. POHLIG JULY 9, 2013 PRESENTATION FOR BIOSTATISTICIAN POSITION THE UNIVERSITY OF DELAWARE COLLEGE OF HEALTH SCIENCES 

Outline n  Background ¡  Education ¡  Previous Research n  Philosophy & Approach ¡  As a statistician ¡  In life n  Role as a Biostatistician ¡  Services ¡  Policies 

Background Graduated December 2012 with my Ph.D. in Research Methodology This past spring taught the Business Statistics Sequence at West Chester University- Quantitative Business Analysis I & II (Eco251 & Eco252) 

Educational Background n  Undergrad University of Scranton ¡  Psychology Major n  University of Pittsburgh ¡  Graduate School of Education ¡  Psychology in Education Department ¡  Research Methodology Program 

Graduate School n  Three areas of focus ¡  Research Design ¡  Statistics ¡  Psychometrics n  Majored in Statistics ¡  Dissertation used Structural Equation Modeling (SEM) n  Minored in Psychometrics 

Applied Statistics n  Interest is in applying the quantitative methods and statistical models ¡  My dissertation examined the performance of two methods across a number of conditions for estimating latent interaction effects ¡  Compared Power & Type I error to see how violation of assumptions affected performance n  Fortunate enough to work on a number of different research projects in a variety of fields requiring an array of analyses 

Social Work & Mental Health n  A series of grants from Allegheny County Department of Human Services Office of Children, Youth, and Families n  Evaluated the impact of the various health services the county provides and the impact on youth involved in the Justice System n  A number of different papers and presentations ¡  Cluster Analysis ¡  Exploratory Factor Analysis ¡  SEM (Path Analysis testing mediation) ¡  General Estimating Equation (repeated measures logistic regression) ¡  ANOVA & ANCOVA ¡  Propensity Scores Analysis [on going] 

Education n  Predicting risk for dropping out of High School ¡  Collected validity evidence showing Adolescent Development Screening Inventory (ADSI-E) is effective in predicting drop outs 

Nursing n  Examined relationship between Cytokine levels in Pancreatic Cancer patients & their quality of life ¡  Sequential Regression n  IL-6, TNF-alpha significantly positively related to quality of life n  IL-8, significantly negatively related to quality of life n  IL-10, not significantly related to quality of life 

Bioengineering n  Examined performance of three manual wheelchair propulsion training programs ¡  General Linear Mixed Model (GLMM) with a Time-Varying Covariate ¡  Complex Contrasts n  Found the two experimental methods (Multi- Media Presentation, and Real-Time feedback) were significantly better than the standard training (informational materials) 

Equivalence-Clinical Trials n  Ran analysis for a Clinical Trial testing a joint supplement for rheumatoid arthritis ¡  GLMM n  No significant difference between Placebo, Experimental Supplement, Loading Dose Supplement, Glucosamine-Chondroitin 

Epidemiology n  Estimated prevalence of Diabetes and Metabolic Syndrome in young adults & in the lean ¡  Secondary analyses of a national database using Complex Survey Design (NHANES data) ¡  IFG found in 28.2% ¡  16% had MetS, not with IFG ¡  12% of lean had pre- diabetes 

Non-credited Research n  Advised and consulted ¡  Physical Therapy n  Created a test for treatment of Lower Back Pain ¡  Rehabilitation Science n  Assess efficacy of tele-rehabilitation for wheelchair seating n  Evaluated the psychometric properties of the wheelchair seating tool ¡  Bioengineering n  Evaluated a model that simulates car-crashes while factoring in BMI ¡  Higher Education n  Examining the impact of federal standards in online degree programs’ organizational culture 

Non-credited Research Cont. n  Currently, on a dissertation committee for a Kinesiology and Sport Sciences student at The University of Miami n  “Bicycle shoe insoles and their effect on EMG-indicated leg muscle activation balance in cyclists” ¡  Examining the impact of the shoes/cleats on bikers’ (pedal-cleat-foot) joints and their pedaling motion for injuries ¡  Design and analysis are not finalized 

Previously Helped Researchers n  Research Design and formulating Research Questions ¡  Mediation vs Moderation ¡  Alternative quasi-experimental designs for causality n  Propensity scores n  Regression discontinuity n  Measurement and Psychometrics ¡  Creating a survey ¡  Finding appropriate survey ¡  Evaluating psychometric properties n  Statistical Analysis ¡  Appropriate methods given variables in play ¡  Alternative Strategies for addressing issues 

Software n  Working knowledge of ¡  SPSS ¡  SAS ¡  R ¡  Mplus ¡  EQS ¡  MS office (Excel) n  Learn new ones for convenience of the researcher ¡  Currently learning Stata for the propensity score analysis 

Approach & Philosophy 

Goal for each project n  Establish a study design that will answer the Research Question n  Find and use the most appropriate analysis n  Researcher should be able to explain their research and communicate their results ¡  Conceptually understand what the method and model do ¡  Interpret the results 

