U.S. patent number 10,902,946 [Application Number 15/158,770] was granted by the patent office on 2021-01-26 for medication adherence device and coordinated care platform.
This patent grant is currently assigned to WATCHRX, INC.. The grantee listed for this patent is WatchRX, Inc.. Invention is credited to Arun Buduri, Jayanthi Narasimhan.
United States Patent |
10,902,946 |
Narasimhan , et al. |
January 26, 2021 |
Medication adherence device and coordinated care platform
Abstract
A method of supporting user adherence to a medication regimen
includes storing received medication information associated with
one or more medications using a patient wearable device and a
coordinated care platform. The method also includes displaying
identification and dosing information associated with the
medications, generating a notification indicating that the
identification information is being displayed, monitoring a
response input indicating that the medications have been
administered. For each of the one or more medications displayed,
the method includes generating a medication status describing that
either (i) the medication has been administered or (ii) the
response has not been received. The method further includes
generating a reminder notification when the medication status
indicates that the response has not been received. The method may
further include gathering behavioral data points associated with
the user for submission to the coordinated care platform, which
performs predictive analytics based on the data points.
Inventors: |
Narasimhan; Jayanthi (Acton,
MA), Buduri; Arun (Somerville, MA) |
Applicant: |
Name |
City |
State |
Country |
Type |
WatchRX, Inc. |
Acton |
MA |
US |
|
|
Assignee: |
WATCHRX, INC. (Acton,
MA)
|
Appl.
No.: |
15/158,770 |
Filed: |
May 19, 2016 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20160342767 A1 |
Nov 24, 2016 |
|
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
62164307 |
May 20, 2015 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G16H
40/63 (20180101); G16H 40/67 (20180101); G16H
20/10 (20180101); G06F 1/163 (20130101); G06N
5/04 (20130101) |
Current International
Class: |
G06F
1/16 (20060101); G06N 5/04 (20060101); G16H
40/63 (20180101); G16H 20/10 (20180101); G16H
40/67 (20180101) |
Field of
Search: |
;705/2,3 |
References Cited
[Referenced By]
U.S. Patent Documents
Foreign Patent Documents
|
|
|
|
|
|
|
WO 2016/187374 |
|
Nov 2016 |
|
WO |
|
Other References
Ilias Maglogiannis and Panayiotis Tsanakas, "Mobile Reminder System
for Furthering Patient Adherence Utilizing Commodity Smartwatch and
Android devices", Jan. 2015 (Year: 2015). cited by examiner .
Daniel Rosner, Andrei-Tiberiu Jurba, Razvan Tataroiu, Constantin
Ilas; "Wearable Medication Reminder Architecture Enhancement." 2015
20th International Conference on Control Systems and Science (Year:
2015). cited by examiner .
International Search Report and Written Opinion of
PCT/US2016/033181 dated Aug. 5, 2016 entitled "Medication Adherence
Device and Coordinated Care Platform". cited by applicant .
Maglogiannis, I. et al "Mobile 1-20 Reminder System for Furthering
Patient Adherence Utilizing Commodity Smartwatch and Android
Devices", Proceedings of the 4th International Conference on
Wireless Mobile Communication and Healthcare, 11 Transforming
Healthcare Through Innovations in Mobile and Wireless Technologies,
2014, pp. 124-127. cited by applicant .
Kannan, S., Wheats: A Wearable Personal Healthcare and Emergency
Alert and Tracking System 11, European Journal of Scientific
Research, Sep. 1, 2012 (Sep. 1, 2012), pp. 382-393, XP55291845,
Retrieved from the Internet:
URL:https:jjwww.researchgate.netjpublication/236268162 Wheats a
Wearable Personal Healthcare and-Emergency Alert and Tracking
System [retrieved on Jul. 28, 2016]. cited by applicant .
Remath, S., et al. , "Wearable Internet of Things: Concept,
Architectural Components and Promises for Person-Centered
Healthcare" Proceedings of the 4th International Conference on
Wireless Mobile Communication and Healthcare"Transformi NG
Healthcare Through Innovations in Mobile and Wireless
Technologies", Jan. 1, 2014 (Jan. 1, 2014). cited by applicant
.
UnaliWear, "Stay Active and Independent" http://www.unaliwear.com/
retrieved from Internet Aug. 18, 2016. cited by applicant .
TREATA Smart Solutions--"SmartCare Watch" www.treatasolutions.com
retrieved from Internet Aug. 19, 2016. cited by applicant .
VEESAG, "Mobile Personal Emergency Response System"
http://www.veesag.com/ retrieved from Internet Aug. 18, 2016. cited
by applicant .
Engadget "PillDrill does smart medication tracking in style"
http://www.engadget.com/2016/04/19/pilldrill-smart-medication-tracking/
retreived from Internet Aug. 18, 2016. cited by applicant .
International Preliminary Report on Patentability of
PCT/US2016/033181 dated Nov. 21, 2017 entitled "Medication
Adherence Device and Coordinated Care Platform". cited by
applicant.
|
Primary Examiner: Weisenfeld; Aryan E
Assistant Examiner: Blanchette; Joshua B
Attorney, Agent or Firm: Hamilton, Brook, Smith &
Reynolds, P.C.
