Short Courses-at-a-Glance

Full Day 9:00am - 5:30pm
SC-FD1	"Introduction to Medical Decision Making"	Beginner
SC-FD2	"Psychology of Medical Decision Making"	Beginner
SC-FD3	"Meta-Analysis: Statistical Methods for Combining the Results of Independent Studies"	Beginner
SC-FD4	"The Basics of Estimation and Uncertainty in Cost-Effectiveness Analysis"	Intermediate
SC-FD5	"Why Aren't Physicians' Practices Evidence-Based? Cognitive and Environmental	Intermediate
	Challenges to Evidence Based Practice"
Morning Half Day 9:00am - 12:30pm
SC-AM1	"Computational Modeling of Decision Processes"	Beginner
SC-AM2	"Introduction to Infectious Disease Transmission Modeling"	Beginner
SC-AM3	"Medical Decision Management"	Beginner
SC-AM4	"Multicriteria Methods for Medical Decision Making"	Beginner
SC-AM5	"Introduction to Conjoint Analysis of Healthcare Programs"	Intermediate
SC-AM6	"Methods for Causal Analysis Using Observational (Non-Experimental) Data:	Intermediate
	Addressing Selection Bias and Other Forms of Endogeneity in Comparative Effectiveness Analysis"
SC-AM7	"Microsoft Excel: Advanced Skills for Efficiency and Modeling"	Advanced
SC-AM8	"How to Respond to an NIH Review"	Advanced
SC-AM9	"Psychology of Medical Decision Making" First Module	Beginner
Afternoon Half Day 2:00pm - 5:30pm
SC-PM1	"Becoming a Strong Clinician-Investigator Mentor"	Beginner
SC-PM2	"Causal Inference and Causal Diagrams in Medical Decision Making"	Beginner
SC-PM3	"Introduction to Discrete-Event Simulation for Healthcare"	Beginner
SC-PM4	"Medical Decision Making with Bayesian Networks"	Beginner
SC-PM5	"Mathematical Modeling of Infectious Diseases: Beyond the Basics"	Intermediate
SC-PM6	"Prediction Models in Medicine: Development, Evaluation and Implementation"	Intermediate
SC-PM7	"Economic Assessment in Clinical Trial"	Advanced
SC-PM8	"How to Obtain a Career Development Award"	Advanced
SC-PM9	"Psychology of Medical Decision Making" Second Module	Beginner

Complete detailed abstracts for each Short Course can be found below

Background:
This course is intended for individuals new to decision analysis who wish to learn the basic principles of formulating and analyzing clinical decisions. The course is "hands-on" and uses in-class exercises to teach the building blocks of decision analysis. These blocks are Bayes' rule, interpreting the results of diagnostic tests, formulating a medical decision problem, measuring utilities and risk attitude, calculating expected utility, and performing sensitivity analysis, as well as (at an elementary level) cost-effectiveness analysis.

Objectives and Course Description:
First participants are made familiar with the methods of MDM through teaching and exercises. Then the course o brings these methods home by applying them to participants' clinical or research-questions o in addition, in this way aims to show what MDM may or may not do, and thus interactively demonstrate its strengths and limitations at the level of the individual, of patient groups and of society

To this end course participants are invited to send in their own clinical or research decision questions by email before the course,(a selection of) which will be used for teaching purposes.

Format, Requirements, Target Audience:
Participants will be asked to send in a short form with their own decision problem(s) by email.

Short Course Sunday, October 18, 2009
Full Day 9:00am - 5:30pm
SC-FD2 Psychology of Medical Decision Making
** This course will be taught in two half-day modules which can be taken as a full day course or as two half day courses.
Course size limit: none Level: Beginner
Faculty:
Alan J. Schwartz, PhD, University of Illinois at Chicago
Olga Kostopoulou, PhD, MSc, King's College London, UK
Robert Hamm, PhD, University of Oklahoma Health Sciences Center, Oklahoma City, OK

Jamie Brehaut, PhD, Ottawa Hospital Research Institute, Ottawa, Canada

Background:
This course introduces participants to descriptive findings and psychological theory related to making decisions in health and medicine. Knowledge of the psychology of medical decision making can be used to improve explanations and predictions of patient and physician behavior, and to design behavioral interventions.

Objectives and Course Description:

• To understand the psychological processes involved in medical decision making by patients and physicians, in the contexts of probability/diagnosis/risk and preference/choice.
• To describe research methods used for studying the psychology of medical decision making.
• To understand patient and physician vulnerability to cognition based errors.
• To develop approaches to support decision makers based on psychological theory.

Concepts, skills, or experiences that the participants will acquire by attending the course.
The characteristics of the cognitive system: large memory, limited attention span, pattern recognition ability. Differences between analytical and intuitive processing and their implications for judgment and decision making The judgments needed for decision making, and the strategies people use to draw on their knowledge in making judgments. Types of errors physicians and patients may make when thinking about decisions. Strategies for minimizing the impact of errors and for supporting accurate thinking about decisions.

Format, Requirements, Target Audience:
This course will be taught in two half-day modules which can be taken as a full day course or as two half day courses. The first module will be offered in the AM will focus on the psychology of probability, including diagnosis and risk perception.
The second module will be offered in the PM will focus on the psychology of preference and choice, including valuation and descriptive models of decision making.

The course involves brief lectures, demonstrations, and small and large group discussions. Attendees should expect to be actively involved in discussions of psychological phenomena as they relate to their clinical, teaching, or research interests.

