Posted by: RedaChouffani
EHR, iOS 5, mhealth, mHealth integration, SIRI
With the release of the iPhone 4S and the deployment of the new Siri application as part of iOS 5, it is clear that the new voice assistant will ultimately find a place in the health care industry. The only question that remains is, “in what capacity?”
This scarily accurate functionality is currently being offered as part of the new iPhone and all future iPhone versions using Natural Language Processing (NLP) in the back end to process voice commands and identify what the user is really asking for. And based on the spoken commands or requests, the system then identifies the appropriate response or action to fulfill. In the past week or so, I have been experimenting with the new SIRI — or as I call her, “Melissa,” my new virtual assistant — and within a day or so of using this technology, it was very clear to me that it would be here to stay. The genius of the technology is not in the technical capabilities, though, as it has already been proven and implemented elsewhere previously. The innovation here is the how NLP is so well integrated with the different components of the device, and how well it can interact with all of them and act as a communication liaison.
Without a doubt, the majority of physicians will soon be utilizing a virtual assistant with vocal command functionality to help them with many of the tasks that currently require too many clicks — as in the kinds that one typically performs within an EHR or other like health care application.The differentiating benefit of the virtual medical assistant is that it is interactive, and in the future, this will allow physicians to request charts, view lab results, see prescription lists, schedule appointments, request best available treatments, identify the closest clinical trials clinics for a patient condition and even provide feedback on a given treatment plan.
Clearly there are many advantageous uses of the technology that can lead to tangible cost savings — not only by increasing efficiencies, but also by bringing system integration to a whole new level. Let’s take few of the use cases in which a computerized assistant with medical intelligence, or “CAMI”, can be used:
1.CAMI basic selects and lookups:
“CAMI, pull the patient John Doe’s medical chart and get me the last patient summary.”
In this use case, the system will be able to identify if there is an accurate match for the patient name John Doe, and if there are multiple matches returned, then CAMI will notify the physician to provide further identifiers in order to narrow down the selection. Fortunately, this querying capability has been available in different platforms such as C-Phrase and has become very advanced. This means that health care systems can perform very complex look-ups against databases using NLP.
2.CAMI basic actions:
“CAMI, please schedule this patient for a follow up appointment in two weeks for a check up.”
“Also CAMI, prescribe for John Doe Metoprolol 25 mg 1 tablet twice daily, and Plavix 75 mg 1 tablet daily”.
In this scenario, the system pulls the schedule, sets up a follow up appointment on the physician’s schedule and sends the patient the appointment details as well.
For the second step, CAMI will start the e-prescribing session. The system will first use a national medication database to lookup the drug names and ensure that what was requested is in fact available, and to ensure that the prescribed drug does not have any adverse affects due to drug/drug interactions with any existing medications that the patient might be on.
3.CAMI scheduling coordinator:
“CAMI, when is my next surgery, and how long would it take for me to get there from here?”
For a surgeon who is seeing patients in multiple locations (ASCs, hospitals, physicians offices), having access to the patient list and daily schedule is critical. But what makes the use of CAMI even more valuable is the fact it is interactive, and can be used to alert a physician when the next appointment is. Not only that, but if the physician is offsite or having lunch far away from where their next patient appointment is, then the system will notify them if the travel time so they won’t miss their appointment.
4. CAMI and communications:
“CAMI, please let nurse Kimberly know that we are ready for the patient to get x-rays right away”.
For this case the system will be simply be used to communicate quickly with coworkers via emails, SMS, or through the EHR system.
5. CAMI and correspondence:
CAMI, please dictate the following message and send along side with John Doe’s post Op notes to his referring physicians.”
In this example, the system will transcribe the physician’s intended letter, gather some of the patient’s health record information and send it via the preferred method (eFaxing, or secure email) to the referring physicians.
6. CAMI Integrations:
“CAMI, please get me the latest X-rays for John Doe.”
In this case, the system will need to interact with a digital imaging viewer and PACS server. It will be able to send DICOM file requests and display them appropriately for the physicians on their smart device.
As we explore the possibilities of using NLP and virtual assistants in health care, we must note that there have been steps moving toward using this powerful technology in the health care field for a while now. Recently we saw how WellPoint adopted Watson, the super computer, from IBM. The intent here is similar, in that the application combines data from patient electronic charts, hospital notes, WellPoint’s history of medicines and treatments and Watson’s incredible library of textbooks and medical journals to help recommend the appropriate treatment for the patient.
In reality, in order to get the capabilities needed to accomplish a successfully CAMI (who is SIRI’s cousin who went to medical school), the solution would need to be customized and configured specifically for each health system. It will have to allow the organization to implement their specific workflows and processes as part of the intelligence of this application.