HomePod and Multi-User Siri in Health Care: It’s how to significantly reduce patient wait times

Since the release of version 2.0 of the Patient is in app, Siri has been supported in both the Patient is in iPhone app and the Patient is in Apple Watch app providing doctors with a voice first user interface. With version 2.5 of the Patient is in, multi-user Siri is now supported on Apple’s HomePod wireless speaker.

Wireless patient announcements with AirPlay speakers

If you deploy an AirPlay-compatible speaker such as Apple’s HomePod on your network, the Patient is in iPad app can wirelessly announce patient assignments. As seen in these screen shots, the charge nurse selects the AirPlay control from the navigation bar:

ipad-navbar-accepted-1

and a pop up dialog will display available AirPlay-compatible speakers such as the HomePod in the Anesthesiologists’ Lounge and the Apple TV in Dr. Quinn’s office:

airplay-controls-anesthesiologist-lounge-selected

 

Once an AirPlay-compatible speaker is selected as in the above example, then whenever the charge nurse assigns a doctor to treat a patient in a room, the HomePod in the Anesthesiologists’ Lounge will announce the patient assignment. As we will see in the video at the end of this blog post, the charge nurse assigns an anesthesiologist to Recovery Room 1 and the HomePod will announce: “The charge nurse says that the anesthesiologist has an assignment in Recovery Room 1”. And to ensure HIPAA compliance, any patient notes provided by the charge nurse will not be read aloud.

Multi-user Siri and HomePod

Much like on the Patient is in iPhone app, the doctor may use Siri to hear her patient assignments with the phrase: “Hey Siri, read my patient messages” and on both Patient is in apps for iPhone and Apple Watch, the doctor may respond to the patient assignment with “Hey Siri, send a patient message saying I’ll go to the exam room in 10 minutes” or “Hey Siri, send a patient message saying I’ve finished surgery”; now doctors can listen to their assignments on HomePod and respond to the charge nurse with Siri. As we see in this video, the charge nurse assigns the on-duty anesthesiologist to different recovery rooms and different anesthesiologists respond using Siri on the same HomePod:

 

For more information about Siri support in the Patient is in app, visit these links:
Siri integration for a voice interface and Et tu Apple Watch: Integrating Siri with an AI engine to reduce patient wait time

Apple Watch Series 3 with Cellular: Patient Assignments on the Go

Since its introduction, the Patient is in watch app supported WiFi connections allowing the doctor to use her Apple Watch while away from her iPhone. With the release of version 2.5 of the Patient is in, the doctor will now receive patient assignments over a cellular connection. The doctor may respond to the patient assignment directly from the notification as seen in theses screen shots:

watch-notification-recoveryRoom2-assignment-1-casewatch-notification-recoveryRoom2-assignment-2-casewatch-notification-recoveryRoom2-assignment-3-case

or directly from the main watch app as she normally would as seen in these screen shots:

watch-mainApp-new-assignment-OR1-1-casewatch-mainApp-done-assignment-OR1-1-case

 

As in previous versions of the Patient is in watch app, Siri as well as FaceTime audio conference calls are also supported over cellular as seen in these screen shots:

watch-siri-both-3-casewatch-notes-facetime-available-casewatch-todayStaff-1-case

 

For more information on the the Patient is in watch app, please visit:
An Apple Watch App for Doctors

Reducing Patient Wait Times with AI and Microsoft’s Azure Cloud: Integrating Virtual Assistants, Wearables, and IoT

The Patient is in reduces patient wait times by integrating a series of technologies including AI and natural language processing, the Siri virtual assistant, the Apple Watch wearable, and the iBeacon IoT micro-location technology.

In this blog post, we will discuss two features of the Patient is in app which use Microsoft’s Azure cloud: Apple Watch notifications and Skype for Business. The Azure cloud was selected over Amazon’s AWS cloud as well as Google’s and IBM’s cloud for many reasons but the most important is that Azure supports the latest iOS and Apple Watch technologies and provides greater flexibility in managing devices.