Short Course Full Day Sunday, October 18, 2009
9:00am - 5:30pm
SC-FD3 Meta-Analysis: Statistical Methods for Combining the Results of Independent Studies
Course size limit: none Level: Beginner
Faculty:
Ingram Olkin, MA, PhD, Stanford University, Stanford, CA
Thomas Trikalinos, MD, PhD, School of Medicine (Clinical), Tufts Medical Center, Boston, MA

Background:
Meta-analysis is a formal, systematic method to synthesize the results of independent studies, considering and integrating the combined weight of evidence to determine the effect of an intervention. Meta-analysis is being used increasingly in the medical and health sciences to inform and guide practice and policy, in areas as disparate as estimating the effectiveness of mammography in detection of breast cancer and the consistency of gene-disease association studies. A Google Scholar search on meta-analysis identified 589,000 hits in medicine, 293,000 in health policy, and 102,000 in genetics. The information explosion in almost every field coupled with the movement towards evidence-based decision making and cost-effective analysis has catalyzed development of more rigorous procedures to synthesize the results of independent studies.

Objectives and Course Description:
This workshop will provide an historical perspective of meta-analysis, and discuss methodological issues such as various types of bias and heterogeneity on the conduct and interpretation of meta-analyses. There will be extensive discussion of the appropriateness and use of statistical methods for combining data across studies, including nonparametric and parametric models; effect sizes for proportions, fixed versus random effects, regression and ANOVA models; multivariate models for proportions and standardized mean differences, treatment of zero cells, models with missing data, and special methods and issues in genetic applications.

• Understand the potential value of and theory underlying the conduct of meta-analysis of independent studies
• Understand conditions under which meta-analyses can be performed and common factors that limit or confound the metaanalysis conduct and interpretation.
• Learn and understand a range of statistical methods for analyzing and interpreting meta-analysis studies

Format, Requirements, Target Audience:
Didactic lectures and interactive discussion of theory, potential confounders and limitations, and statistical methods, using case study examples from published medical literature.

Short Course Sunday, October 18, 2009
Full Day 9:00am - 5:30pm
SC-FD4 The Basics of Estimation and Uncertainty in Cost-effectiveness Analysis
Course size limit: none Level: Intermediate
Faculty:
Jeffrey S. Hoch, MA, PhD, Centre for Research on Inner City Health, St. Michael's Hospital, University of Toronto, Toronto, ON, Canada
Katia Noyes, PhD, MPH, Division of Health Services Research, Community & Preventive Medicine, University of Rochester, Rochester, NY
Ahmed Bayoumi, MD, MSc, Centre for Research on Inner City Health, St. Michael's Hospital, University of Toronto, Toronto, ON, Canada
Elisabeth Fenwick, BA, MSc, PhD, University of Glasgow, Glasgow, Scotland

Background:
Many funding agencies and academic publications now require cost-effectiveness estimates and measures of uncertainty around those estimates. This course will introduce the concepts behind probabilistic sensitivity analysis and net benefit regression and provide participants with practical skills to produce numeric and graphical outputs that reflect estimates and their uncertainty. This course is directed towards health outcomes researchers in academia, industry, government or regulatory bodies without special training in health economics. Participants will be expected to know how to perform simple regeression analysis.

Objectives and Course Description:
This course will provide participants with the conceptual background to understand why it is important to produce cost-effectiveness estimates and measure their uncertainty and will present some commonly used methods to do this. Specific objectives are:

1. To understand the conceptual basis for measuring uncertainty in costs, effects, and cost-effectiveness estimates.
2. To understand the relationship between the scatter plot of incremental costs and incremental effects and how these relate to confidence ellipses and cost-effectiveness acceptability curves
3. To be able to build confidence ellipses and cost-effectiveness acceptability curves using Excel or Stata
4. To understand how regression analysis can be used to estimate cost-effectiveness in person-level data using the net benefit approach.
5. To understand how regression analysis can be used to characterize uncertainty in person-level data using the net benefit approach.
6. To understand the strengths and weaknesses of alternative methods for presenting uncertainty in cost-effectiveness analyses.

Format, Requirements, Target Audience:
The course will consist of didactic learning (theory bursts) followed by interactive tutorials ("hands on" exercise). Participants are required to bring their own laptops loaded with either Excel or Stata. Using guidance and data files provided by the instructors, each participant will learn to generate cost-effectiveness acceptability curves (CEACs) and confidence ellipses in Excel and Stata.

Short Course Full Day Sunday, October 18, 2009
9:00am - 5:30pm
SC-FD5 Why Aren't Physicians' Practices Evidence-Based?
Cognitive and Environmental Challenges to Evidence-Based Practice
Course size limit: none Level: Intermediate
Faculty:
Roy M. Poses, MD, Alpert School of Medicine, Brown University, Providence, RI, and the Foundation for Integrity and Responsibility in Medicine, Warren, RI
Wally R. Smith, MD, Division of Quality Healthcare, Virginia Commonwealth University, Richmond, VA

Background:
Evidence-based medicine (EBM) integrates the best scientific evidence with clinical expertise and patient values. However, physicians often fail to practice in accord with EBM principles and many attempts to change physician behavior to make it more evidence based have failed.

Objectives and Course Description:
This course will examine reasons physician practice at times fails to adhere with EBM principles and explore promising interventions to improve EBM-based practice. The impact of human thinking strategies designed to cope with inherent cognitive limitations that may lead to judgments and decisions that fail to conform with normative ideals (with emphasis on judgment and decision biases and heuristics) will be discussed for each stage of the evidence-based decision making process: identifying options and their outcomes; assessing probability of outcomes; assessing value of options; and combining information to make a decision. The course will examine the impact of organizations and culture on medical decision making and practice, which increasingly is provided from within large organizations whose leadership, structures, processes, incentives and environments (e.g., time and economic pressure; conflicts-of-interest) may undermine EBM-based care. The growing presence and impact of stealth marketing, special politically correct pleadings, suppression and manipulation of research, perverse bureaucratic and financial incentives, and intimidation and coercion that may challenge and undermine elements of an EBM approach will be identified, described and discussed.
The course will conclude with review and exploration of promising approaches based on findings in the cognitive psychology to address physicians' human cognitive limitations that may help physicians practice more in accord with EBM, as well as general approaches to defend evidence-based decision practice from health care environment threats.

