Case Managers spend their time managing Cases and leveling and balancing resources across Cases. 
Popular examples of Cases are Patients in healthcare, Investigations in law enforcement but, in general, a Case is just a cursor position at a post-relational database management system. So, we can have “Cases” where the focus is on a supplier, a customer, an insurance claim, a helicopter receiving periodic maintenance.
The methodology of choice for Case Management is ACM (Adaptive Case Manager).
ACM recognizes that a Case ends up as a mix of structured protocols (i.e. process fragments) plus ad hoc interventions. Each Case is typically unique. Each has goals and objectives and their presence is essential to closing any Case.
If you are planning on developing a Case Management software suite or planning to acquire one, here is a list of fifteen (15) “must-have” features:
- Official interfaces only (normal users, casual users, import/export engine)
- Case Hx (longitudinal view)
- Case Hx (workflow view)
- Case Goals/Objectives
- FOMM for assessing progress toward Case Goals/Objectives
- Post-relational dbms
- Menu of Services (for selecting “best practices” protocols)
- User Workspace (InTray, capable of hosting “best practices” protocol template instances)
- Background orchestration from BPM process template instances
- 3-Tier Scheduling (system, users, supervisors)
- Skip/Jump at Process Steps
- Insert an ad hoc intervention at a Case
- Break Glass (for emergency takeover of an in-progress intervention)
- Re-assign/take back a process step that is not being worked on
- Advanced Security (who can do what, when)
Whether you decide to build or buy, don’t drop any of the above items from your shopping list without careful consideration of the consequences.
Feel free to call Civerex at 800 529 5355 if you have questions.
KWKeirstead's Blog 
There is not that much difference between a process fragment (linked set of steps) and an ad hoc intervention at a Case. The way to avoid making a special “case” of “ad hoc intervention” is to say that each is a “process of one step”.
We can debate must-haves at process steps – the basics are that each step to be performed by a human user needs to have a name, at least one attached form (for data display and much of the time for data capture), plus a routing (skill designation).
LikeLike
As usual, it’s not what IS there that is the problem, it’s what IS NOT there that is the problem.
Two remedies
1. Donald Rumsfelds “known knowns, known unknowns, unknown knowns, unknown unknowns”
2. 3D Kbase searches where you get to see what matched and what did not match.
LikeLike
It’s not obvious that a post-relational DBMS is a must-have. Please explain why you think so.
LikeLike
What this has meant to me since we started using them in 1990 is that you can, at a cursor position, accommodate: structured data, free-form (i.e. memo) data, .pdf, .doc, images, videos AT the cursor position, in a Binary Large Object (BLOB) field.
This greatly simplifies security (i.e. if the user does not have access to the database record, then they cannot get to any of the “embedded” objects).
It is essential at Cases to be able to accommodate a mix of digital objects – a good way for users to look at a Case is as a “bucket” or “container”
Hope this helps – If not, I would be pleased to phone you and hopefully provide a better explanation.
LikeLike