Realtime patient assignments and updates

From the perspective of your doctor, the charge nurse sends a patient assignment from the Patient is in iPad app to the doctor’s Apple Watch and iPhone. On her Apple Watch, the doctor can accept the assignment and provide an estimated time of arrival, reject the assignment, and update the charge nurse that she has completed her assignment and thereby allowing the cleaning staff to prepare the room and consequently reduce the wait time for the next patient.

In this screenshot, we see the charge nurse using the Patient is in iPad app to visualize doctor availability and room status:

ipad-assignments

and here we see the charge nurse creating a patient assignment by assigning a doctor to treat a patient in Recovery Room 3:
1_ipad-eo-assigned-recoveryRoom3.png

The patient assignment is encrypted and securely sent from the Patient is in iPad app to the Microsoft Azure cloud which notifies the doctor on her Apple Watch of the assignment as we see here:

watch-notification-OR1-assignment-1-case
The Patient is in watch app will alert the doctor with:

  • A discrete vibration felt on the doctor’s wrist
  • An optional, audible sound

The doctor can then update the charge nurse with her estimated time of arrival and likewise updated the charge nurse after she completes her assignment as seen here:

 

 

In large clinics and hospitals multiple iPads may be deployed to administrators and charge nurses and of course, a new iPad may be running side-by-side an older iPad during repairs. The Patient is in iPad app will securely synchronize its patient assignments to all other Patient is in iPad apps via iCloud and thereby provide a realtime global view of:

  • Doctor availability
  • The status of each doctor assignment
  • The status of rooms
  • Location of off-site doctors with live map and traffic view
  • Location of on-site doctors using iBeacon micro-location technology

In this screenshot, we can see that the doctor is onsite and near Nurse Station 6. The doctor’s onsite location is provided by the doctor’s iPhone proximity to a charge nurse’s iPad all of which use iBeacon micro-location technology to determine proximity. A doctor may enable or disable this feature from the Patient is in iPhone app.

beacon-sibley

Group conference calls with Skype for Business

The Patient is in iPhone app is integrated with Skype for Business which is Microsoft’s HIPAA compliant audio and video conferencing solution for groups. Unlike Microsoft’s consumer version named Skype, Skype for Business enables medical staff to record their conference calls and comply with the various HIPAA requirements regarding conference calls. While hospitals may choose to install their own Skype for Business servers on their own premises, smaller organization can choose the Microsoft Azure cloud-based solution which costs a few dollars a month per staff member.

Much like how the charge nurse sent a patient assignment with Patient is in iPad app, she can also send a conference call request to the doctor and the doctor can join the conference call directly from the Patient is in iPhone app as seen here:

 

iphone-sfb-notification-bezel

 

and here is a clip showing the integration between the Patient is in and the Skype for Business iPhone apps:

 

 

 

Et tu Apple Watch: Integrating Siri with an AI engine to reduce patient wait time

The Patient is in apps integrate with Siri to provide a voice interface for messaging between the doctor and the charge nurse. This includes listening to patient assignments, rejecting an assignment, accepting an assignment and providing an estimated time of arrival, and lastly notifying the charge nurse of assignment completion which allows the staff to start the cleaning process to reduce the wait time for the next patient.

On both the iPhone and Apple Watch, this enables the doctor to use AirPods and a few other Bluetooth headsets to remotely manage patient assignments exclusively with her voice. For shorter distances, the “Hey Siri” voice trigger also works well and that is the exclusive technique for hands-free voice control on Apple Watch as of watchOS 3.2. Click on this link to learn about Siri support in the Patient is in.

In this article, we will look at the underlying natural language processing (NLP) engine built for Siri integration with the Patient is in messaging features.

Natural Language Processing Theory

NLP, is one of the core AI speech technologies along with text-to-speech (TTS) and speech-to-text which is also known as speech recognition. Advances in text-to-speech has led to more natural sounding computer voices and advances in speech recognition has led to better audio transcription.