• Understand principles and impact of human cognitive behavior used to cope with cognitive limitations at each stage of the clinical decision making process.
• Understand organizational challenges to EBM practice.
• Understand environmental and social/cultural threats to increasing EBM practice.
• Understand promising approaches for facilitating physician EBM practice.

Format, Requirements, Target Audience:
Didactic review of the cognitive psychology and organizational theory literature and interactive critical review and discussion of case studies and interventions to improve evidence-based practice.

Short Course AM Half Day Sunday, October 18, 2009
9:00am - 12:30pm
SC-AM1 Computational Modeling of Decision Processes
Course size limit: none Level: Beginner
Faculty:
Joseph G. Johnson, PhD, Miami University, Oxford, OH

Background:
Will present a primer on understanding and describing individual decision processes using computational models, with a specific (sequential sampling) example.

Objectives and Course Description:
This Short Course will survey recent advances in process-oriented models of individual decision behavior, and provide an in-depth tutorial in one popular and successful framework (sequential sampling). Rather than resorting to paramorphic utility representations that are meant primarily to predict decision outcomes, I will introduce several contemporary formal models of the underlying (cognitive) processes that are assumed to produce these outcomes. The benefits of this approach will be presented and specific applications within the framework of Decision Field Theory (see Busemeyer & Johnson, 2004) will be included.

• Understand basic characteristics of computational process models, including key benefits
• Understand the range of dependent variables (choices, confidence ratings, response times, process-tracing measures, etc.) available to critically test these models
• Understand basic logic for model testing and comparison, using these multiple indicators
• Conduct simulations of an exemplar computational model to understand its behavior
• Apply an exemplar computational model to real data to understand model fitting

Format, Requirements, Target Audience:
This Short Course will involve both lecture and hands-on applications. Basic understanding of probability and matrix algebra will be beneficial, but not required. Basic knowledge of programming in MATLAB or another language will be beneficial, but not required. Step-by-step examples will be covered using programs written in MATLAB. These will also be made available such that participants with personal laptops with MATLAB installed will be able to follow along and also conduct additional hands-on applications. The course can be conducted at a differential skill level, with advanced coverage and auxiliary material for those who desire.

Short Course AM Half Day Sunday, October 18, 2009
9:00am - 12:30pm
SC-AM2 Introduction to Infectious Disease Transmission Modeling
Course size limit:20 Level: Beginner
Faculty:
Chris Bauch, PhD, University of Guelph, Guelph, Canada
Beate Sander, RN, MBA, MEcDev, The University of Toronto, Canada

Background:
Many issues related to infectious disease control currently challenge decision makers in public health and medicine. Modeling the effectiveness and cost-effectiveness of interventions may serve as valuable tools in the formulation of policy. However, in order to be credible, such models need to capture important facets of infectious diseases, such as changes in risk of infection over time as susceptible individuals become vaccinated and infected; the tendency of infectious disease epidemics to occur at regular (e.g. seasonal) intervals, and the potential benefits associated with such phenomena as "herd immunity".

Objectives and Course Description:
This course will introduce participants to the principles of infectious disease modeling. This Short Course will review key concepts in infectious disease epidemiology such as the natural history of infection in individuals, the basic reproductive number, and herd immunity. The session will continue with an overview of different infectious disease modeling techniques, describe how they differ compared to non-infectious disease modeling techniques and when and when not to apply various modeling techniques. In the hands-on section, participants will construct and analyze a simple compartmental model using the modeling software. Finally, a case study (cost-utility analysis of Hepatitis A vaccination) will be discussed.

• Review infectious disease epidemiology.
• Describe infectious disease modeling techniques.
• Show how to construct and analyze simple infectious disease models, including examples of required software.
• Discuss case studies of infectious disease models supporting health policy decision making.

Format, Requirements, Target Audience:
This course will use a variety of formats including didactic lecture, individual and group work, problem solving and discussion of real world examples. Material for the course will be provided electronically including step-by-step guidance on model construction. The example model and case study will be provided electronically for future reference. This course is aimed at beginners. However, participants should have a very basic level of college-level calculus including basic differential equations. Participants should bring a laptop computer with a CD drive and/or USB storage device. Instructions on how to obtain the software and reading material for the case study discussion will be sent to participants prior to the course.

Short Course AM Half Day Sunday, October 18, 2009
9:00am - 12:30pm
SC-AM3 Medical Decision Management
Course size limit: none Level: Beginner
Faculty:
J. Frank Yates, PhD, University of Michigan, Ann Arbor, MI
Lee A. Green, MD, MPH, University of Michigan, Ann Arbor, MI

Background:
Good "decision managers" help decision makers themselves work through all the issues that must be resolved in the process of making any practical decision. They do this by drawing on any and everything they know about prescriptive and descriptive decision science (e.g., what patients and physicians are naturally inclined to do that compromises their interests and values). It provides an important and comfortable complement or alternative to approaches such as decision analysis and the shared decision making paradigm. The result is decision making that is more effective, including easier, than what occurs in everyday practice (cf. J. F. Yates, Decision Management, Jossey-Bass, 2003).

Objectives and Course Description:
At the end of this course, the participant will:

• understand essential concepts and effective approaches in decision management as they appear in medical practice
• know how to apply and continually develop his or her personal decision management skills indefinitely after the course is completed

These aims will be achieved primarily through guided, collaborative exercises built around a scenario representative of those common in health care (that of a group practice of primary care physicians whose patients have an excessively high rate of admissions for uncontrolled asthma).

