# The Challenges of Automating Case Management [[Case Management]] Source [[Case Management Combining Knowledge With Process]] The article talks about technology supporting case managements. However as pointed out in [[Case Management#^3c87c4]], case's often doesn't have a clear outcome/end goal, thus it's hard it's sequence/design. Related to [[HOPES Project Index]], because in the building in a algorithm to support case managers when an early relapse signs are detected. The cases that clinical case managers faces are often unpredictable. The characteristics of Case Management we outlined earlier indicate where the challenges lie when trying to more fully automate this style of work. As with the automation of any business processes, the technology has to support “exception handling, collaboration, decision making, unstructured information, negotiations and paper flows”xviii. But the fundamental challenge in automating case management is using the technology to support the unpredictable ways cases progress and people work in practice. Traditional automation assumes that a sequence or pattern can be determined in advance by careful initial analysis and catered for by good design. More formally, it assumes that the logical flows can be understood a priori. As noted by van der Aalst et alxix, in a traditional approach the designer has to specify what is permitted. Any routing that is not specified at design time will not be supported by the system at runtime. However, when seeking to automate cases, there is no predetermined sequence, and new tasks and processes can be added at any point during the life cycle of the case as the need for them arises. The emphasis must be on supporting the ad hoc nature of cases. As Lucy Suchman puts it “The trick .... is to introduce bits of automation that will fit in to the work and do useful things, and then make it possible for people to work with those bits of automation embedded in the systems while leaving them the discretionary space to exercise the kind of judgment they need to exercise to really get the work done”xx ## The challenges specific to automating Case Management include: 1. Striking a balance between Practice and Procedure: 1. Almost any job today has both clearly defined, predictable elements and less well defined, more ambiguous aspects where workers exercise their judgment. Different types of job, and the activities within a particular job, can be thought of as a spectrum running from ”well defined procedure” to loosely defined ”discretionary practice”. Someone working in a call center generally doesn’t exercise great discretion in how they carry out their job, while a senior investment analyst probably does.![[Screenshot 2021-11-13 at 11.57.15 AM.png]] 2. There is a danger of aiming to fully proceduralize processes when reengineering knowledge work, ignoring the ”Practice” side of the spectrum. However, this is counter productive, because when a case occurs that doesn’t match the rules prescribed in the procedure, workers are either brought to a halt or forced to create an unofficial workaround.xxiii A careful balance must be maintained between prescribing defined procedures for fully understood and repeatable aspects of work, while respecting the parts of work that should be left to the discretion of knowledge workers. Every effort to change how work is done needs a dose of both process – the design for how work is to be done – and practice, an understanding of how individual workers respond to the real world of work and accomplish their assigned tasks.xxiv 3. It should be a goal that as particular workers become more experienced, they will be able to formalized and standardize aspects of what they do, and these standardized procedures can then be made available for use by colleagues.xxv 2. Capturing Implicit Rules and Tacit Knowledge: 1. This challenge is related to the point made about Practice versus Procedure. Many case management processes will never have been previously automated. They rely on paper forms and tacit and implicit rules governing how cases should be managed in addition to documented and explicit policies and procedures. Examples of implicit rules will include how staff cope with particular exceptions, how they make decisions at particular stages of a case and how they deal with unstructured information. The challenge here is to discover these implicit rules and tacit knowledge and, where appropriate, try to support their automation, while leaving room for those steps and decisions that should continue to depend on individual discretion. 3. Formalizing Experience 1. Supporting Learning: A good Case Management solution should help an organization learn from previous cases. This learning could be exhibited in the definition of new processes, new procedures, better online help etc., where lessons learned by knowledge workers during a previous case were quickly applied to process definitions to improve them. We use the phrase ”formalizing experience” to describe this process of changing a practice into an automated step where appropriate, or supporting some other action that will assist in the processing of future cases. There is a related area of research, called Case Based Reasoning (CBR), which has some obvious applicability to Case Management. Case Based Reasoning is the process of solving new problems based on the solution of similar past problems: it’s based on two tenets, “similar problems have similar solutions. Consequently, solutions for similar prior problems are a useful starting point for new problem-solving” and “the types of problems an agent encounters tend to recur. Consequently, future problems are likely to be similar to current problems”xxvi. CBR consists of four steps – look at a given case or problem and try to think of a previous case or cases it might match; consider how the solution to the previous case could be applied to the current case; test the solution on the existing case and revise it if necessary; then record the new case, and the solution that eventually worked, for future reference. 4. Supporting Ad Hoc Change 1. It is not possible to fully analyze and define at design time how a case will actually unfold at ”runtime”. While the overall pattern of a typical case may be known, any case management system must allow for entirely new and unpredictable process paths being required at execution time. This is due to the nature of cases, where a range of outcomes may arise at each stage in the case, unpredictable at design time, and these outcomes then determine the next stage or stages. Also because cases are long-running, change may be introduced into the policies governing how cases are handled while previous cases are being processed. For example, mid way through the execution of an application for social welfare benefits, an entirely new set of steps may be initiated to comply with new procedural guidelines, or to cope with an unplanned external event such as the completion of a related case. 5. Involving Participants in the Design of Knowledge Processes: 1. the challenge here is to let knowledge workers influence the design of those processes they participate in, helping them make changes to processes or innovating new ones. Any solution needs to support fast, easy change initiated by a worker. 6. Supporting Collaboration : 1. Collaboration is a key requirement, but it is not simply a matter of enabling instant messaging or document sharing. Case workers need to share everything related to a case, including history, discussions, correspondence and previous decisions. The collaboration support for case management must ensure that the correct information is made available to team members at the correct time, without losing the context or current state of progress of the case. This requires a ”smart” system that knows who needs what when. Conversely, it is important that irrelevant information is not provided at the wrong time, and that confidential information is not made available to inappropriate recipients. 7. Supporting Decisions: 1. Case workers are the key decision makers in determining how a case will progress, supported where appropriate by automated rules. Automation of cases must recognize that control will continue to reside with human case participants, rather than seeking to encapsulate everything in an increasingly complex rule-base. 8. Effectively Coordinating Participants: 1. related to the point above, effective case management requires that work is routed to participants at the appropriate time and in the appropriate sequence, given the history of the case to date. This coordination requires sophisticated workflow routing, synchronization of process flows at various points, ensuring overall milestones are monitored and met, and ensuring that delays are identified and exceptions raised where necessary. 9. Managing Complexity: 1. Information and data has to be organized and presented to all case workers in a useful way so they do not become overwhelmed or confused by the various pieces of documentation, records and notes related to a case. The interface used by case workers to interact with automating systems is a key determinant of the success of any automation. 10. Managing Artifacts 1. Beyond the presentation of case information to the user, there is a need to effectively store, manage and retrieve information related to a case. Case history and associated records may need to be retained for specific periods, as a result of legislation or organizational policy. Content may be structured or unstructured, and may reside on multiple supporting systems such as databases, content management systems and electronic record management systems. Any case management solution must manage this content efficiently and effectively. 11. Integrating Disparate Systems 1. While key aspects of case management are poorly automated, there are almost always some important legacy systems in use at an organization that will be part of any solution. Effective case management requires the smooth integration of these existing systems into any future solution.