Missing data present a common problem for prevention research and improperly handling missing data can severely compromise the validity of a study’s inferences. In this Methods: Mind the Gap webinar, Dr. Little highlights the power and utility of modern principled treatments for missing data to optimize inferences. He discusses the three issues that occur when data go missing and the three mechanisms that give rise to missing data.
The problem of overdiagnosis, the detection by screening of latent cancers that would never have surfaced, has been much in the news lately. What is overdiagnosis, and how significant is the problem? In this Methods: Mind the Gap webinar, Dr. Etzioni examines how overdiagnosis arises and discusses what it takes to validly estimate its frequency.
This course discusses machine learning (ML), which has become a core technology underlying many modern applications, especially in health care. Machine learning techniques provide powerful methods for analyzing large datasets such as medical images, electronic health records, and genomics. Recent advances in deep learning (DL) provide an analysis framework that can be used to automatically classify images and objects with (and occasionally exceeding) human-level accuracy.
The goal of this FAES course is to introduce biomedical research scientists to R as an analysis platform rather than a programming language. Throughout the course, emphasis is placed on example-driven learning. Topics include: installation of R and R packages; command line R; R data types; loading data in R; manipulating data; exploring data through visualization; statistical tests; correcting for multiple comparisons; building models; and generating publication-quality graphics. No prior programming experience is required.
This is a free, 7-part, self-paced online course with instructional slide sets, readings, and guided activities.
- Part 1 provides an introduction and overview of the three kinds of randomized trials and their distinguishing characteristics.
- Part 2 considers the design of group-randomized trials (GRTs).
- Part 3 discusses analytical approaches to GRTs and individually randomized group-treatment trials (IRGTs).
- Part 4 explores important power and sample size considerations for GRTs.
- Part 5 provides examples of GRTs from the Health Care Systems Collaboratory, a project funded by the NIH.
- Part 6 reviews recent practices in GRTs and IRGTs based on literature reviews.
- Part 7 examines alternative designs to evaluate interventions in comparison to GRTs and IRGTs.
This training walks participants through writing programs that would help them solve scientific problems. During the course, participants get a brief introduction to the programming concepts, followed by hands-on walkthrough in writing scripts using the Unix Shell, R, Perl, and Python. The course will cover reading the data through processing and saving the processed data.
This course is designed for researchers, clinicians, students, and academics who are aspiring to learn to write their own scripts and programs, or for scientists in program administration interested in learning the standard tools and techniques for biomedical programming.
In this Methods: Mind the Gap webinar, Dr. Linda Collins discusses why behavioral interventions are important in many areas of public health, for example, smoking cessation, drug abuse prevention, treatment of obesity, management of heart failure symptoms, and promotion of physical activity.
Behavioral interventions are typically developed and evaluated using a treatment package approach, in which the intervention is assembled a priori and evaluated by means of a randomized controlled trial (RCT). Dr. Collins reviews an alternative approach called the Multiphase Optimization Strategy (MOST), an engineering-inspired framework for developing, optimizing, and evaluating behavioral interventions. MOST includes the RCT, as well as other empirical steps aimed at intervention optimization.
In this Methods: Mind the Gap webinar, Dr. Jennifer Croswell demonstrates methods to critically assess the quality of published systematic reviews of clinical or public health interventions.
In this Methods: Mind the Gap webinar, Dr. Monica Taljaard explains the unique characteristics of the stepped wedge cluster randomized design and its implications for sample size calculation and analysis, and discusses its strengths and weaknesses compared to traditional designs. Emphasis is on application, with examples in disease prevention and health promotion research.
In this Methods: Mind the Gap webinar, Dr. William Shadish reviews illustrative studies that demonstrate the direction such work is taking and the results that seem to be emerging in regard to nonrandomized control group designs, regression discontinuity designs, and interrupted time series designs.
In this Methods: Mind the Gap webinar, Dr. Walsh presents preventive strategies that integrate clinical data science, informatics, and mental health expertise in an attempt to prevent suicidal thoughts and behaviors. He explains basic concepts in applied predictive modeling relevant to an audience interested in disease prevention. He also shares examples of active research and operational efforts in this domain in civilian and active duty military environments.
The objective of this course is to provide a thorough grounding in the conduct of randomized clinical trials to researchers and health professionals interested in developing competence in the planning, design, and execution of randomized clinical trials involving behavioral interventions.