Parent Case Text
RELATED APPLICATION(S)
This application claims the benefit of U.S. Provisional Application
No. 62/164,307, filed on May 20, 2015, the entire teachings of
which are incorporated herein by reference.
Claims
What is claimed is:
1. A method of supporting adherence, of a user, to a medication
regimen, comprising: by a processor-based wearable computing device
that is one of wrist-worn, pendant worn or armband worn, and
capable of stand-alone, independent and local execution of
processor instruction code and having local data storage configured
to store at least medication information and scheduling
information, the wearable computing device including a display and
an embedded cellular transceiver, the wearable computing device
configured to communicate and cooperate with a caregiver component
and a webserver component, which together with the wearable
computing device comprise a coordinated care system; storing
received medication information that characterizes one or more
prescribed medications, including at least medication
identification information including at least image information
sufficient to render a recognizable image of the medication to be
taken, the recognizable image being a photographic image captured
from the medication that was filled by a pharmacy, the recognizable
image depicting a quantity of the medication to be taken, and
prescribed dosing information selectively including (i) when food
is required to be taken with, before, or after the medication, (ii)
when the medication is to be taken at a specific time, and (iii)
personalized voice instructions, comprising a medication name, the
quantity of the medication, and/or instructions on how the
medication should be taken, that are voiced in a family member's
voice, associated with each of the one or more medications, the
medication information received from one or both of the caregiver
component and the webserver component; locally generating and
storing, by the wearable computing device and independent of
network connectivity, a visually-renderable reminder schedule based
on the received medication information, prescribed dosing
information, and a meal schedule associated with the user, the
reminder schedule generated to coordinate the meal schedule with
when food is required to be taken with the medication and/or when
the medication is to be taken at a specific time; displaying (i)
the medication identification information including at least the
recognizable image of the medication and (ii) dosing information
selectively including when food is required to be taken with,
before or after the medication, and when the medication is to be
taken at a specific time, associated with at least one of the one
or more medications according to the reminder schedule; generating,
based on the reminder schedule and independent of network
connectivity, an initial notification indicating that the
medication identification information is being displayed;
monitoring a response input instrumentality configured to receive a
response indicating that the at least one of the one or more
medications has been administered; for each of the one or more
medications displayed, generating a medication status describing
that either (i) the medication has been administered or (ii) the
response has not been received; and when the medication status
indicates that the response has not been received, (i) activating a
camera integrated into the wearable computing device, and (ii)
conveying the medication status to a caregiver, and receiving a
subsequent reminder from the caregiver in response to the conveyed
medication status.
2. The method of claim 1, further including generating a reminder
notification when the medication status indicates that the response
has not been received, by: re-displaying medication identification
information and dosing information associated with one or more
medications referenced in the reminder; generating a subsequent
notification indicating that the medication identification
information is being displayed; monitoring the response input
instrumentality; and for each of the one or more medications
referenced in the reminder, generating an updated medication status
describing that either (i) the medication has been administered or
(ii) the response has not been received.
3. The method of claim 2, further including repeating the reminder
notification periodically until: (i) all of the at least one
medications have been administered; or (ii) a predetermined time
period has elapsed since generation of the initial
notification.
4. The method of claim 3, further including providing the reminder
notification by playing a voice recording generated by a
caregiver.
5. The method of claim 3, further including repeating the reminder
notification at intervals of about 10 minutes.
6. The method of claim 3, further including notifying a caregiver
after the predetermined time period has elapsed.
7. The method of claim 1, further including automatically accepting
an incoming communication conveying the subsequent reminder,
generating a subsequent notification indicating that the medication
identification information is being displayed, and monitoring the
response input instrumentality configured to receive a response
indicating that the at least one of the one or more medications has
been administered.
8. The method of claim 7, further including automatically accepting
an incoming telephone call when the incoming communication
acceptance input instrumentality does not receive an accept
response within a predetermined time interval.
9. The method of claim 7, wherein the incoming communication is a
real-time communication configured to one or both of (i) survey a
patient and (ii) check in on the patient.
10. The method of claim 1, further including automatically
connecting to a healthcare device, receiving health information
from the healthcare device, the health information comprising one
or more of (i) blood glucose, (ii) blood pressure, and (iii) heart
rate, storing the health information and forwarding the health care
information to a caregiver.
11. The method of claim 1, wherein the identification information
and dosing information includes an image associated with the at
least one of the one or more medications, the image including a
depiction of a vehicle by which the medication is administered, the
vehicle instantiated a number of times corresponding to the dosing
information.
12. The method of claim 1, further including activating a camera
integrated into the wearable computing device, the activating
occurring in response to a camera stimulus provided by the
caregiver.
13. The method of claim 12, wherein the camera stimulus includes
one or more of (i) a camera activation communication received from
a caregiver, (ii) the medication status indicates that the response
has not been received.
14. The method of claim 1, further including synchronizing the
wearable computing device by receiving and storing updated
medication information.
15. The method of claim 1, further including determining
geolocation of the wearable computing device in relation to a
predetermined area, and notifying a caregiver when the geolocation
is determined to be outside of the predetermined area.