Format, Requirements, Target Audience:
As is common (and effective) in decision management classes, participants will be randomly assigned to small teams. For the focal scenario, each team will be required to collaboratively address each of the core tasks demanded in any decision management enterprise. These include interpreting a real medical problem in terms of standard decision management constructs, settling on what "decision quality" means (or should mean) in the given situation, and developing effective means of meeting every standard decision making demand, e.g., surfacing non-obvious potential decision outcomes, acquiring accurate judgments, and achieving decision acceptance. Toward the end of the course, there will be an exercise and discussion concerning how participants can assure that they continually sharpen their decision management skills and habits indefinitely after returning home. This discussion will include pointers toward developing a manageable yet effective custom of reading accessible, high-value decision management and decision science sources (including from the Web and even the popular press).

Short Course AM Half Day Sunday, October 18, 2009
9:00am - 12:30pm
SC-AM4 Multicriteria Methods For Medical Decision Making
Course size limit: 30 Level: Beginner
Faculty:
James Dolan, MD, University of Rochester, Rochester, NY

Background:
Many medical decisions involve complex trade-offs between alternatives that offer different sets of advantages and disadvantages. These decisions are further complicated by the ever-present uncertainty about future outcomes, insufficient data, practical constraints, and the involvement of multiple stakeholders. A variety of multi-criteria decision making methods have been developed to help both individuals and groups make better decisions in circumstances like these. They have been successfully applied in both medical and non-medical settings to improve the quality of the decision making process, enhance communication among involved stakeholders, and to study decision-making.

Objectives and Course Description:
The overall goal of this course is to provide participants with a practical introduction to the use of multi-criteria methods in medical decision making research. The course objectives are:

1. To acquaint participants with the range of multi-criteria methods available.
2. To familiarize participants with published medical applications of multi-criteria methods.
3. To teach participants how to analyze a medical problem using the Analytic Hierarchy Process (AHP), a widely used multi-criteria method.
4. To discuss current issues regarding the use of multi-criteria methods in medical decision research and collaborative approaches to addressing them.

Format, Requirements, Target Audience:
The course will be divided into three sections. The first will be a brief, didactic overview of multi-criteria methods and previous medical applications. The second will consist of an interactive, hands-on illustration of how to use the AHP to address a medical decision. The third will be a discussion of the strengths and weaknesses of multi-criteria methods and approaches for addressing current research issues. Course materials will include pertinent references, sources of ongoing information about multi-criteria methods, and information about available software tools.

Short Course AM Half Day Sunday, October 18, 2009
9:00am - 12:30pm
SC-AM5 Introduction to Conjoint Analysis of Healthcare Programs
Course size limit: 20 Level: Intermediate
Faculty:
David N. Fisman, MD, MPH, The Research Institute of the Hospital for Sick Children, Toronto, Canada
Victoria Ng, BSc, PhD (cand), The Australian National University, The Research Institute of the Hospital for Sick Children, Toronto, Canada

Background:
When making a decision to utilize a health service or fund a health program, consumers and administrators alike need to consider multiple trade-offs associated with the choice of a given option over a competitor. Such trade-offs may relate to cost, convenience, safety, aesthetic appeal, or any number of other attributes. The "conjoint analysis" technique is an emerging approach to the measurement of preferences in the face of multiple trade-offs in healthcare. Developed largely as a market research tool, but recently adopted by health services researchers, the conceptual basis of conjoint analysis is that individuals' choices of "goods" are influenced by many different characteristics of a given "good". Conjoint analysis provides a methodological framework that allows the relative importance of program attributes to be evaluated through presentation of multiple vignettes in which attributes are varied simultaneously. The impact of this variation on strength of preference is assessed using multivariable regression methods. Incorporation of cost dimensions into vignettes also allows estimation of willingness-to-pay for specific program attributes. Although the apparent simplicity of execution of conjoint analysis is appealing, there are a number of potential pitfalls in study design and execution. This course seeks to provide information (as well as hands-on experience) on optimal conduct of conjoint analysis studies to interested investigators.

Objectives and Course Description:
This course will provide participants with a general overview of the basic concepts that underlie conjoint analysis, of the potential applications of this methodology in health-related decision making, and to practical issues involved in the design and conduct of a conjoint analysis. The course is taught using a mix of didactic lectures and practical, hands-on demonstrations, utilizing a teaching dataset derived from a study of vaccination preferences in healthcare workers. The specific objectives of the course are as follows.
Participants will:

• Gain an understanding of the conceptual framework that surrounds the use of conjoint analysis.
• Develop an appreciation of important considerations in the framing of research questions for conjoint analysis, including the use of focus groups and systematic reviews to identify important program attributes and domains.
• Gain practical experience in the creation of a study instrument for conjoint analysis, including the use of orthogonal planning algorithms to minimize the number of distinct scenarios that need to be incorporated into the instrument.
• Gain experience in the use of relevant analytic methods, including ordinal logistic regression.
• Understand how willingness-to-pay estimates for specific program attributes can be derived using conjoint analysis.
• Become aware of available software packages and statistical tools that can be used for conjoint analysis study design, instrument preparation, and data analysis.

Format, Requirements, Target Audience:
The course will begin with a 1-hour overview of basic concepts and methods. The subsequent two hours will be spent in hands-on problem-based learning exercises, with participants identifying attributes to incorporate into a hypothetical conjoint analysis, utilizing an orthogonal planning approach to minimize the number of scenarios presented to study subjects, and analyzing a dataset derived from a study of vaccination preferences in healthcare workers, in order to gain a nuanced understanding of model building, troubleshooting, and interpretation in the context of conjoint analysis. Prospective participants should have a reasonable level of facility with the use of standard statistical software such as SAS, SPSS, or Stata, including some familiarity with multivariable regression methods.

