This course provides a foundation in the main components of alternative approaches to individual randomized controlled trials for evaluating interventions in real-world settings. An overview of the history of experimental and observational design sets the stage to understand design variants. For each randomized (cluster-randomized and stepped-wedge randomized trials) and quasi-experimental design (QED) presented, scholars assess the key features, common pitfalls, and possible strategies to improve internal and external validity. QEDs covered in this class include pre-post designs and interrupted time-series designs. Scholars are also introduced to the implications of design decisions for analytic approaches to data analysis, and how to prepare to meet with a biostatistician to take design ideas to the next level. Scholars are challenged to determine which design is best suited to a range of 'real-world' implementation settings and circumstances, and to choose between designs to maximize overall study quality.
In addition to the core study design elements described above, at the start of the course, scholars may select to participate in a complementary Analysis Lab. Each week, learners enrolled in the Analysis Lab apply a given study design to compute sample sizes and analyze data, generating and interpreting results. The lab component is optional. Students who previously opted into the lab may opt out at any time.
Note for UCSF Graduate Division Students: The core course is worth 2 units, and the optional lab is worth 1 unit. The lab may be completed only as a supplement to the core course. It is the student's responsibility to manage enrollment with the registrar's office (e.g., enrollment in 2 vs. 3 units).
At the end of the course, scholars will be able to:
- Describe the key characteristics of - and rationale for choosing among - common non-randomized quasi-experimental study designs (e.g., pre-post, interrupted time series) and randomized study designs (e.g., pragmatic, cluster, stepped wedge, factorial, MOST, SMART, and choice/preference) used in real-world implementation research.
- Identify key threats to internal validity across study designs, including confounding, mediation, effect modification, selection bias, information and reporting bias, and random error, and propose design or analytic strategies to mitigate each threat.
- Design hybrid implementation-effectiveness studies and define appropriate implementation and effectiveness outcomes for each study design.
- Create clear visual diagrams for each study design for inclusion in protocols, presentations, and grant proposals.
- Develop a plan for discussing sampling strategy, sample size determination, and analytic approach with a biostatistician for each study design.
Optional Analysis Lab, in addition to the objectives above:
- Identify appropriate analytic methods for each study design covered in the course. Methods reviewed include regression, segmented regression, autoregressive integrated moving-average (ARIMA) models, generalized estimating equations, and mixed-effects models.
- Calculate and interpret sample size estimates for each study design using STATA software.
- Conduct and interpret analyses in STATA using example datasets and code provided for each study design.
Audience
Clinicians, public health practitioners, and researchers wishing to gain knowledge and skills in translating evidence into practice.
Offered: Spring Term
Faculty
Course Directors
Starley B. Shade, PhD, MPH, is an Associate Professor in Epidemiology and Biostatistics at UCSF. Her research focuses on quantitative and economic evaluation of community- and clinic-level interventions to improve health outcomes for those with HIV, TB and Malaria. Her current projects include cluster-randomized trials and adaptive study designs in Kenya and Uganda, as well as the evaluation of demonstration interventions to improve engagement in HIV care among people seen in publicly funded programs in the U.S.
Joelle Brown, MPH, PhD, is an epidemiologist and Associate Professor in the Department of Epidemiology and Biostatistics and the Department of Obstetrics, Gynecology, and Reproductive Sciences at the University of California, San Francisco. She has over 20 years of experience conducting health research in sub-Saharan Africa. Her research and expertise include reproductive health and the prevention of sexually transmitted infections, including HIV, clinical trials, implementation science, and safer conception strategies for women and couples living with HIV.