16. The method of claim 1, further including: (i) detecting that
the user has fallen by analyzing an output of one or more
accelerometers embedded in the wearable computing device; (ii)
notifying a caregiver upon detecting that the user has fallen.
Description
BACKGROUND
Medication adherence has been a major concern in the senior age
group (e.g., 65 years old and older) for several decades,
accounting for one-third of all senior hospital admissions.
Medication non-adherence is major contributing factor to health
care costs, especially among seniors with multiple chronic
conditions. Such seniors often are re-admitted to hospitals due to
not following medication regimen, and/or are advised to get
admitted to nursing homes, home care or assisted living facilities.
The major reasons for non-adherence include forgetting to take
their medications and getting confused with the complexity of the
medication regimen.
There are relatively few existing solutions available for ensuring
medication adherence in elderly patients. Examples include
electronic pill packs such as MedMinder.TM., electronic caps on
prescription containers, WisePill, and various smartphone apps.
Many of these adherence techniques either blink light emitting
diodes (LEDs) or produce a sound (e.g., beep) when it is time to
take medicine. A major issue with such designs is that the patient
needs to be in close proximity to see the lights or hear the beeps.
For example, the patient may miss a reminder if the medication
adherence device is in the kitchen while the patient is in the
bathroom.
Existing smartphone apps typically generate reminders based on SMS
text messaging. Smartphone usage, however, is less than 10% among
seniors 75 years and older, and is less than 5% among seniors 80
years and older. Furthermore, many seniors get confused navigating
through complex smartphone and app menus, and the apps usually
require complex setup through the host smartphone.
Clearly there is a need for a medication adherence solution that is
easy to use for seniors with multiple chronic conditions and
dexterity, vision and hearing issues.
SUMMARY OF THE INVENTION
The described embodiments include a Coordinated Care System (CCS),
which provides a medication adherence solution for patients who are
not completely capable of functioning independently. The CCS may
include a patient wearable device (PWD), a caregiver mobile app,
and a webserver. The webserver implements a coordinated care
platform that seamlessly integrates the PWD, the caregiver mobile
app, the webserver, and various third party providers (e.g.,
pharmacy API integration, Electronic Health Records software such
as EPIC). The coordinated care platform receives, processes, and
analyzes information from the PWD and the caregiver mobile app, to
provide status and predictive information to the caregiver and/or
other relevant healthcare providers.
In one embodiment, the PWD is a wrist-worn apparatus (or other form
such as pendant or armband) with a display screen, audio input and
output (e.g., speaker and microphone) and user input mechanisms
(e.g., buttons and/or touch-sensitive display screen). The
embodiment may include a single System on a Chip (SoC) architecture
that includes at least a chronometer (i.e., a watch), a cellular
telephone, a GPS receiver, a camera, and one or more sensors (e.g.,
gyroscope, accelerometer and/or magnetometer).
The described embodiments may provide a dedicated PWD that helps
elderly persons remember to take their medication on time, and also
assists in various daily life activities and help elderly people
remain independent. In the descriptions herein, the person wearing
the PWD may be referred to as a "patient"
In some embodiments, a PWD may provide visual and audible
medication reminders including a display of medication name, an
image of the medication, and dosage with voice instructions,
displayed text, or both. Such a PWD may provide substantially
constant connectivity to caregivers to allow the caregivers to
receive real time status notifications, and a mechanism to call
back and immediately check the patient's condition. The PWD may
also combine a cellular telephone, GPS tracking and fall detection
to provide automatic calling of a caregiver or healthcare
professional in an emergency situation.
In general, a PWD may help seniors live independently by assisting
them with medication reminders, Activities of Daily Life (ADL)
reminders (such as drinking water, eating food and exercising) and
enabling them to call their family whenever and wherever they need
help. The caregivers are notified with the status and alerts, so
they always know the status of the patient. The coordinated care
platform integrates personalized treatment programs in situations
where the PWD is provided to patients by the treatment providers
such as hospitals, pharmacies, or institutional care facilities
(such as independent or assisted living). In some embodiments, the
PWD may provide a periodic (e.g., daily) "are you okay?" (R U OK)
check to the patient.
The described embodiments may provide a single, integrated
coordinated care environment (also referred to herein as a
coordinated care platform). This coordinated care platform may (i)
manage the patient's medication regimen, (ii) coordinate
communication between the patient, and the caregiver, and other
relevant healthcare professionals, (iii) integrate with pharmacies
for automatic download of patient's medication regimen into the
PWD, (iv) coordinate updates of the medication regimen from the
caregiver and/or the pharmacy, (v) provided an executive dashboard
for hospitals to provide an aggregated real-time, current status of
the patient, and/or (vi) implement an Early Warning Predictive
Analysis (EWPA) of behavioral information captured and conveyed
from the PWD.
The behavioral information, based on engagement with the PWD (e.g.,
tapping the touch-sensitive display to acknowledge taking a
medication) and movement/position of the PWD (e.g., patient in a
reading position), may include a significant number of behavior
points (for example, 250 or more behavioral points). Once the
medication reminder sets off, every move by the patient such as
lifting the hand, reading the watch, tapping to acknowledge taking
the identified medicine, for example, is captured, and captured
data is conveyed through the coordinated care platform periodically
(e.g., every 15 minutes). The EWPA may compare the captured data to
previously received historical data to predict trending, and/or to
identify/predict an anomaly. Results of the EWPA may be documented
in summaries and/or detailed reports, which may be made available
for use by the caregivers, physicians and/or other healthcare
providers.