Short Course AM Half Day Sunday, October 18, 2009
9:00am - 12:30pm
SC-AM6 Methods for Causal Analysis Using Observational (Non-Experimental) Data: Addressing Selection Bias and Other Forms of Endogeneity in Comparative Effectiveness Analysis
Course size limit: none Level: Intermediate
Faculty:
Joseph V. Terza, PhD, University of Florida, Gainesville, FL

Background:
Biases due to self-selection and other forms of endogeneity pervade empirical studies of causal effects using observational data. If left uncorrected, such biases render statistical results from non-experimental data useless for causal inference. This is particularly important in the context of comparative effectiveness research - "the direct comparison of existing health care interventions to determine which work best for which patients and which pose the greatest benefits and harms" (Slutsky and Clancy, American Journal of Medical Quality, 2009). In this Short Course, we will provide a simple but generic regression-based framework for characterizing self-selection and other forms of endogeneity in causal effects models. This framework is comprehensive but practical in that it affords relatively simple estimation methods designed to account for endogeneity in commonly encountered nonlinear regression contexts, e.g., binary response models; count data models; survival-duration models; and two-part models of healthcare utilization and expenditure. Details of endogeneity correction methods for these and other relevant modeling contexts will be presented, along with detailed demonstrations of corresponding easy-to-use Stata 10® software for implementation.

Objectives and Course Description:
The course is designed to:

• Introduce the student to the concept of endogeneity (self-selection) in non-experimental sampling,
• Demonstrate how endogeneity can bias causal interpretation of conventional statistical results in the context of comparative effectiveness analysis,
• Discuss conventional (linear) methods typically used to correct for endogeneity bias (instrumental variables methods) and other forms of confounding (propensity scores),
• Offer real world examples of these conventional methods implemented in Stata 10®,
• Introduce a generic and comprehensive but simple modeling framework for characterizing endogeneity and its consequent biases,
• Discuss newer and more generally applicable (nonlinear) statistical methods for dealing with endogeneity bias
• Offer real world examples of these newer nonlinear methods implemented in Stata 10®
.

Format, Requirements, Target Audience:
The course will follow the standard lecture format with ample time set aside for student questions and comments. The course will be designed for the intermediate student - i.e. researchers having some training and experience with estimation of causal effects via linear and nonlinear parametric regression models.

Short Course AM Half Day Sunday, October 18, 2009
9:00am - 12:30pm
SC-AM7 Microsoft Excel: Advanced Skills for Efficiency and Modeling
Course size limit: none Level: Advanced
Faculty:
Greg Zaric, PhD, Richard Ivey School of Business, University of Western Ontario, Canada
Lauren Cipriano, BSc, BA, PhD student, Stanford University, Stanford, CA
David Hutton, BA, MSc, PhD candidate, Stanford University, Stanford, CA

Background:
Microsoft Excel has become the tool of choice for quantitative analysis in MBA programs because of its power, flexibility, ease of use, transparency, and widespread availability. However, it has not been fully exploited by the decision analysis community. Many decision problems can be analyzed in Excel. Excel can be used to organize and model input parameters, build small or large models, analyze and graphically display results, and perform sensitivity analysis. Simple models can easily be built in Excel and more complicated models can be built through the use of powerful add-ins or through the use of Excel's built-in programming language, VBA. Additionally, several common decision analysis software tools integrate with Excel. In this course we demonstrate how to use Excel to build a number of common types of decision-analytic models. All examples will be done with Excel only and will not require the use of any add-in packages.

Objectives and Course Description:
The goals of this Short Course are to:

1. Present advanced spreadsheet functions that will enable participants to build transparent decision-analytic and Markov models in Excel.
2. Present methods for performing deterministic and probabilistic sensitivity analysis.

Participants will acquire the following skills from this Short Course:

1. Build a simple tree in Excel.
2. Build a cohort semi-Markov model in Excel with time varying probabilities.
3. Use data tables for deterministic sensitivity analysis of spreadsheet models
4. Use uniform, beta, gamma, and normal distribution functions to conduct probabilistic sensitivity analysis.
5. Construct a first-order simulation model in a spreadsheet.

Format, Requirements, Target Audience:
This is a "hands-on" course. Participants will work through structured examples using their own computers. Data sets and files needed for the course will be available for download prior to the conference. Participants should be confident using basic Excel functions including matrix multiplication functions (MMULT, SUMPRODUCT), and database lookup functions (VLOOKUP) and the use of data tables.

Short Course AM Half Day Sunday, October 18, 2009
9:00am - 12:30pm
SC-AM8 How to Respond to an NIH Review
Course size limit: none Level: Beginner
Faculty:
Nananda Col, MD, MPH, MPP, FACP, Maine Medical Center
Steven Fox, MD, SM, MPH, Center for Outcomes and Evidence, AHRQ

Background:
Course faculty will include senior investigators who are experienced in the NIH review process and can guide junior investigators in ways to understand and respond to NIH reviews.

Objectives and Course Description:

• To help prepare trainees and junior faculty to successfully respond to grant reviews from the NIH and AHRQ. Non-profit foundations and professional societies' grants will be discussed, but the focus will be on federal grant reviews.
• To describe the new changes in the NIH and AHRQ review process

The course will cover essential aspects of successfully responding to a grant review, including discussing the new NIH and AHRQ review criteria, the process of peer review and what occurs during a study section, the process of receiving written feedback, and whether to resubmit a grant based on the written feedback. Senior experienced investigators will compare and contrast strategies and approaches to responding to an NIH review, and common mistakes junior investigators make when resubmitting proposals.

Format, Requirements, Target Audience:
Combination of didactic lecture, individual and panel discussion, and case examples. Ample time will be provided for interaction between faculty and course participants. The course is targeted to trainees and junior faculty seeking independent research support.

Short Course AM Half Day Sunday, October 18, 2009
9:00am - 12:30pm
SC-AM9 Psychology of Medical Decision Making
** This course will be taught in two half-day modules which can be taken as a full day course or as two half day courses.
Please see SCFD2 for the course details.

