Ira Kirschenbaum, Chairman, Orthopaedics at Bronxcare Health System
It is unclear to me exactly when the Electronic Health Record (EHR) jumped the shark. The problem with EHR’s is not that they are difficult to use. It is not that they are cumbersome during office hours. It is not that they are filled with extraneous information. While all these things are true, the real problem is that they simply are useless repositories of partially recorded snippets of events and no real business or clinical intelligence can be harvested from them.
There is a ton of information in an EHR but so what? In 2000, I went completely paperless in my office. I was in private practice at the time and there was no real EHR available. I did this with off the shelf products. My notes were dictated with Dragon and I had templates I used with basic copy and paste. Patient files were organized using the computer operating system file/folder hierarchy, scanning was done with a great program called PaperPort and that was it. Fast forward to today where in my hospital we use one of the big EHR companies. The notes are written using templates I constructed (not them) with Dragon. The patients are filed using a Medical Record search system. Internal labs and results are posted (an actual advance). The system does not allow for point-of-care scanning of documents received by the patient. In the end I have about the same system- just millions of dollars later.
"We have to remember that what we observe is not nature herself, but nature exposed to our method of questioning.- Werner Heisenberg"
The problem is that when I want to know which patients I operated on had hepatitis and what was their liver enzyme profiles (just an example) - I can’t do that myself- I need to put a “ticket” in and wait 2 weeks for a report to be done by the Health Information Technology department. In fact, if I want to know who was discharged from my service last week-not possible. We are not mining the right kind of data because the EHRs do not allow us to ask the questions we want.
In reality, EHRs represent a great opportunity for big data questions and answers. The problem is that EHRs have limited query environments. When I took over as Chairman of Orthopaedics at my hospital, I embarked on a quest to find ways to collect and analyze data with greater speed and granular accuracy. I summarized this journey to some extent at a recent Stanford MedicineX meeting (Fall 2017). In the lecture, I summarized the deployment of a powerful low-code database environment developed by the company Trackvia. This database program allowed me to use all the clinical information hiding in my EHR and fill the gaps so sorely needed. After scraping or outsourcing the input of key data we built smart applications. Now every doctor knows where every patient is. Patients seen in the ER are posted to Trackvia and the Attending can remotely see the progress of the treatment. We manage OR products through this- requests and negotiating pricing. Surgical scheduling, not a part of our EHR, are done using this. Our docs know who they are operating on this week, next week, and the foreseeable future. The platform even tells us which cases have not been coded or billed. We do clinical research on this platform. We even built an application to have referral docs send us patients through this database system. I recognize that I am a low-code developer as well as a surgeon and don’t expect everyone to have the platform that we developed but with the millions of dollars spent on my EHR it is amazing how our parallel relational database fills the “questions” gap. We ask questions and we get answers.
EHRs unfortunately think like computers and not like clinicians. I hope that future generations of EHRs have the ability to cull data and answer the types of questions clinicians need.