Adverse Drug Reactions (ADRs) are significant factors contributing
to emergency situations and increased healthcare costs. Some
embodiments may address ADRs by tracking Over-The-Counter (OTC)
medications (dosage and time taken) consumed ad hoc by the patient,
and evaluating the OTC medications with respect to the prescription
regimen being followed by the patient.
In some embodiments the software of the PWD is customizable, which
allows for personalized therapy with disease specific intervention
messages to the patients; this feature can be used for clinical
trials and motivational messages during a therapy along with
medication reminders. In some embodiments, the PWD may support
automatic setup, and real-time synchronization (i.e., update) of
prescription and filled medication information. The distinction
between prescription medication and filled medication is that
filled medication refers to the medication that was actually filled
based on the prescription. For example, the prescription may be for
a brand name medicine, while the filled medication may end up being
a generic form, in which case the color or dosage may be different
from what was actually prescribed. Also, sometimes not all
prescribed medications are actually filled, as some of the
prescribed can be "as needed" medication.
The described embodiments may provide what amounts to 24 hours per
day, 7 days per week (24.times.7) connectivity of the patient to
one or more caregivers, via a mobile app running on a platform such
as a smartphone or tablet at the caregiver's location. Such
24.times.7 connectivity facilitates (i) staying connected with an
elderly parent, (ii) getting missed medication and emergency
alerts, (iii) receiving a real-time feed of the current status of
the patient.
In an application of the described embodiments, the Coordinated
Care Platform may be used for Phase 3 clinical trials for new drug
application to study the effectiveness of a medication. In Phase 3
clinical trial the medication or treatment is given to large groups
of people to confirm its effectiveness, monitor side effects, and
collect information such as side effects, pain levels etc. to allow
the drug or treatment to be used safely. Currently the clinical
trials are monitored based on manual methods such as patients
logging into electronic diary and providing feedbacks on side
effects during in person visits. This method depends on patient's
ability to remember all the information to provide and also all
vitals are collected only during in person visits, which can be
very expensive. The PWD in the described embodiments can be used to
(i) convey medication reminders with instructions, (ii) record
feedback and side effects and (iii) collect vitals data from Blood
Glucose, Blood Pressure and heart rate monitors and provide the
data in real time. Use of the described embodiments may also help
to reduce patient "in-person" visits and reduce cost of clinical
trials.
In another application of the described embodiments, the
Coordinated Care Platform may be used for providing targeted
therapy for conditions such as ADHD, substance abuse (e.g., smoking
cessation), where in addition to a medication regimen, motivational
messages are also needed for continued treatment, as well as
improving effectiveness of the treatment. As part of the treatment
plan, the content of the messages and the schedule can be provided
to the coordinated care platform, and then the coordinated care
platform can download the messages and schedule to the PWD, to be
delivered to the patient at the appropriate time.
In one aspect, the invention may be a method of supporting
adherence, of a user, to a medication regimen. The method may
include, by a wearable computing device capable of stand-alone,
independent operation, the wearable computing device including a
display and an embedded cellular transceiver, storing received
medication information that characterizes one or more medications,
including at least identification information and dosing
information, associated with each of the one or more medications.
The method may further include displaying identification
information and dosing information associated with at least one of
the one or more medications. The method may also include generating
an initial notification indicating that the identification
information is being displayed. The method may also include
monitoring a response input instrumentality configured to receive a
response indicating that the at least one of the one or more
medications has been administered. The response input
instrumentality may include, for example, a touch-sensitive display
screen or a pushbutton on the PWD. The method may further include,
for each of the one or more medications displayed, generating a
medication status describing that either (i) the medication has
been administered or (ii) the response has not been received.
One embodiment may further include generating a reminder
notification when the medication status indicates that the response
has not been received, by re-displaying identification information
and dosing information associated with one or more medications
referenced in the reminder, generating a subsequent notification
indicating that the identification information is being displayed,
monitoring the response input instrumentality, and, for each of the
one or more medications referenced in the reminder, generating an
updated medication status describing that either (i) the medication
has been administered or (ii) the response has not been
received.
Another embodiment may further include repeating the reminder
notification periodically until (i) all of the at least one
medications have been administered, or (ii) a predetermined time
period has elapsed since generation of the initial
notification.
Another embodiment may further include providing the reminder
notification by playing a voice recording generated by a caregiver.
One embodiment may further include repeating the reminder
notification at intervals of about 10 minutes. Yet another
embodiment may further include notifying a caregiver after the
predetermined time period has elapsed.
One embodiment may further include detecting an incoming
communication from a caregiver, generating a notification of the
incoming communication, and monitoring an incoming communication
acceptance input instrumentality configured to receive an accept
response that causes the wearable computing device to accept the
incoming communication.
Another embodiment may further include automatically accepting an
incoming telephone call when the incoming communication acceptance
input instrumentality does not receive an accept response within a
predetermined time interval.
