Error component analysis for PACS: operational sources of data error in real world PACS for DICOM Series, Study, and Patient level identifiers

摘要

The gold standard for modern PACS workflow is now considered to be a complete integration of HIS, RIS and PACS datastreams. However, a number of current obstacles exist in regard to achieving this level of integration in real-world practice environments, given the current level of modality support for more advanced DICOM services such as MWL, MPPS, and Storage Commitment. In this paper, many of the more common sources of disparity between HIS/RIS and PACS data are discussed, as well as possible DICOM-based mitigations and possible PACS workflow models not requiring modality upgrades. These are discussed in the context of the IHE model, and compared to the ideal of complete IHE workflow. Various operational sources of modality-based patient demographic data for existing PACS are discussed and analyzed from a DICOM perspective, comparing modality inputs for a typical large PACS to RIS data references. Disparities between PACS and RIS data are emphasized, with the goal of PACS/RIS data reconciliation. The existing obstacles to this integration are discussed, and the application of DICOM Modality Worklist, Storage Commitment, and Performed Procedure Step Services are discussed in reference to these error sources. As the subject site utilizes DICOM Modality Worklist, those operational sources of error that remain after Modality Worklist was implemented are analyzed in the context of the existing modality workflow limitations. The intrinsic disparity between CPT-based billing and reporting structures and existing DICOM modality models for CT, MR, US, R&F, and CR are discussed, and examples given for operational challenges in integrating these elements using existing modality design. Conclusions included that the Modality Worklist, in itself, is inadequate to drive RIS-integrated PACS workflow, and that current modality limitations preclude complete PACS/RIS datastream integration at this time. Several alternative mitigation models are discussed using existing modalities, as well as suggestions for improvements in modality workflow design.