Short Course PM Half Day Sunday, October 18, 2009
2:00pm - 5:30pm
SC-PM1 Becoming a Strong Clinician-Investigator Mentor
Course size limit: 50 Level: Beginner
Faculty:
Joel Tsevat, MD, MPH, University of Cincinnati, Cincinnati, OH
Carol M. Mangione, MD, MSPH, David Geffen School of Medicine at University of California, Los Angeles, CA

Background:
This interactive workshop is geared towards clinician-investigators who are planning or starting to mentor trainees and junior faculty.

Objectives and Course Description:
Topics to be covered include: mentorship basics; benefits (and hazards) of being mentored; expectations of mentors; characteristics of great mentors; expectations of mentees; funding, promotion, and tenure considerations for mentors; number of people one can mentor at a time; co-mentorship; long-distance mentorship; formal mentorship programs; special issues involving sex, race, and age; professional and personal relationships between mentor and trainee; and mentorship skill-building. Real-life case studies of problematic mentor-mentee relationships will be presented and discussed.
After completing this course, participants will be able to:

• Learn the basic requirements for effective mentorship of clinician-investigators
• Learn how to obtain funding to be a mentor
• Become familiar with common situations encountered during mentorship and how to deal with them
• Hone their mentorship skills

Format, Requirements, Target Audience:
The format will consist of presentations by faculty followed by audience discussion. No mentorship experience is required.

Short Course PM Half Day Sunday, October 18, 2009
2:00pm - 5:30pm
SC-PM2 Causal Inference and Causal Diagrams in Medical Decision Making
Course size limit: none Level: Beginner
Faculty:
Uwe Siebert, MD, MPH, MSc, ScD, Institute for Public Health, Medical Decision Making and HTA, University of Health Sciences, Medical Informatics and Technology, Eduard Wallnoefer Center, Hall i.T, Austria

Background:
One of the most important tasks of decision makers is to derive causal interpretations using both statistical analyses of original datasets and decision analysis. Often an intervention, action or risk factor is modeled to have a "causal effect" on one or more model parameters (e.g., probability, rate, or mean of outcome). Therefore, both the biostatistician and the decision analyst need tools to check: (1) when effect estimates have a causal interpretation and when they do not; and (2) the appropriate methods to derive causal effects instead of merely statistical associations.

Objectives and Course Description:
This course will provide an introduction to the principles of causation and causal diagrams, with focus on Directed Acyclic Graphs (DAG) and a brief introduction to methods for causal inference including g-formula, marginal structural models (inverse probability of treatment weighting), and structural nested models (lecture - exercises - discussion). Published cardiovascular, HIV, nutrition and obstetrics examples will be used to:

• Adjust for compliance in randomized clinical trials, where both "intention to treat" and "per protocol" analyses can fail to yield the true causal intervention effect;
• Assess the "fallibility of estimating direct effects" (i.e., adjusting for intermediate steps);
• Adjust for time-independent confounders in observational studies (i.e., confounder affects both risk factor and disease), where standard stratification or regression analysis yield valid causal effects if all confounders are measured, and
• Adjust for time-dependent confounding in observational studies (i.e., the confounder simultaneously acts as an intermediate step in the causal chain between risk factor and disease), where standard regression analysis fails and "causal methods" such as marginal structural models or g-estimation must be used.

Objectives:

• Define causal interventions and actions, draw and interpret causal diagrams, and apply the rules of causal diagrams to distinguish causal from non-causal statistical associations.
• Decide which biostatistical/epidemiological methods must be used in different situations to derive causal effect parameters.
• Use causal diagrams to estimate the direction of bias in "non-causal" models.

Format, Requirements, Target Audience:
The course will consist of lectures, exercises drawn from the published literature and interactive discussion. The intended audience includes researchers from all substance matter fields, statisticians, epidemiologists, and decision analysts interested either in methods of causal analysis or causal interpretation of results based on the underlying method. Requirements: Basic knowledge in epidemiologic methods (confounding).

Short Course PM Half Day Sunday, October 18, 2009
2:00pm - 5:30pm
SC-PM3 Introduction to Discrete-Event Simulation for Healthcare
Course size limit: none Level: Beginner
Faculty:
James Stahl, MGH Institute for Technology Assessment, Boston, MA

Background:
This is a hands-on course that tries to balance theory with practice and will use in-class exercises to teach the building blocks of discreteevent simulation. Participants will also learn the basics of queuing theory and learn how to compare different simulation scenarios.

Objectives and Course Description:
Discrete-event simulation (DES) is a method for modeling systems where competition for resources is important. It has been successfully applied in industrial engineering since the 1960s. Advances in software and hardware have brought this capability to the desktop computer and the decision scientist. In DES, simulation models having approximately the same cause and effect relationships replace real, proposed or conceptual system. Experimentation is carried out through "what-if" experiments, where you can vary the structure of the model, and sensitivity analysis where parameters are varied. Conclusions are about the target system are inferred by studying the behavior of the model under normal and abnormal conditions. DES is most useful when analyzing problems that involve resource constraints or competition for resources, involve closely interdependent events, understanding emergent behavior and to illustrate processes. DES also allows the investigator to explore different system scenarios before committing resources, drill down to understand why a sequence of events occurs, diagnose system level problems, identify constraints and specify requirements for new resources, visualize and can serve as a vehicle to communicate ideas and build consensus among stakeholders. Basic concepts include entities, events, attributes, resources, queues and delays. Basic statistical concepts discussed will include the difference between observational, e.g.- waiting times, flow times, counts, and time-persistent statistics e.g., status of resource, number of entities in system, queue length. The class will describe where DES is best applied and if time permits demonstrate modeling Markov models within the DES framework.

Goals: 1) Understand basic queuing theory 2) Learn basic modeling techniques 3) Learn basic simulation statistics.

Format, Requirements, Target Audience:
Requirements: Participants will require an Intel based computer running Windows 95 or higher. Participants may share laptops. Discrete-event simulation models will be constructed using Arena™. Other software may also be introduced. Demonstration versions of this software will be made available during and before the course. No previous knowledge is necessary though the textbook "Simulation with Arena" by Kelton and Sadowski and "Simulation Modeling and Analysis" by Law and Kelton are recommended.