NLP attempts to understand the meaning or intent of a sentence and to do that, an NLP engine must first be able to determine in which language is the sentence. Next, the part of speech of the words used in the sentence, as well as other components of a language such as word stems and contractions must be identified. For example, the doctor’s sentence: “I’ll go to the front office” would be decomposed or parsed into the pronoun “I”, the verb “will”, the verb  “go”, the preposition “to”, the determiner “the”, the adjective “front”, and the noun “office”.  So even though the doctor spoke the word “I’ll” the NLP engine had to understand the concept of American English language contractions and process the two words “I” and “will”. When the doctor annunciates correctly, then everything works well as we see in these screen shots:

Siri in Healthcare: Combinators and the Patient is in

Beyond contractions, advanced natural language processing algorithms must also consider the accent with which a user speaks. From the NLP engine’s point of view, everyone has an accent whether its the slow-taking southerner’s accent, the neutral mid-atlantic accent, the fast-taking New Yorker’s accent, a Bostonian accent, or perhaps the accent of a partial deaf therapist. And of course, when we are tired, we all tend to mumble a bit making us sound inarticulate. This also happens to surgeons and anesthesiologist after a middle-of-the-night emergency surgery. So, it should not be surprising that the intended word “I’ll” can be misspoken just enough to be transcribed as the word “all” as we see in the following screen shot:

watch-siri-all-go-1-case

The NLP engine used in the Patient is in apps on both Apple Watch and iPhone compensates for these common transcription errors because the algorithms were tailored for a doctor’s use of Siri.

Other transcription errors can seem insurmountable to deduce the doctor’s intent. For example, can you fix the phrase “go to low or one”? The app’s NLP engine was able to correctly determine that the doctor intended “I’ll go to OR 1” as we see in these screen shots:

You should also notice that the NLP engine also understands American Language homophones as in “8 and eight and ate”, “4 and for and four”, et tu “2 and to and too and two”.

If we reexamine the classic NLP pipeline advocated in AI research as mentioned at the start of this article, we learned that it started with first identifying the context language and then parses the text into parts of speech. When we applied this to the doctor’s message we saw that the phrase “Exam Room” was decomposed and identified as the adjective “front” and the noun “office”. While linguistically correct, that does not help us because the app has to notify the charge nurse that the doctor has accepted the assignment to go to the specific room named Exam Room to treat a patient or perform surgery.

Unlike the Siri messaging support in apps such WhatsApp or Messages which simply relay the transcription provided by Siri’s speech recognition, the Patient is in must determine which room is the doctor discussing, if the doctor is rejecting an assignment, completing an assignment, or accepting an assignment and providing an estimated time of arrival which may be provided in hours or minutes.

The NLP engine uses multiple algorithms and text processing techniques to best ensure that the doctor’s intent is correctly captured even if the doctor’s words were incorrectly transcribed. The approach to use multiple algorithms and extraction techniques is relatively new and is called parser combinators. The NLP engine used in the Patient is in apps on both the iPhone and Apple Watch adds many Siri-specific and doctor-specific algorithms to the classic NLP approach.

Due to the uncertainty in processing speech, a new user interface idiom has evolved. Called the conversational user interface, it enables Siri to mediate a conversation between the user and the app and more abstractly between the doctor and the charge nurse.

Introducing Conversational Interfaces

Like in real life, rarely is information unambiguously clear. If you ask a taxi driver to drive you from the John Wayne Airport in Santa Ana, California to your office “on Main and MacArthur in the next town over” which is a few blocks away in the city of Irvine, he may instead drive you a few miles further to Main and MacArthur in the city of Costa Mesa as both cities are adjacent to Santa Ana and not only have the same street names but are actually the exact same streets which intersect in two different cities a few miles apart.

With voice interfaces, the app has to support a conversation to clarify the user’s words by asking for more information from the user either because the user has not provided enough information or has provided ambiguous information. The app must also confirm its understanding before taking action on behalf of the user.

Siri mediates this conversation between the user and the app and handles the speech recognition and passes the text transcript to the app which processes that with natural language algorithms and other text processing technologies. If the app needs additional information, it asks Siri to prompt the user by providing Siri with text to read to the user. Siri uses text-to-speech to ask those questions. After the app is satisfied that it understands the user’s request, the app asks Siri to ask the user to confirm or reject those assumptions and with the user’s permission, the app finally processes the user’s request.

What’s next in Conversational Interfaces?

