In 1982, when I started my radiology residency, doctors didn’t use computers. Most of us couldn’t type, but we did have some long-lost skills. We could flip a film onto a view box and shape a catheter using steam from a kettle. For greater reading efficiency, we used “alternators,” a wonderful contraption consisting of 25 or more glass panels driven by conveyor belts so that pre-hung films could be displayed in sequence. Nonetheless, alternators left plenty of room for improvement - they often mangled films, for example. Murphy’s Law mandated that the referring doctor on the phone always wanted to discuss the exam on the glass panel furthest away from display.
By 1992, the future arrived. When we started using our first PACS, a DOS-based workstation called “The Dominator,” we suddenly gained random access to all of our imaging exams. Wow! We beamed, “The Dominator is the terminator of the alternator.” It turned out that we were right.
PACS helped us in many ways. We could read faster, compare better, and communicate more efficiently. We stopped losing information. We could read exams in any order from just about anywhere. We could cherry-pick exams in our sub-specialty. The young file room worker whose job it was to find “lost films” in the neurosurgeon’s automobile trunk went to school, learned about computers, and got a job as a PACS administrator. By 1994, the multi-modal computer enabled us to record an audio clip of each exam’s key findings, so that the summary interpretation was available to the referring doctor even before the full report was transcribed. Later, in the late 1990’s, the miracle of speech recognition enabled us to create reports nearly instantly. Reading rooms got quiet. Patient care sped up. Technology resulted in ever-more-efficient storage, display, interpretation, and communication.
Today, we stand on the threshold of yet another new era. Cognitive computing will increasingly help us better understand how to diagnose and treat patients. Within the next decade, cognitive computing will help determine the work-up, prioritize exams and images within an exam, assign exams to the proper reader, better understand the patient’s differential diagnosis, and help us provide evidence-driven recommendations. Computers will help us understand how to achieve the best patient outcomes and avoid complications.
Age-old reading paradigms will change. Today, the patient’s history, images, and reports compete for the reading physician’s visual and cognitive attention. Soon, our technology will present compiled clinical data that is relevant to the current complaint. The reading physician will be presented with the report pre-populated with the differential diagnosis, including the probabilities of each item. Reading physicians will view images and annotate the abnormalities via a combination of speech and other inputs. Gone will be the days of turning one’s attention to the report instead of the images. Automated anatomical localization and reporting technology will assist in creating an organized report, including tracking of previously observed index lesions. Instead of looking at a chest radiograph, MRI, or mammogram and then turning to the report, the reading physician will keep his or her eyes on the images, mark the findings, and thus complete the report. The differential diagnoses and likelihoods will be dynamically modified based on the marked imaging findings, and appropriate reference material will be automatically presented. Reports will then be automatically analyzed to weed out mistakes and pro-actively identify disease trends.
Beyond the benefits of cognitive computing, workstations will also increasingly be designed based on our growing understanding of human perception, cognition, learning, and emotion. To enhance job satisfaction, our workstations will promote the psychological state of flow, achieved when one is deeply concentrating on a challenging yet achievable task. The reading station will evolve to a single pane of glass with optimized lighting and ergonomics. Information and images will appear when and where needed, without tiresome head and eye movements. Input devices will gain in sophistication. Most of all, the software will evolve like a video game, allowing the beginning user to get started easily, while also promoting ever-increasing self-learning and mastery.
As with all change, there will be challenges and unintended consequences. Get ready…it’s going to be an interesting ride. Now is the time to identify the visionaries in your organization and empower them to lead the next generation of technological adoption in healthcare.
Murray A. Reicher, MD FACR, is the Chief Medical Officer at Merge Healthcare.