Agile Science: Generic
The generic agile science process is conceptualized as appropriate for a variety of health interventions including pharmaceuticals, medical devices, behavioral interventions, policies, and combination approaches. We also view this as appropriate for different targets such as supporting individuals, clinics and hospitals, and even cities and beyond. Finally, we are also examining how agile science processes can be enacted by and for patient-scientists who are working to solve their own problems and patients like them.
For these other intervention types, targets, and different types of researchers we are supporting, including patients themselves, we are working developing concrete structures that are specified as well as we have done so for behavior change (see below and the resources page). For up-to-date information about our work extending agile science behavior behavior change, please contact us!
For these other intervention types, targets, and different types of researchers we are supporting, including patients themselves, we are working developing concrete structures that are specified as well as we have done so for behavior change (see below and the resources page). For up-to-date information about our work extending agile science behavior behavior change, please contact us!
Agile Science: Behavior Change
We started to develop agile science for behavioral interventions. As such, we have the most concrete specification for supporting behavioral interventions.
The agile science process for behavior change focuses on creating useful and usable behavior change interventions and corresponding usable evidence for making decisions. It borrows from agile software development, user-centered design, lean start-up, and data science and modeling techniques for simulating complexity.
The process starts with creating many variations of plausibly useful behavior change interventions for a “niche,” i.e. specified people, places, and times. This is followed by optimizing those behavior change interventions for that targeted niche.
Optimization tests whether interventions produce the desired real-world success, with definitions of success and failure called optimization criteria. The methods of this phase are inspired by Linda Collins et al's MultiPhase Optimization Strategy and other optimization methods. If interventions are useful for a given niche, they are repurposed for others who might benefit from them.
If behavior change intervention components are useful for a given niche, they are repurposed for other niches. This occurs via:
The agile science process for behavior change focuses on creating useful and usable behavior change interventions and corresponding usable evidence for making decisions. It borrows from agile software development, user-centered design, lean start-up, and data science and modeling techniques for simulating complexity.
The process starts with creating many variations of plausibly useful behavior change interventions for a “niche,” i.e. specified people, places, and times. This is followed by optimizing those behavior change interventions for that targeted niche.
Optimization tests whether interventions produce the desired real-world success, with definitions of success and failure called optimization criteria. The methods of this phase are inspired by Linda Collins et al's MultiPhase Optimization Strategy and other optimization methods. If interventions are useful for a given niche, they are repurposed for others who might benefit from them.
If behavior change intervention components are useful for a given niche, they are repurposed for other niches. This occurs via:
- modularizing an intervention to its smallest, meaningful, and self-contained element,
- engaging in a science of matchmaking that systematically studies the decision policies used to match interventions with other people, places, and times (what we call a "niche").
- within a clinical context, taking advantage of implementation science methods, such as pragmatic clinical trials, to iteratively improve upon processes within real-world contexts.
