Interpreting radiology studies used to be simple - too simple - because radiologists were given too little information for adequate diagnosis. We were handed a chest x-ray with a clinical indication of just “pain” or “please do in the AM, thanks” and that was it. We looked at the image and gave an interpretation, but had no idea why the patient was in pain. Did he fall? Did it come on suddenly? When we saw a suspicious nodule, we reported it but we didn’t know if the patient was a smoker, or if he had a family history of cancer. Could that have had an impact on our recommendations for follow-up of our findings? You bet it could. But we didn’t have the information, so it couldn’t happen. The full clinical history was in the patient’s paper chart, which stayed with the patient who was nowhere near the x-ray picture and radiologist at the time of interpretation.
All of this changed with the advent of electronic medical records (EMRs) over the past 10 years. Suddenly, all of this wealth of patient information which was shackled in paper records burst forth when the recording of that data transitioned to digital. Different EMRs evolved for different health care settings, most notably the Emergency Department (ED), the inpatient setting and the outpatient setting. Each of those settings has different needs with regard to the management of patient information. The ED must either admit or discharge a patient within 8 hours, so they need to keep the patients progressing to their end point at a reasonable pace. The in-patient setting has a slower pace, with careful attention to disease progression on a daily basis and many medication and consultation orders and results to juggle in order to resolve the acute disease phase so the patient can safely be discharged. The out-patient EMRs handle episodic encounters, with problems handled on a much longer time scale, over the course of weeks or months.
"As the wealth of information available from EMRs continues to increase, so must the mechanisms that present that information become more savvy, so as to allow doctors to provide better care from it rather than be overwhelmed by it"
Which type of historical information do doctors need in order to best treat a patient - ED, in-patient or out-patient? The answer is obvious - all of them. Every time a doctor starts treating a patient, there are some basic pieces of information they want to know regardless of the type of encounter they came from, like the patient’s problem list, recent laboratory and radiology results and procedure reports.
To that end, health systems began creating centralized warehouses of all patient data to allow every healthcare worker to see whatever they needed to know about the patients they treat. Rather than have to login to each EMR separately, clinicians can login to the central warehouse, sometimes called a health information exchange, to see all available information about a patient, including ED, in-patient and out-patient visits.
Once the floodgates were opened, the information started pouring in. Every problem, every health issue, every laboratory value and every clinical note became available. The creators of these warehouses quickly realized that without a strong clinical viewer to present the information in a clear and organized fashion, the information cannot be used effectively. They moved quickly to build an organizational framework for this vast amount of patient data, to allow clinicians to traverse it and drill down to the sections of interest with the minimum number of clicks. They sought to present the most recent information first, and categorize all the data so it can be navigated easily.
These clinical viewers have been a boon to all clinicians, including radiologists, who can now see the full clinical history of a patient who came in from the ED, including the ED Provider Note, recent laboratory results, problem lists and procedure reports. It sounds great, and it is great. But can we do better?
Of course we can. Why should the radiologist have to sift through the irrelevant parts of the medical record to find the relevant parts when it is already known what she or he needs to see? Even with this information organized as best as it can, it takes time to drill down to the relevant parts because they are found in different parts of the record, and degree of relevance depends on the clinician and their use case. What is most relevant to the radiologist is not necessarily what is most relevant to the surgeon, for example.
Therefore, at Northwell Health we created a function called Relevant Clinical Info which automatically scans the EMR for the basic info that a radiologist needs to provide the best interpretation of a radiologic study. It gathers the History of Present Illness, most recent and relevant progress notes, laboratory results, anatomic pathology reports, procedure reports, operative reports and problem lists, and presents them immediately to the radiologist with one click. Since the information is extracted ahead of time, at the time the patient is being scanned, it is ready in an instant as a single web page to the radiologist upon demand. This saves time, and can make sure that important information is not missed when looking through the medical record.