Voice technologies and conversational interfaces empower app developers to use AI to provide alternative ways for users to access their app. These technologies, also extend access to a larger set of users. Through a voice interface, perhaps app developers will finally realize that accessibility and usability are fundamentally intertwined or as Tim Cook recently said in an interview marking Global Accessibility Awareness Day 2017,  accessibility is a human right.

Why Build an App for Blind Therapists?

Today is Global Accessibility Awareness Day #GAAD #gbla11yday and I am pleased to announce that my app the Patient is in provides the best accessibility features in a real time, HIPAA-compliant messaging app for the Apple Watch and Siri. Built over 15 months with leading doctors, therapists, and practitioners to support the needs of blind therapists, the Patient is in made its public debut at last week’s Digital Health Demo Day.

Many people ask me why did I first implement accessibility support in this ground breaking real time, HIPAA-compliant messaging app rather than add a localized version of the app for French, German, or Spanish speaking doctors and nurses. They correctly note that there is a larger market for real time medical messaging for French, German, and Spanish speaking doctors and nurses than there is for English speaking blind therapists and their nurses. But the fact is that most app developers struggle with just creating a useful and usable app that they just don’t care about such a small group of potential users. Unfortunately, the imagined profits from the many outweigh the needs of the few.

Here’s a secret that only a handful of app developers know about and the real secret why Apple has invested billions of dollars in accessibility to support the needs of a small group of its users: if you design accessible products both your underlying technology and product will be more usable by the majority of your users.

This does not discount the obvious compassion which has driven Apple engineers and managers to support all types of users with both accessible features as well as foreign language support since the 1980s. As Apple watcher Rene Ritchie from iMore.com often says: “Apple engineers create products which their families use so they try to build the best and usable products”. This is why Apple products have soul and many products from Google and Microsoft lack soul. How many CEOs have you seen brought to near tears when discussing the empowerment of blind users of their products? Watch the end of the then new CEO Tim Cook’s first WWDC keynote in 2012 and you will witness a soulful man.

If you are a press person, query your contacts to see how many of Apple’s top human factors/user experience engineers worked on the vibrations and sounds used on the Apple Watch. And their requirements were supported by great hardware engineers who worked on the Taptic engine. By the way, notification of a new or urgent patient assignment through “discrete vibrations felt on the doctor’s wrist” continues to be the most useful app feature for most doctors demoing the Patient is in and that feature was originally built to support notifying a blind therapist that he had a new patient assignment.

Beyond the engineering benefits in building accessible apps, here is another reason why we should think about blind therapists on Global Accessibility Awareness Day 2017: blind therapists support our blind children as they start learning about the world and assist them as they transition into adulthood. Blind therapists also help our parents during the transition brought on by age-related macular degeneration.

 

The Patient is in accessibility features

The Patient is in provides access to patient assignments and updates in 4 ways, all of which have been designed to support low-vision and blind doctors and therapists:

  • The watch app supports VoiceOver and provides a great user experience.
  • Patient assignments from the charge nurse can be automatically announced over  AirPlay compatible speakers such as Apple’s HomePod. This removes the step in which the user must first access the iPhone notification to find out about the assignment details
  • Conversations with the charge nurse using Siri on the iPhone: “Hey Siri, read my Patient messages” allows the therapist to hear her patient assignments and “Hey Siri, send a Patient message saying I’ll go to Therapy Room 2 in 10 minutes” allows the therapist to respond to assignments from either her Apple Watch or iPhone with an estimated time of arrival. Notice of assignment completion is supported from both the Apple Watch and iPhone with “Hey Siri, send a Patient message saying I’ve completed my assignment in Recovery Room 2”. All of these statements are processed by the Patient is in natural language processing (NLP) engine fixing homophones (“Exam Room 2″ vsexam room to” vsexam room too” vsexam room two”) and other linguistic and transcription impediments to create a structured message upon which the charge nurse’s iPad app can visually display and drive real world processes such as cleaning the room so that the next patient’s wait time is significantly reduced.
  • FaceTime support is integrated directly into all of the apps allowing the charge nurse, therapist, and other doctors to instantly connect. This includes initiating a FaceTime audio call directly from the watch app.

… just my bit to inspire other app developers for Global Accessibility Awareness Day 2018