The curriculum will enable participants to:
- Describe the principles underlying the conduct of unbiased clinical trials
- Identify the unique challenges posed by behavioral randomized clinical trials (RCTs)
- Evaluate RCT designs in terms of their appropriateness to scientific and clinical goals
- Select appropriate strategies for enrollment, randomization, and retention of participants
- Understand methods for monitoring, coordinating, and conducting RCTs
- Develop strategies for appropriate statistical analyses of RCT data
- Evaluate the quality of behavioral RCTs and interpret their results
- Design an RCT as part of a working group on a specific topic.
Dr. Sterman discusses systems approaches in public health, including the concepts of policy resistance, implementation feedbacks, and model boundaries and explores how these ideas can be applied to effect change in a complex system. He includes examples from healthcare and public health such as implementation of formulary drug lists and SARS epidemic modeling.
Dr. McLeroy discusses adoption of systems methodology, including multiple levels of analysis, utility for identifying points of change, testing models against reality, and applications to program evaluation and various research designs, including community-based participatory research and randomized clinical trials.
Systems Network Analysis: Using Connections and Structures to Understand and Change Health Behaviors
Dr. Faust presents a non-technical overview of methods used to analyze networks, with an emphasis on social networks. Topics include: formal representations of social networks (graphs and sociomatrices), social network data considerations, and methods for analyzing social networks (connectivity, centrality, cohesive subgroups, equivalences and blockmodels, subgraphs, and structural hypotheses).
Dr. Valente describes methods for using network analysis to elucidate the antecedents and consequences of health-related behaviors. To do this, he draws from a number of examples of his applied work in the areas of substance abuse prevention and treatment, contraceptive choices, and community coalitions, among others. He also describes how applied research utilizing network analysis methods can be used to stimulate improvement in individual, community, and organizational behavior change programs.
This series of online lectures covers a range of diverse topics in data science such as data management, data representation, computing, data modeling, and other overarching topics. This series is an introductory overview that assumes no prior knowledge or understanding of data science.
In this Methods: Mind the Gap presentation, Dr. Selvin discusses the importance of epidemiologic evidence in informing strategies and cut points for screening and diagnosis of diabetes. A focus is on the evidence supporting the importance of the hemoglobin A1c (HbA1c) test and current controversies regarding screening and diagnosis of prediabetes.
Johns Hopkins Bloomberg School of Public Health
In this Methods: Mind the Gap webinar, Dr. Kay Dickersin reviews models of how systematic reviews are being used globally to plan, implement, and derive recommendations from comparative effectiveness research (CER). She also reviews some of the existing challenges to using systematic reviews and methods being used to address these challenges.
Time-varying effect modeling (TVEM) is a novel method that enables health, behavioral, and social scientists to examine developmental (i.e., age-varying) and dynamic (i.e., time-varying) associations. In this Methods: Mind the Gap webinar, Dr. Stephanie Lanza discusses potential research questions that can be addressed using TVEM, and provides resources for researchers interested in using the models in their own work.
The series provides an overview of analytic approaches, methods, and statistical applications for analyzing tobacco regulatory science (TRS) data. The presenters include an esteemed group of scientists, well known for their work in methodological research dealing with casual inference. The webinars are intended for any investigator funded by the Center for Tobacco Products.
The National Cancer Institute is hosting this training institute to provide participants with a thorough grounding in conducting D&I research with a specific focus on cancer, across the cancer control continuum. In 2020, the institute will use a combination of online coursework (six modules with related assignments) and a 2-day in-person training to be held August 3 and 4, 2020, at the NCI campus in Bethesda, MD. Faculty and guest lecturers consist of leading experts in D&I theories, models, and frameworks; intervention fidelity and adaptation; stakeholder engagement and partnership for D&I; research methods and study designs for D&I; and measures and outcomes for D&I. This training institute has been adapted from the broader Training Institute for Dissemination and Implementation Research in Health (TIDIRH), organized by NIH and the VA over the past nine years.
This training is designed for investigators at any career stage interested in conducting D&I research with a focus on the cancer control continuum. There is no cost associated with the training. Invited participants are required to cover related travel expenses to the Washington D.C. area for the in-person meeting. More answers to common questions can be found on the site FAQ.