## human behavior - rationality - compliance - media effect e.g. misinformation ## development and supply - ## 1. characterisitc of health care - there exist individual and population level, depending on aggregation level (transfer learning, pattern sharing) - physical + behavioral factor (service provider and receiver, behvioral factor to reported death rate during covid) - data: mobile data ([onnela lab](https://www.hsph.harvard.edu/onnela-lab/research/), digital phenotyping) and J-clinic mammograph (recalibration) - methodology: - hierarchical: recalibration to different hospitals and age group - dynamic: behavior, adjustment, perception, - parameterization decision tree [[Pasted image 20230124032958.png]] three types of independent Binomial sampling. Cross-sectional sampling involves sampling with the goal of making inferences about the proportions of individuals in distinct populations that exhibit a particular characteristic at a particular point in time. For example, suppose there are two types of medical device that can be used for a specific procedure. An experiment is performed to compare the proportions of “successful” procedures under the two types of device. The second type is called prospective cohort sampling. In general, this type involves following one or more groups of individuals through time with the goal of relating an outcome/endpoint, like death from some particular cause, acquisition of a particular disease, or acquisition of a particular infection, etc., to a collection of potential risk factors. In the simplest case, interest would focus on assessing whether the two levels of a binary “exposure” variable were associated with different levels of risk corresponding to the particular outcome/endpoint under study.