Consider adding a casestudy entry to service_index

[johannehouweling]

Currently, there is no casestudy entry in service_index.

Strictly speaking, it might not be necessary, as the regulatory question defines the case study. However, this relationship is now being used in search functionality, and other indexes (e.g., methods/data) do include a case study entry.

It might be worth considering adding a casestudy entry explicitly to the data model for consistency and clarity, even if regulatory questions already express this relationship.

The current mapping can already be used to go from Reg. Q -> Case study

Case study	Reg Q fields
Kidney	reg_q_1a OR reg_q_1b
Parkinson	reg_q_2a OR reg_q_2b
Thyroid	reg_q_3a OR reg_q_3b

---

Pros of adding casestudy explicitly to service_index:

• Simpler filtering/querying: search and UI logic can filter directly on casestudy without translating from reg_q_*.
• Consistency with other indexes: methods/data already expose this relation, so the model becomes more uniform.
• Clearer semantics: casestudy is easier to understand than inferring it from regulatory-question fields.
• Less application-side logic: avoids hardcoded mapping logic in search.
• Probably stable enough: if regulatory questions are very specific and unlikely to change, duplication risk is lower.

Cons of adding casestudy explicitly:

• Data duplication: the same relation exists implicitly via regulatory-question fields and explicitly via casestudy.
• Risk of misalignment: if one side changes and the other is not updated, data can become inconsistent.
• More maintenance: requires validation or conventions to keep both representations aligned.
• Potential model ambiguity: it may become unclear which field is the source of truth.
• Strictly not necessary: if reg_q_* already fully determines the case study, this adds redundancy rather than new information.

Reasonable conclusion
Adding casestudy likely makes sense for usability and consistency, especially if:

• you already depend on this relation in search, and
• the regulatory-question mapping is relatively stable over time.