Chapter 14
Other Designs and Approaches to Know
7 videos · about 32 minutes · watch as a YouTube playlist →
- 1N-of-1 studiesN-of-1 studies, also called single-case designs, use a single person as their own control to ask what works for that one person. The ABAB design, randomizing the order of phases, a multiple baseline study of fathers’ drinking in Eldoret, Kenya, why history and maturation are weak explanations here, and the strengths and limitations of the design.6 min
- 2Ecological studiesWhen the data exist only for groups, the unit of analysis becomes the group. What makes a study ecological, when it is the right tool, and the ecological fallacy, worked through Robinson’s 1950 analysis of literacy in the 1930 U.S. Census.3 min
- 3Mathematical modeling studiesWhen no experiment can say in time where an outbreak will spread next, mathematical models can. The SEIR model, deterministic, stochastic and agent-based models, and what real-time forecasts of Zika and weekly Ebola forecasts in West Africa got right and wrong, including why the longer forecasts became unreliable.5 min
- 4The synthetic control methodSouth Africa raised cigarette taxes sharply from 1994, but consumption was already falling. How a synthetic control builds a comparison from a weighted combination of untreated countries, checking the estimate with pre-intervention fit, leave-one-out and in-space placebos, and when to use the method, including the mortality impact of U.S. global health investments.6 min
- 5Economic evaluationsOnce an intervention works, the next question is whether it is a good use of resources. Cost analysis, cost-effectiveness, cost-utility and cost-benefit analysis, DALYs and QALYs for comparing interventions across diseases, and the incremental cost-effectiveness ratio.1 min
- 6Diagnostic accuracy studiesHow we judge a diagnostic test: sensitivity, specificity and predictive values from the same two-by-two table, why predictive values depend on prevalence, where to set a threshold, and how to design an accuracy study that gives trustworthy numbers, worked through Xpert MTB/RIF against smear microscopy and culture for tuberculosis.9 min
- 7Delphi studiesThe Delphi method turns expertise spread across continents into consensus. How a Delphi study runs in rounds, why anonymity and controlled feedback matter, and when it is a strong choice for setting priorities, indicators, case definitions and guidelines.2 min