Reporting Workflow Revision - Brief Proposal
1. Proposed Workitem: Reporting & Processing Workflow Update
- Proposal Editor: Kevin O'Donnell
- Editor: Kevin O'Donnell?
- Domain: Radiology (& Cardiology?)
2. The Problem
Radiology workflows are getting more distributed; data collection, processing & reporting are more commonly spread across multiple locations, even multiple organizations.
Radiology workflows are getting more complex; more steps such as clinical processing, 3D reconstruction, and more combinations and permutations.
Radiology workflow logic is getting more sophisticated; there are examples of push, pull, centralized, distributed, data-driven, event-driven, etc.
This makes it harder for people to know who is supposed to do what next or find out what has been done. Worklists and performed procedure steps would help, but GP Worklist has proven to be difficult to implement and does not address some of the logic patterns.
Costs from manual workflow likely include slower turnaround (a key metric for radiology), extra manpower for the manual tasks, lost/dropped actions, poor resource balancing, missing data, poor coordination, poor collection of performance metrics, poor tracability, etc.
Proprietary workflows sometimes address some of the above, but at a cost in flexibility and limiting choice/ability to use preferred systems.
3. Key Use Case
An imaging procedure is ordered which should be followed by running a clinical analysis package on the images, performing a CAD analysis and generating a 3D visualization. All four datasets should be considered in the interpretation and report.
- the three post-processing steps are performed by different people on different workstations
- there is more than one protocol that could be run in each step and they need to know which
- the techs need to know when the input data is available for them to start their work
- they should not have to go manually retrieve their inputs
- the output data should be identified and made available to the next step
- the radiologist may not know whether to expect all 4 of the datasets or only some of them
- the radiologist does not know when each is complete or when the data is available
- when there is a failure, there should be enough information for a management system to be alerted/intervene
- there should be information in the audit trails to give some idea of how the department is performing and where the bottlenecks, if any, are
Among other things, we would like to be able to:
- request reading or consultation by another facility
- provide references and access paths for images to read, images for priors
- precisely describe the requested action
- possibly indicate who you would like to do it
- provide avenue for requester to monitor progress and get pointers to results
4. Standards & Systems
Systems include RIS, PACS, modalities, post-processing, reporting, HIS, EMR
Standards include DICOM Unified Procedure Step (Sup 96)
This is basically about updating the existing Reporting and PostProcessing Profiles (which seem to have had very limited traction), to be easier to implement and use, in part through using UPS.
- Radiotherapy implemented Sup 74 based on Sup 96 and successfully tested at two Connectathons
- currently in the process of moving from Frozen Draft to Final Text
- due to changes/improvements during review will be re-balloted in October
- retains many of the attributes used in GP-WL for easier transition
- much simpler object management logic
- supports push workflow, pull workflow, self-scheduling, etc.
- provides a subscription model for monitoring procedure steps
- improves referencing of input/output data, handling local network, media or XDS retrieval
- SIIM is working on mapping/modeling radiology workflow steps and patterns (first draft completed)
- good potential for collaboration on defining needs, designing solutions and promoting the result
- Should loosen the linkage to SINR as the output.
- Is this a clean compliment to the Reporting Templates?
- Dovetail fairly easily. Don't have to be joined at the hip.
- Should we address MWL?
- (While UPS would work, shouldn't mandate replacing MWL, but a useful compliment - perhaps add as an option to SWF)
- we should confirm the need/business case does exist for these profiles. They may not have failed due to technical issues.
- Feeling is that the business case is sound. There is a need.
- Were there other inhibitors than the complexity of the GP implementation
- Did people implement GPWL and balk at using SR for the output?
- we should be careful not to stray into trying to design universal workflow management. Even while limiting to Radiology, we should still be aware of general solutions (e.g. BPEL, etc).
- do we need a RESTful interface to the UPS Service?
- Given that it is DICOM-based, how do we keep the EMR in the loop?
<This is the brief proposal. Try to keep it to 1 or at most 2 pages>