We live in extraordinary times. Imagine if someone told you 10 years ago that Blockbuster and Circuit City would cease to exist, and that a whole new realm of communication called social media would rise up and influence billions of lives across the globe. In the span of a mere decade, we have witnessed the fall of giants and the rise of new ones, founded by those living in a dorm room or working out of a garage. Times they are a changing, and it will be fascinating to watch the next several years unfold.
In 2004, Will Smith starred in a movie titled iRobot about the rise of anthropomorphic robots built to serve their human counterparts. In this article, I will attempt to look into the future and determine how this type of artificial intelligence (AI) may serve the medical community, and how those services could improve our healthcare delivery system. For the sake of this article, please allow me to speculate specifically on cardiology.
Cardiologists often rely on diagnostic tests to pinpoint the reason a patient may be experiencing certain symptoms such as shortness of breath or heart palpitations. These tests hold thousands of clues about what may be happening to the patient on a holistic level. The problem? Somebody needs to analyze all those clues and keep up with the latest evidence-based guidelines to decipher the best course of action. This task falls on our nation’s very talented heart doctors, and I foresee an opportunity that will present itself to lend a hand.
There is so much that needs to be accomplished prior to a cardiologist sitting down to read a CV diagnostic test. Depending on the type of test being read at the time, the physician and staff must cobble together all information relevant to that procedure.
In the near future, AI will be able to gather relevant data in real time whether the test is an echocardiogram, vascular ultrasound, cardiac catheterization procedure, holter monitor or myocardial perfusion imaging test. But why should we stop at simply assimilating data? AI will take it one step further and intelligently compare and analyze previous data sets to current data (be it intra or inter-modality) and present that information to the physician in a clinically relevant and contextually meaningful format. Now, with the right data at the right time and presented in the right manner, the cardiologist will have actionable intelligence at the point of care.
At this point, the AI has assisted several parties: the physician who once had different processes when reporting different types of diagnostic tests, the office staff formerly assigned with pulling old reports and scanning records, and the patient who is able to take comfort that his or her data has been effectively presented to the cardiologist for proper diagnosis. But we still have work to do. The cardiologist still needs to read the test and produce a report that satisfies numerous parties including the patient, referring physician, payers, accrediting agencies, and the hospital and/or clinic where the test was performed.
It is 2014, and the process to accomplish this generally takes place by a physician dictating into a recording device, a transcriptionist typing the report, the report going back to the physician for review and signature, and then being scanned into the patient record and faxed to the necessary parties. There is so much inefficiency here that AI could help solve. What if the AI could, upon analysis of the results described earlier, begin to build out the report on behalf the physician thus eliminating dictation and transcription?
This report would eliminate transcription errors, decrease time to finalization (no time wasted on dictation/ transcription/ manual signature) and be presented in such a manner that the cardiologist would never want to read a diagnostic test any other way, and the referring physician would never want to send his or her patient anywhere else. Upon finalization, the AI would also deliver the report directly into the patient record within the hospital or clinics own EHR, eliminating the need to scan reports. And finally, this AI would hold all those thousands of data points for future retrospective searches or prospective alerting based on the criteria that are important to your practice.
There are certainly other pain points such as a physician prescribing medication that is not on a patient formulary, or trying to keep up with ICD-9/ICD-10/Snomed Codes. How about accurate E&M coding or the challenges associated with obtaining diagnostic data to assist with clinical decision support and clinical research? How long must you tolerate such inefficiencies within your practice?
Now, what if I told you that this article is not a glimpse into the future, but a glimpse into today. At OMS, we have designed technology exclusively focused on cardiology, and we have partnered with some of the finest companies in our industry to deliver a world-class cardiology experience. We do not need to wait and see what the future holds…the future is available now.
For more detailed information regarding this capability, please see our previous article titled: Condition Critical: How Cardiologists Forced EHR to Evolve.