In one embodiment, the incoming communication may be an R_U_OK
alert.
Another embodiment may further include automatically connecting to
a healthcare device, receiving health information from the
healthcare device, storing the health information and forwarding
the health care information to a caregiver. Such a healthcare
device may include, for example, a blood glucose meter, a blood
pressure measuring device, a weight measurement device (i.e., a
scale), or a pulse measuring device. Other such vitals measuring
devices may also provide data to the wearable computing device.
In one embodiment, the identification information and dosing
information include an image associated with the at least one of
the one or more medications, the image including a depiction of a
vehicle in which the medication is administered, the vehicle
instantiated a number of times corresponding to the dosing
information. The vehicle may include a pill, a capsule, or other
such vehicles known in the art for conveying medication to a
patient.
One embodiment may further include activating a camera integrated
into the wearable computing device, the activating occurring in
response to a camera stimulus. In one embodiment, the camera
stimulus includes one or more of (i) a camera activation
communication received from a caregiver, (ii) the medication status
indicates that the response has not been received.
One embodiment may further include synchronizing the wearable
computing device by receiving and storing updated medication
information. Another embodiment may further include determining
geolocation of the wearable computing device in relation to a
predetermined area, and notifying a caregiver when the geolocation
is determined to be outside of the predetermined area.
One embodiment may further include detecting that the user has
fallen by analyzing an output of one or more accelerometers
embedded in the wearable computing device, and notifying a
caregiver upon detecting that the user has fallen.
In another aspect, the invention may be a method of monitoring and
analyzing behavior of a patient, and providing predictive analytics
therefrom, comprising, by a coordinated care system including a
patient wearable device (PWD), a caregiver application and a
webserver, gathering, by the PWD, behavioral data associated with
the patient and conveying the behavioral data from the PWD to the
web server. The method may further include analyzing, by the
webserver, the behavioral data to determine one or more behavioral
trends. The method may further includes comparing the determined
behavioral trends to one or more historical behavioral information
to identify a behavioral anomaly. The method may further include
sending an alert to the caregiver application when a behavioral
anomaly is identified.
In one embodiment, analyzing the behavioral data includes
determining a compliance index and an engagement index using an
exponential weighted moving average.
In another embodiment, the behavioral data includes one or more of
position information based on sensor outputs, response information
based on reaction of the patient to reminder information, and
vitals information (such as weight, heart rate, blood pressure,
blood glucose level, for example) collected by the PWD.
Another embodiment further includes determining, based on the
behavioral data, that an emergency event has occurred, and in
response to the determining, notifying appropriate emergency
response personnel.
BRIEF DESCRIPTION OF THE DRAWINGS
The foregoing will be apparent from the following more particular
description of example embodiments of the invention, as illustrated
in the accompanying drawings in which like reference characters
refer to the same parts throughout the different views. The
drawings are not necessarily to scale, emphasis instead being
placed upon illustrating embodiments of the present invention.
FIG. 1 shows an example embodiment of a coordinated care system
according to the invention.
FIG. 2 shows a block diagram view of an example embodiment of a PWD
according to the invention.
FIG. 3 shows example usage of an embodiment of the PWD shown in
FIG. 2.
FIG. 4 illustrates a functional block diagram of an example
embodiment of the PWD shown in FIG. 2.
FIG. 5A shows a flow diagram of an example medication reminder
process according to the invention.
FIG. 5B shows an outline of an example medication reminder process
according to the invention.
FIG. 6 shows example screen shots of a PWD displaying medication
reminder and battery charging reminders.
FIG. 7 illustrates an outline of an example setup procedure of a
PWD according to an embodiment of the invention.
FIG. 8 illustrates an example graph that displays potential actions
with respect to combinations of compliance index values and
engagement index values.
FIG. 9 illustrates an example prediction data pipeline of the CCS
predictive analytics platform according to the described
embodiments
DETAILED DESCRIPTION OF THE INVENTION
A description of example embodiments of the invention follows.
The teachings of all patents, published applications and references
cited herein are incorporated by reference in their entirety.
An example embodiment of a Coordinated Care System (CCS) 100, shown
in FIG. 1, includes a patient wearable device (PWD) 102, a
caregiver mobile app 104, and a webserver 106. The caregiver mobile
app 104 runs on a commercially-available platform such as a
smartphone or electronic tablet. The webserver 106, which runs on a
commercially-available platform such as a computer system or
workstation, implements a coordinated care platform that seamlessly
integrates the PWD, the caregiver mobile app, the webserver, and
various third party providers (such as pharmacies and healthcare
providers). The webserver 106 includes a web frontend, a server
backend, a backend database and an application program interface
(API). Target users for each of these components are shown in Table
I.
TABLE-US-00001 TABLE I Component Target User Patient Wearable
Patient Device (PWD) Smartphone or Tablet Nurse/Caregivers, Family
members, application (Caregiver Physicians,Medical Coordinators,
Mobile App) among others Webserver, backend Family members,
Physicians, Hospitals, application with a Pharmacy, Caregivers,
among others web frontend
The software stack technologies for the CCS components and
sub-components are shown in Table II.
