Quarto reporting

[azimov]

The use of shiny apps is helpful for the general purpose exploration of data.
However, this is often time consuming and difficult to share results.

To that end, any Cohort definition should be able to turn into a standard set of static reporting tools with 2 modalities:

• Simple boilerplate reports that give a broad overview of a Phenotype based on a cohort definition
• An API that gives access to results in a consistent manner that allows data to be exposed and plots/tables to be made in ways that users can customize
  (for example, you may only wish to look at certain time windows or domains related to a cohort). Doing this in a general purpose way would be both restrictive to users and overly complex for developers, with every possible combination requiring a considerable amount of time.

With this in mind, this issue is intended to be a broad goal of ideas that will be split up as and when the initial modalities of usage are identified.

Boilerplate report

My suggestion for inital top line reports is here.

For a given cohort, the following top line stats should be available in a boilerplate manner

• Cohort counts
• Cohort inclusion rules
• Index event break down
• Visit context
• Time distributions
• calendar year based incidence
• Table 1 demographic features

Customizable elements

• Cohort overlap
• Custom cohort features
• Cohort comparison (cross database or cross cohort pairs)

Note the exclusion of orphan concepts here as this is probably now best done in Atlas.