JACR: Data reconciliation has its place in medical imaging
While the most common medical application of data reconciliation is in pharmaceutical reconciliation, this concept can be applied to the delivery of medical imaging services, which begins with order entry and ends with reporting and communication, according to an article published in the September edition of the Journal of the American College of Radiology.
“On a more in-depth analysis, data reconciliation within medical imaging should extend beyond the imaging cycle and include clinical data that are critical to ensuring imaging data optimization,” wrote Bruce I. Reiner, MD, Paul B. Hall Regional Medical Center in Eden Prairie, Minn. “Clinical data are critical to the steps of ordering, protocol optimization, image processing, interpretation and reporting. Medical imaging in the absence of relevant clinical data is the equivalent of practicing radiology in a vacuum, with the potential for adverse clinical outcomes.”
The relevant data sources for medical imaging reconciliation consist of longitudinal clinical and imaging data, Reiner stated. “For reconciliation of contemporaneous (current) data, the clinical data requirements include the clinical indication for the ordered imaging examination (chief complaint), the presumptive diagnosis (or diagnoses) and relevant clinical support data (laboratory and physical examination findings).”
Three data components contribute to imaging data analysis:
In addition, the derived data can in turn be used to create data-driven, best-practice guidelines and computerized decision support tools. In addition to automated extraction of predefined associated data, a radiologist could manually input specific database queries, which could then prompt the computer agent to search the clinical and imaging databases for corresponding data, Reiner wrote. “Once these data have been recognized as having relevance to a given context, end user or patient, they would become automatically incorporated into the automated data extraction software.”
Medical imaging is well suited for data reconciliation applications because a well-defined process occurs in the delivery of medical imaging, beginning with order entry data and concluding with report data, concluded Reiner. However, the challenges to creating such a technology are significant and include the standardization and integration of multisource data, which are largely disparate and proprietary in nature. “A collective effort by the medical imaging community and technology providers can overcome these challenges if united, with the goal of promoting continuity of care and improved clinical outcomes,” Reiner stated.
“On a more in-depth analysis, data reconciliation within medical imaging should extend beyond the imaging cycle and include clinical data that are critical to ensuring imaging data optimization,” wrote Bruce I. Reiner, MD, Paul B. Hall Regional Medical Center in Eden Prairie, Minn. “Clinical data are critical to the steps of ordering, protocol optimization, image processing, interpretation and reporting. Medical imaging in the absence of relevant clinical data is the equivalent of practicing radiology in a vacuum, with the potential for adverse clinical outcomes.”
The relevant data sources for medical imaging reconciliation consist of longitudinal clinical and imaging data, Reiner stated. “For reconciliation of contemporaneous (current) data, the clinical data requirements include the clinical indication for the ordered imaging examination (chief complaint), the presumptive diagnosis (or diagnoses) and relevant clinical support data (laboratory and physical examination findings).”
Three data components contribute to imaging data analysis:
- Technical (numeric);
- Image (pixel); and
- Report (textual).
In addition, the derived data can in turn be used to create data-driven, best-practice guidelines and computerized decision support tools. In addition to automated extraction of predefined associated data, a radiologist could manually input specific database queries, which could then prompt the computer agent to search the clinical and imaging databases for corresponding data, Reiner wrote. “Once these data have been recognized as having relevance to a given context, end user or patient, they would become automatically incorporated into the automated data extraction software.”
Medical imaging is well suited for data reconciliation applications because a well-defined process occurs in the delivery of medical imaging, beginning with order entry data and concluding with report data, concluded Reiner. However, the challenges to creating such a technology are significant and include the standardization and integration of multisource data, which are largely disparate and proprietary in nature. “A collective effort by the medical imaging community and technology providers can overcome these challenges if united, with the goal of promoting continuity of care and improved clinical outcomes,” Reiner stated.