Ethical Research n  Ethics ¡  In design ¡  During implementation and while carrying out the research project n  Ethical treatment of data ¡  Check assumptions ¡  Document and disclose all manipulations ¡  Proper adjustment (Post-hoc, use of covariates) and interpretation of statistics ¡  Changing of Research Questions 

Flexibility n  Adaptability ¡  Constraints may be applied: financial, review boards, time ¡  Accommodate reviewers’ opinions n  Opportunities for learning experiences ¡  Learn to work with individuals ¡  Iterative: learn from experiences and apply that going forward n  Novel Research ¡  Unique research might call for unique solutions ¡  Might have to learn a new method ¡  Or might have to instruct the researcher on a method new to them 

Personal Philosophy n  Attitude ¡  Positive “can do” attitude ¡  Integrity ¡  Respect for others ¡  Focus on the long term n  Balance immediate needs with long term n  Solicit ideas from experts and have multi- discipline interaction ¡  Often times perspectives from a related or a foreign field can bring insight 

Work Philosophy n  Team oriented ¡  Balance individual strengths and weaknesses n  Methodical ¡  Prefer to think about and research answers before asserting ideas n  Work-Flow ¡  Actively manage time on the various projects ¡  Immediacy: prioritize based on need ¡  Forecasting upcoming and potential projects 

Authorship n  Authorship policy ¡  APA guidelines: "formulating the problem or hypothesis, structuring the experimental design, organizing and conducting the statistical analysis, interpreting the results, or writing a major portion of the paper" ¡  Authorship earned in a project by contributing directly to the work or in the writing process ¡  I do not require authorship to help with your research 

BIOSTATISTICS Long term vision 

Research n  When thinking about this position I asked a number of people who I had previously worked with and in the Health Sciences, what they would want in a Biostatistician (n = 11) n  Results were fairly consistent centering around four areas 

Results n  Willingness to bring expertise ¡  “[T]here are advanced methods that I don't use/consider because I'm not familiar with them.” ¡  “Important knowing how to ‘frame’ the methods in a way that is appropriate for the research subject matter.” 

Results cont. n  Consistency ¡  “Be consistent, the biostatistician changes the analysis every time you meet with her and is very frustrating.” ¡  “I met with a person, and then 2 hours later they emailed me telling that the analyses were all wrong.” 

Results cont. n  Get in early and stay engaged ¡  “[W]ork with the same statistician throughout so they become familiar with the plans we are trying to propose.” ¡  “Helpful to have a statistician's help in the proposal writing phase... Being able to see the big picture and forecast any issues” ¡  “It would be really helpful for grant submissions since we rarely have access to a statistician for proposals and then we get hammered with that methodology or propose things we can't actually do.” 

Results Cont. n  Accessibility ¡  Use of technology- Skype, Online chat, Conference calls ¡  Availability during non-traditional hours including weekends and nights n  Either in person on the phone, or via email n  “It's amazing that you're nearly instantly reachable. I can't say the same of many professors.” 

At University of Delaware n  Would like to contact those who I would be working with, and find out what they envision the position’s function would be n  Going to adapt to working with different researchers and keep their expectations in mind n  Evaluation and Feedback 

Biostatistician Role n  Providing support for researchers ¡  Study design and methodology ¡  Write analysis plans for grants ¡  Review documents for publications ¡  Power and sample size calculations ¡  Selecting and evaluating instruments and surveys ¡  Data management and cleaning ¡  Statistical analysis ¡  Result reporting ¡  Tables and figure creation 

Accessibility n  Open Door policy ¡  Researchers can seek help when needed n  Hold open “office hours” on a day ¡  Allow faculty and students to come by n  Discuss methodology n  Ask statistical advice n  Aid data management n  Help with analysis or results interpretation 

Educational Training n  Presentations and Workshops ¡  New or “Hot” methods ¡  Find commonalities in the research and present on those methods and analyses ¡  Software usage n  Lectures and seminars in Biostatistics, Psychometrics, Research Methodology and Ethics ¡  Reach out to other departments that offer similar courses or training 

Community n  Establish relationships within and outside the University ¡  Help facilitate research and funding n  Institutional Review Board n  Research Advisory Boards and Councils ¡  Broaden the base of co-researchers n  Promote Multidiscipline research n  Businesses n  Active in Scholarly Committees ¡  Dissertation and Thesis Committees 

Resources n  Continue to build my network of specialists n  Outside resources ¡  I have no problem getting help from others ¡  Tapping resources for ideas and their knowledge n  I am willing to help find answers for people n  Even if it is not for the statistics n  Can call on experts in ¡  Statistics ¡  Measurement (Psychometrics) ¡  Econometrics ¡  Biomechanics ¡  Sports Medicine ¡  Rehabilitation Science & Technology ¡  Education ¡  Social Work & Mental Health ¡  Online assessments/Surveys