TABLE-US-00002 TABLE II Component Technology Target User
Hosted/Runs On PWD Android Patient PWD hardware Mobile App Android,
Caregiver Caregiver's mobile iOS device (smartphone or tablet)
Server Java, CSS Admins Google App Engine Backend Hibernate Web
HTML5, Medical Staff, Google App Engine Frontend CSS3 Admin,
Physicians Backend Google CSS Admins Google App Engine Database
CloudSQL API Java Mobile apps, Google App Engine Frontend, Pharmacy
integration
Patient Wearable Device (PWD)
As described below, the PWD 102 supports cellular communications,
which enables connection to the webserver 106 through mobile
Internet. A SIM card within the PWD 102 is associated with a voice
and data plan activated for calls with caregivers and selected
family members. As the PWD connects to the webserver 106 using
cellular data service, there is no need for external/additional
controller, hub or a smartphone in the patient's home or carried by
the patient for the PWD to connect to the internet. The PWD is
fully functional by itself, and can therefore be used anywhere
inside or outside the home, as long as cellular connectivity is
available.
The PWD is water resistant with Ingress Protection (IP) IP67 or
IP68 certification, so it can be used in the shower, kitchen or
rain, even while travelling or sleeping. The first digit 6 in IP67
implies that it is dust proof and the second digit 7 implies water
resistant and 8 implies water proof. These certifications are
provided International Electrotechnical Commission (IEC) based on
tests, and show that the device has proper sealing and enclosures
to protect from dust or water.
Patients using the PWD are expected to be seniors typically 70
years or older. In one embodiment, the PWD and its patient
interface may be developed with characteristics specifically
targeted for this age group to accommodate their attendant vision
and dexterity issues.
As shown in FIG. 2, an example embodiment of the PWD 102 may
include a Central Processing Unit (CPU) 202, Graphical Processor
Unit (GPU) 204, and a touch-sensitive display 205. The display 205
is capable of presenting configurable color schemes that optimizes
the visual effect experienced by the patient. In some embodiments,
the screen is relatively large (e.g., 1.5 inches by 1.5 inches).
Displayed text font is generally large and bold so that elderly
patients will have an easier time reading the text. As described
herein, patients may respond to medication reminders by tapping the
display screen. For such responses, the user can tap the display
screen at any point on the surface of the display screen, rather
than a specific sub-region of the screen.
The PWD 102 may also include a cellular modem module 206,
supporting, for example LTE protocols, 3GPP protocols, GSM/EDGE
protocols, combinations thereof, or other wireless wide area
communication protocols. The cellular modem 206 coordinates with an
RF interface 208 for connection to a cellular wireless network.
The PWD may also include a Global Navigation Satellite
System/Global Positioning System (GNSS/GPS) module 210 to determine
the position of the PWD 102 based on satellite communication,
although other positioning systems known in the art may also be
used. The PWD 102 may also include audio and video CODECs 212 for
encoding and decoding voice and video information. The PWD 102 may
also include a multi-media processor 214 for supporting PWD
activities such as voice calls, PWD activation, video calls,
playing video, among others.
A sensor hub 216 manages and collects data from one or more sensors
(e.g., an accelerometer, a magnetometer, a motion sensor, a
gyroscope, a stress/strain sensor, and a pressure sensor, among
other types of sensors know in the art). The sensor hub 216 samples
data from the one or more sensors and stores the collected
information in memory 232. Software or firmware, stored in memory
232 and executed by the CPU 202, may cause data to be extracted
from the memory 232, assembled into a predetermined format, and
sent to the back-end server within the CCS for various types of
processing, such as the Early Warning Predictive Analysis (EWPA)
described herein. The sensor hub 216 may be implemented by a
processor running instruction code, a collection of analog and
digital circuitry, or a combination of both.
The PWD memory 232 may also be used to implement a database to
store medication regimen, instructions and messages. In an example
embodiment, the following attributes may be stored for each
medicine: Timeslot (i.e., times when the patient takes Breakfast,
lunch, snack, dinner, wake up time and bed time) Medicine name
Dosage (quantity--e.g., two tablets, one capsule, one-half fluid
ounce) To be taken before or after food Days of week (e.g., every
day, every other day, once per week) Image information for
depicting/displaying the medicine Voice instructions
In one example embodiment, the memory 232 is approximately 4 GB.
The memory 232 may be implemented in a single memory component, or
distributed across two or more memory components. The information
that needs to be stored is less than 1 GB including, the medicine
image information. Having a local storage in the PWD 102 also helps
in delivering the reminders consistently while reducing the risk of
failure due to a communication outage between the PWD 102 and the
webserver 106. Thus, once the medication details are downloaded to
the PWD 102, the PWD 102 may operate substantially autonomously,
without requiring constant supporting communication with the
webserver 106.
The main functionality of the PWD is implemented with a PWD app. In
one embodiment, the PWD app is instruction code, stored in the
memory 232 and executed by CPU 202. In one embodiment, the PWD app
is a dedicated (kiosk mode with root access) application using
Android OS and Android application framework. The term "dedicated"
as used herein implies that there are no other applications running
on the watch and the user need not navigate through any menus or
apps.