Short Course PM Half Day Sunday, October 18, 2009
2:00pm - 5:30pm
SC-PM4 Medical Decision Making with Bayesian Networks
Course size limit: none Level: Beginner
Faculty:
Oguzhan Alagoz, PhD, University of Wisconsin, Madison, WI
Elizabeth Burnside, MD, MPH, MS, University of Wisconsin, Madison, WI
Mehmet U. S. Ayvaci, MS, University of Wisconsin, Madison, WI

Background:
Bayesian Network (BN) is a probabilistic graphical model used for data classification, identification of causal relationships, and output prediction. BNs allow a domain expert to model uncertain relationships between a variable of interest with unknown values (e.g., modeling uncertain relationships to predict risk of a disease) and clinical findings/observations (known variables) and are particularly useful for medical diagnosis (e.g., estimating breast cancer risk using mammography findings). Attractive features of BNs include encoding dependencies among all variables, thereby addressing problems with incomplete data; informing causal relationships, thereby increasing understanding about a problem domain and predicting consequences of treatment; combining prior knowledge (which often comes in causal form) and available data; and user friendliness of graphical representations.

Objectives and Course Description:
The course will begin with a general description of BN theory and statistics, demonstrate how BNs differ from decision trees, causal diagrams, and other statistical and data mining techniques such as logistic regression and artificial neural networks. Use of NETICA software for BN construction and its application to various medical decision-making problems will be demonstrated, focusing on breast cancer risk prediction using mammography observations and patient demographic factors. The session will conclude with discussion of limitations and extensions of BNs.

Objectives:

• Understand the theory underlying Bayesian Networks (BNs).
• Understand similarities and differences between BNs and other modeling and analytic techniques.
• Application of BNs to medical decision-making problems.
• Understand BNs' limitations and potential extensions.

Format, Requirements, Target Audience:
This is both a conceptual and a hands-on course. Although a beginner-level course, participants should be comfortable with basic notions of probability. Additional theory such as conditional probability and Bayes' Theorem will be introduced during the course. Participants should bring a Windows/Macintosh-based PC for computational tutorials. A guest license for NETICA and the tutorial examples will be provided to course participants.

Short Course PM Half Day Sunday, October 18, 2009
2:00pm - 5:30pm
SC-PM5 Mathematical Modeling of Infectious Diseases: Beyond the Basics
Course size limit: 20 Level: Intermediate
Faculty:
David N. Fisman, MD, MPH, The Research Institute of the Hospital for Sick Children, Toronto, Canada
Amy L Greer, PhD, The Research Institute of the Hospital for Sick Children, Toronto, Canada

Background:
Control of infectious diseases poses an important challenge to healthcare and public health decision-makers. Traditional models used for the evaluation of cost-effectiveness of disease control interventions have ignored the fundamentally transmissible nature of these diseases, leading to distorted estimates of program cost-effectiveness. However, over the past decade there has been a marked increase in the use of "compartmental transmission models" in the projection of the effectiveness and cost-effectiveness of communicable disease control interventions. While this development is desirable, standard compartmental models assume that transmission is "density dependent" (that is, that disease transmission risk to susceptible individuals is a function of the number of infectious individuals in the population), and reproduce disease patterns that may be seen when large populations are confronted with communicable diseases with epidemic potential. Many communicable diseases of public health importance, including sexually transmitted infections and hospital-acquired infections, are difficult to represent with traditional compartmental models, due to complex patterns of transmission (that are not density dependent), differential efficiency of transmission depending on who is infected, and the potential for stochastic "die out" of infectious processes due to chance. Such processes may be more easily represented using agent-based models that explicitly identify each individual in the population as an unique entity, and which more readily allow representation of population heterogeneity in models. This course will introduce interested investigators, with some background in infectious disease modeling, to the implementation, manipulation, and interpretation of such models.

Objectives and Course Description:
We will provide a hands-on introduction to agent-based modeling of infectious diseases, using sexually transmitted infections and transmission of antimicrobial resistant organisms in an intensive care environment as teaching examples. After an initial introductory lecture that places individual-based models in context, describes their relation to traditional compartmental models, and provides information on the use of such models for health economic analysis, participants will gain insights into the construction, manipulation, and analysis of individual-based infectious disease models by modifying and examining models of Chlamydia transmission in a network of adolescents, and transmission of Clostridium difficile in an intensive care unit environment. These models will be manipulated using the AnyLogic platform, which we are able to provide for workshop purposes. Specific objectives of this course are as follows:
Participants will:

• Gain awareness of communicable disease entities that are best represented using agent-based models.
• Become familiar with the concept of behavioral heterogeneity in transmission of communicable diseases, and appreciate the differential impact on disease control associated with targeting infected individuals at different levels of risk.
• Gain awareness of the use of network structures to represent communicable diseases, and understand the concept of "directed networks". Appreciate the impact on disease dynamics that results from changes in network structure.
• Understand the potential for both diseases and therapeutic interventions to travel along networks (e.g., partner-delivered therapies for sexually transmitted infections).
• Gain familiarity with the implementation and interpretation of stochastic simulations of disease transmission in small populations (e.g., intensive care unit).
• Appreciate the complexities of host-environment interaction in disease propagation in the healthcare environment.

Format, Requirements, Target Audience:
Aside from the initial introductory lecture, the course will be largely hands-on, with instructors assisting students to develop and manipulate pre-prepared models of disease transmission as described above. Students will gain insights into model construction and disease dynamics by changing model parameters, changing the network structures that result in disease transmission, modifying population and pathogen characteristics, and adding simulated interventions to control diseases. Although we assume no prior familiarity with AnyLogic, students should have a basic familiarity with transmission modeling of infectious diseases (as might have been acquired in earlier MDM courses on disease transmission modeling, for example).

