Written by Dr. Oliver Aalami, MD
We have all heard the adage, “Pictures are worth a thousand words.” Well, this is 100% true. Not only do images pack a ton of information, our brains are designed to interpret them at lightning speed. Dr. Fei-Fei Li of Stanford University showed that 100 milliseconds were enough time to recall significant amounts.
of information from images. Electronic medical records (e.g. EPIC or Cerner) are designed primarily for billing, and documentation has always been via text display. When working in the “trenches” of a hospital with a clinical team, communication and collaboration is key. The primary mode of communication is via text messages. Interestingly, whenever there is a pertinent visual finding noted on physical exam it is most often shared as a photo or video. Why? Because snapping a photo and sending is incredibly fast and efficient and does not leave room for interpretation.
These photos are incredibly valuable and are unfortunately never repatriated with the electronic medical record
1. If we were following up on a rash or wound infection and a patient returns to the office to be seen by another provider, what can one compare the exam to? Written descriptions are often incomplete, and as mentioned above, leave a lot of room to interpretation. Even if I were to see the patient for follow up, would I really take the time to find the thread of texts in which the photo was shared or go back to my photo library of my smartphone to try to find the photo? Nope, this is like finding a needle in a haystack. Discovery of old images taken is a major pain point
2. Images in electronic medical records are typically stored in what is known as the “Media Tab” which often contains consents, scanned nursing documents, advanced directives, etc. The typical workflow for storing images to the medical record can be either
1. uploading jpeg images from a digital camera’s SD card by connecting to a computer, or
2. taking an image using an EHRs mobile app (EPIC Haiku, for example). In settings where images are frequently taken such as the wound clinic, emergency rooms or hospital wards, a mixture of these two workflows is used.
But there are several clinical usability issues to this solution which include,
Images cannot be easily shared with key provider team members, you have to be in the EHR’s media tab for that patient to view the images,
The viewing experience is horrible, comparable to opening up PDFs. When you open up an image you may only see the upper right-hand corner of image and you have to scroll down and to the right to see the rest of the image, nothing close to viewing images on a PACS system,
Image discovery is challenging- the onus is on user to sift through documents/images of media tab to find the image they are looking for,
Images are flat files with little to no structured metadata. There is an option to enter a non-structured free text description of each photo. These are not searchable and more importantly can not be leveraged for artificial intelligence or machine learning tools.
We need a solution that allows for…
Quick image capture
Image annotation with structured metadata for easy search and analysis via algorithms
Association with the electronic medical record/ image archiving system
Easy sharing and archived collaboration for future viewing
Quick image discovery
Intuitive image viewing experience
When working in a clinical setting and you spend your day putting out one fire out after another, every millisecond counts. Integrating medical grade imaging into daily clinical workflow can unlock tremendous speed and efficiency, as Dr. Fei-Fei Li’s work showed, and ultimately improve collaboration and confidence in clinical decision making.