To send "medication taken" status, alerts, behavioral data points,
and other such information, however, the PWD 102 does communicate
with the webserver 106 to some extent. The watch communicates using
JAVA REST APIs to the webserver 106. Such communications may be
done in the background, without interfering with the PWD's regular
functions such as providing medication reminders and collecting
behavior data points. In some embodiments, the webserver 106 may
periodically check the health of the PWD 102, and the status of its
communication capabilities, by executing diagnostic tests. Such
diagnostic tests may be executed external to the PWD 102 (e.g., by
sending test messages from the webserver 106 to the PWD and
requiring specific responses), internal to the PWD (e.g., test
instructions stored within the PWD memory 232 and executed by a PWD
processor).
In addition, the PWD 102 supports a subscriber identification
module (SIM) 218 for designating phone and data service
subscriptions, a near-field communication (NFC) module 220 for
reading and processing NFC tags, one or more audio drivers 221 and
speakers 222 for producing audible sounds, one or more microphones
224 for receiving audible sound and transducing the sound into an
electrical signal, a vibration module 226 for providing haptic
notifications, one or more cameras 228 and a real time clock (RTC)
230 that provides a chronometer function for date and time of day
(i.e., the PWD functions as a wrist watch). The PWD may further
include local communications elements such as a Bluetooth
transceiver 234 and a WiFi transceiver 236. Other communications
protocols (e.g., Zigbee) may also be used.
The PWD 102 may also include an energy storage element (e.g., a
lithium-ion battery or other re-chargeable battery type) and
associated recharging circuitry 238 for efficiently recharging the
energy storage element. Some embodiments provide for an electrical
connection of the PWD 102 to a charging unit through electrically
conducting contacts. Other embodiments provide for inductive
coupling of the PWD 102 to a charging unit.
Information from the patient (e.g., responses to medication
reminders) may be provided to the PWD 102 through the
touch-sensitive display screen 205, or through other input/output
(I/O) components 240, for example push-buttons, sliding switches,
and other I/O components known in the art.
In one embodiment, all of the components of the PWD 102 may be
implemented by a commercially-available single System On a Chip
(SoC) solution. In other embodiments, the functionality may be
implemented by a custom-designed device (i.e., ASIC) or chipset, or
as combinations of both a commercially-available SoC and custom
designed components.
Interaction between the PWD and the patient (i.e., the PWD user
interface) is simple. There is no patient-involved setup required
for the PWD, as the set up will be completed by the caregiver using
the caregiver mobile app (described later herein) or by the
pharmacist once the pharmacy is integrated with the CCS.
Activation of the PWD may also be done by the caregiver using the
caregiver mobile app. The PWD downloads the complete medication
regimen when the PWD is powered on, and the patient can start
wearing the PWD once activated. Typically the only action required
by the patient is to acknowledge an instruction by performing a
single tap anywhere on the display screen. Many aspects of the user
interface, including the voice, display and the vibrations, can be
customized remotely from caregiver mobile app.
An example embodiment of a PWD constructed according to the
invention, with various screenshots, features and PWD app
architecture, is shown in FIG. 3. As shown, the PWD has a local
database where the medication regimen, images, and voice files are
stored. The local processor activates reminders based on a
scheduler, and stores responses from the patient. When the
reminders are completed, the PWD sends an alert or log to the
coordinated care platform at the webserver, including all the
medications taken and missed, along with the time sequence
information associated with the responses. The PWD may communicate
with the coordinated care platform via JAVA REST APIs. For example
a missed medication alert may have following information:
AlertID:
missedTimestamp:
patientID:
missedTimeSlot:
missedBeforeOrAfterFood:
missedMedicationIDs: [{medication detail 1}, {medication detail 2},
{medication detail 3}
Medication details may include, for example, medicine name and
dosage, among other descriptive information. Since patientID is in
the form of a control number, no name will be sent across.
FIG. 4 illustrates a functional block diagram of an example
embodiment of the PWD 102. An example embodiment of a PWD 102 may
provide one or more of the functionalities described below.
Displaying Time of Day--This is the default image displayed by the
display screen 105. Time of day (TOD), as described herein, may
include hours, minutes, seconds, month, day and year. The PWD may
also speak the TOD when the patient taps the display screen (or
otherwise indicates a desire to receive an audible version of the
TOD). In some embodiments, the patient is able to modify the TOD
setting only, with no other menus or navigation requirements.
Providing cellular telephone service--The PWD includes a cellular
telephone that supports incoming and outgoing calls. In one
embodiment, the PWD cellular telephone is implemented with a
commercially available chip set such as the MediaTek 6753 or
Qualcomm Snapdragon 2100. These chipsets may support one or more of
3GPP, LTE and GSM/EDGE protocols.
A SIM card is also provided for phone/data service. The SIM card is
configured to include a set of preprogrammed telephone numbers and
to accept calls only from those numbers. The PWD manages the
incoming calls and outgoing calls from the PWD. In one embodiment,
the patient is able to call pre-programmed family members by simply
pushing a button on the side of the PWD. In one embodiment, a
caregiver can program a sequence of telephone numbers and
sequentially call each subsequent telephone number if the previous
telephone number is not answered.
