I have posted a number of previous notes about integrated diagnostics and integrated diagnostic centers (see: Integrated Diagnostics and Its Relationship to Digital Pathology: A Strategic Analysis; Revisiting Integrated Diagnostics and the Integrated Diagnostic Report; Diagnostic Delay Time (DDT) and Integrated Diagnostics; A Call for the Development of Integrated Diagnostic Centers). The basic concept of integrated diagnostics is to aggregate pathology and radiology results for specific patients in order to: (1) decrease the time to diagnosis by providing guidance about the most efficient path to arrive at a diagnosis and (2) enable the analysis and interpretation of interim test results by which the time to diagnosis is also reduced by reporting "next step" recommendations. In order to pursue the goal of integrated diagnostics in a hospital, very close collaboration between pathology and radiology is necessary. I have suggested in the past that perhaps these two specialities could merge to form a new medical speciality of Diagnostic Medicine (see: Ten Reasons for Merging Pathology/Lab Medicine with Radiology). This latter note was posted in 2006. In the intervening eight years, I have come to understand that the creation of such a new speciality is highly unlikely.
The pursuit of the broad goal of integrated diagnostics in support of cancer diagnoses in a hospital requires the development of integrated diagnostic servers (IDSs) under the control of pathology and radiology. Under this scenario and when radiology or pathology departments first identify patients being worked up for a suspected but undiagnosed malignant lesion, a file will be created for that patient on the IDS and all subsequent pathology and radiology results for that patient will be copied to the server from the LIS, RIS, and PACS. This IDS is thus be a component of a larger, proposed federated, service-oriented IT architecture (SOA) in pathology and radiology.
Installed on the IDS will be a complex set of heuristics, business rules, and algorithms developed within pathology and radiology that will analyze all available diagnostic information for each presumptive cancer patient and then recommend the next set of tests and radiology procedures that are necessary to arrive at a diagnosis in the least amount of time. Usually, these IDS recommendations are forwarded to the patient's clinician. Under a reflex testing testing option if and when ordered by the patient's treating physician, additional test and procedures orders can be ordered automatically by the IDS. This reflex testing option speeds up the time-to-diagnosis by reducing the number of test ordering cycles. I discussed an early form of the IDS reports now operating in pathology at Pitt in a previous note about their so-called comprehensive theranostic summary (CTS) (see: The Comprehensive Theranostic Summary (CTS): A "Must Have" for Surgical Pathology).
Only pathologists and radiologists will have sufficient knowledge of tests and imaging procedures to create these heuristics, business rules, and algorithms. Because of the very rapid advances in the understanding of cancer genomics and cancer biomarkers, they also may change on a monthly basis. They also need to be constantly tested and validated in a hospital environment with the goal of shortening the time to diagnosis. I anticipate that the deployment of of IDSs will also reduce the cost of cancer diagnoses because of the elimination of unnecessary, irrelevant, and redundant testing. As soon as it can be demonstrated that IDSs result in faster, cheaper, and better cancer diagnoses, these heuristics, business rules, and algorithms will become extremely valuable intellectual property.
There is almost no likelihood that integrated diagnostics, as described here, can ever be accomplished by EHRs (see: Genomics-Based EHR: Is This a Realistic Expectation?; Predictions for the Post-EHR IT Era; Business Continuity Challenges; The -Omics Cloud: A Healthcare IT Solution Already Developed for Genomics Research). These systems are too large and unwieldy and are also designed primarily to replicate the paper medical records and generate bills. Also, they will be unable to allocate the necessary computing power to run the complex and ever-changing heuristics, business rules, and algorithms described here. Moreover, EHR vendors don't have ready access to the diagnostic expertise required for their development nor can they install new software with the short turnaround-times that will be necessary to test and validate them.
I will be providing more details about the proposed IDS to the members of the International Society for Strategic Studies in Radiology (ISSSR) on December 4, 2014, that will convene as part of the upcoming RSNA conference in Chicago. Because of the complexity and challenges of the development of IDSs, I will also address other aspects of them in upcoming blog notes. I also invite any readers of Lab Soft News to comment on this idea.