Short Course PM Half Day Sunday, October 18, 2009
2:00pm - 5:30pm
SC-PM6 Prediction Models in Medicine: Development, Evaluation and Implementation
Course size limit: none Level: Intermediate
Faculty:
Michael W. Kattan, PhD., Cleveland Clinic, Cleveland, OH
Ewout Steyerberg, PhD., Erasmus MC, Rotterdam, Netherlands
Andrew Vickers, PhD., Memorial Sloan-Kettering Cancer Center, New York, NY

Background:
Our course will cover the importance of prediction models in medical decision making, traditional and novel methods for their evaluation, and cost effective deployment methods. Attendees are expected to leave our course with improved skills in each of these domains.

Objectives and Course Description:
Prediction models are proliferating and gaining increasing acceptance in clinical decision making. Physicians look to us, as analysts, to guide them on which model should be used in the clinic. At the same time that models have proliferated, several different error measures have been introduced. Traditional biostatistical measures include indicators of overall predictive performance, such as explained variation, calibration and discrimination. Novel measures include reclassification and integrated discrimination, as well as some measures explicitly based on decision-analytic principles, such as net benefit and decision curve analysis. The proliferation of measures has made it more difficult to evaluate prediction models, compare rival prediction models, or assess the value of an additional predictor, such as a molecular marker.
The objectives of this course are:

• Increase appreciation and utilization of medical prediction models.
• Learn more insightful ways of analyzing the accuracy of prediction models, both from a traditional biostatistical point of view and a decision analytic point of view. There will be a focus on novel decision-analytic tools.
• Learn a new tool for rapid web-based deployment of prediction models.

Format, Requirements, Target Audience:
Attendees should have experience with regression models, especially logistic and Cox proportional hazards regression models. Familiarity with R or S-PLUS software would also be helpful, but is not required. The course will be a mix of didactics and demonstrations.

Short Course PM Half Day Sunday, October 18, 2009
2:00pm - 5:30pm
SC-PM7 Economic Assessment in Clinical Trial
Course size limit: none Level: Advanced
Faculty:
Henry Glick, PhD, University of Pennsylvania, Philadelphia, PA
Jalpa Doshi, PhD, University of Pennsylvania, Philadelphia, PA

Background:
Prospective economic evaluation of clinical trials is an increasingly important component of the clinical development program for new clinical therapies (e.g., treatments, behavioral interventions, and drugs). The statistical methods used for analysis of economic data from prospective studies are constantly evolving. In this course, the faculty and participants will explore issues in the design and analysis of economic assessments in trials and introduce both standard and recently proposed statistical methods for these assessments. The course format is primarily didactic; its content is both theoretical and applied (with STATA 8.0 computer software documented to assist in use).

Objectives and Course Description:

• Design, implementation, and analysis issues
• Outline the steps in an economic evaluation and provide an understanding of the strategic issues in the design of economic assessments in clinical trials
• Evaluating patient-level medical costs
• Describe and illustrate issues related to choice of univariate and multivariate methods (OLS, log-OLS, GLM) for evaluating and reporting the effect of treatments on costs.
• Evaluating stochastic uncertainty in cost-effectiveness analysis
• Introduce the large number of methods available for reporting on stochastic uncertainty related to the comparison of costs and effects and highlight the preferred methods for confidence interval estimation
• This course is designed for people with some familiarity with statistics and prospective economic data collection in trials.

Format, Requirements, Target Audience:
This course is designed for people with some familiarity with statistics and prospective economic data collection in trials.

Short Course PM Half Day Sunday, October 18, 2009
2:00pm - 5:30pm
SC-PM8 How to Obtain a Career Development Award
Course size limit: none Level: Beginner
Faculty:
J. Sanford Schwartz, MD, The Wharton School of the University of Pennsylvania, Merion Station, PA
Mark Roberts, MD, MPP, University of Pittsburgh, Pittsburgh, PA

Background:
Course faculty will include senior investigators from a variety of MDM-related disciplines who are experienced in the CDA review process and mentoring junior faculty CDA applicants and awardees. Junior investigators who have received a CDA will provide their perspectives on the process.

Objectives and Course Description:

• To help prepare trainees and junior faculty to successfully obtain a peer-reviewed career development award (CDA) and investigator initiated RO-1 awards.
• Address strategies for developing a project and writing an application for CDAs funded by the NIH, AHRQ, Department of Veteran's Affairs, non-profit foundations (e.g., Robert Wood Johnson Foundation; American Cancer Society) and professional societies (e.g., American Society for Clinical Oncology; American Heart Association).

Course Description:
The course will cover essential aspects of successfully competing for a CDA, including development of a research focus; identification of a research mentor and collaborators; formulation and specification of a research question; developing preliminary findings and writing a research plan and proposal, with identification of key points and common errors, examples and discussion of each CDA proposal section; development of an educational plan and skill development. Faculty will highlight key aspects of the CDA process, with an emphasis on writing a competitive proposal and responding to reviewer comments and suggestions. A panel of senior investigators with a range of experience will compare and contrast strategies and approaches to writing a successful CDA /RO-1 application. Specific topics to be highlighted include: introduction and overview of goals; review of various types of CDAs; identifying appropriate awards; and developing one's own research focus; writing a proposal - specific aims, background and significance, preliminary data and findings, research methods, education and training; finding a mentor, writing the application, and dealing with rejection and persevering.

Format, Requirements, Target Audience:
Combination of didactic lecture, individual and panel discussion, and case studies by both experienced and junior faculty. Ample time will be provided for interaction between faculty and course participants. The course is targeted to trainees and junior faculty seeking CDA and/or initial RO-1 support.

Short Course PM Half Day Sunday, October 18, 2009
2:00pm - 5:30pm
SC-PM9 Psychology of Medical Decision Making
** This course will be taught in two half-day modules which can be taken as a full day course or as two half day courses.
Please see SCFD2 for the course details.