Providing Medication Reminders--With trio feedback
(audio/visual/haptic), the PWD 100 may vibrate, beep and display a
"Reminder" screen according to the patient's medication dosing
schedule, although other embodiments may present alternative
haptic, audio and/or visual stimuli. In response to an
acknowledgement from the patient (i.e., the patient tapping the
touch-sensitive display screen 105 or pressing a button, depending
on the embodiment used), medications are displayed on the display
screen 105 one by one, with brand and medication name, medication
dosage and strength, an image of the medication, and a voice prompt
providing instructions associated with the medication (e.g., how
the medication should be taken). If a pill pack with weekly slots
is used instead of individual pill bottles, then the image and
dosage instructions for the slot may be displayed.
In some embodiments, the voice prompt is customizable to any family
member voice. Some patients, especially Alzheimer or Dementia
patients, are accustomed to only a family member's voice, so that
instructions presented in the family member's voice helps the
patient in taking their medicines.
Once the medication is taken, the patient is required to confirm
that the medicine has been taken by tapping on the display screen.
If the patient does not confirm taking the medication, the PWD will
provide three additional reminders, snooze for 10 minutes and
repeat the above-described reminding process. These repeated
reminders will continue for one hour (or per physician's'
instructions). If even one medication is missed, a missed
medication alert listing the missed medicines is generated and sent
to the coordinated care platform at the webserver. The flowchart
shown in FIG. 5 describes one embodiment of an example medication
reminder process.
FIG. 6 provides the example screen shots of a PWD displaying
medication reminder and battery charging reminders. An initial
reminder screen 602 initially alerts the patient to a medication
dosing session, and begins the medication reminder sequence of
events. In a first medication screen 604, the PWD shows the
medication name (Seroquel), an image of the pill (front and back),
and dosage (400 mg--take 2). An alternative medication screen 606
shows a door from a pill pack (Thursday noon) and dosage (4
tablets). Reminder screen 608 shows the PWD displaying another
reminder, this time indicating a family member (i.e., daughter).
Alert screen 610 shows the PWD displaying a "low battery" alert,
requesting the patient to recharge the PWD (i.e., "Charge Me!"),
and notification screen 612 notifies the patient that the PWD has
been fully charged and is ready to wear again.
When the patient taps the screen, timestamps associated with the
taps are recorded and sent to the predictive analytics engine. If
the intervals between taps are too short, an alert is raised to the
caregiver to check if the patient has taken the scheduled
medicines.
Receiving Reminders from Caregivers--Caregivers may be alerted when
a patient does not respond to one or more of the scheduled
medication reminders. The caregiver can use their mobile app to
submit a subsequent reminder to the patient, call the patient, or
both.
Providing Activities of Daily Living (ADL) Reminders--Some
embodiments may provide reminders to the patient to perform daily
activities. Seniors often forget to take their meals on time, drink
water or exercise. Lack of water causes dehydration. Exercise is
very important for fall prevention, lowering the risk of heart
diseases, injury, and elevated blood pressure. Lack of exercise may
also lead to increased joint pain. Further, some medications
require specific instructions on when to eat food or drink
additional water. In some assisted living facilities or independent
living communities, the number of nurses to patient ratio is often
very small and in general there is not enough nurses to personally
remind patients to remind their activities.
The described embodiments provide a convenient way to remind the
patient to perform ADLs. These reminder messages can be stored
during setup while entering the patient schedule, and the PWD will
display those messages on time and get a confirmation that the
messages were read. This confirmation may be logged in an
activities log, and caregivers can be notified through a status
feed. The resulting ADL performance information can be added to
behavioral information to be evaluated via the EWPA. For example, a
patient that misses meals with increasing frequency may be
suffering from worsening memory loss.
Providing Motivational Messages--An embodiment may provide a
mechanism where any specialized motivational messages or
instructions can be provided to the patient based on a therapy.
These messages are displayed based on a schedule configured in the
database. Embodiments may include daily motivational messages (with
spoken voice playback) encouraging the patient to adhere to their
medication schedule and/or treatment program
Conducting Real-time Check-In--Real-time, bidirectional
communication to survey or check on a patient, such as "Are you
doing okay today?" (R U OK) or "How many did you smoke today?"
The "R U OK" communication is designed for seniors who live alone
at their homes. The PWD displays the "R U OK" message. If the
patient does not answer, an alert is sent to the caregiver or other
healthcare provider. The alert can contain additional information
including if the PWD is worn, information on the patient's recent
activity prior to the alert, and other engagement details.
Requesting HELP--The patient can tap on the PWD 100 to send a HELP
message to caregivers, thereby requesting a phone call-back to the
PWD 100 (or to an alternative line in the home). Some embodiments
of the PWD 100 may require alternative or additional actions from
the patient to reduce or avoid false alarms (e.g., double tap or
triple tap, or tap followed by a button push).
Call Caregiver/family--By pressing the side button (for example),
the patient can initiate a telephone call to the family (or first
level responder), through the PWD cellular module 106 and RF
interface 108. As with the request for HELP, some embodiments of
the PWD 100 may require alternative or additional actions from the
patient to reduce or avoid false alarms.
Incoming Calls from caregiver or a set of preprogrammed
numbers--The PWD allows authorized numbers (such as caregivers) to
call the PWD directly. All other incoming or outgoing calls are
blocked to limit potential misuse. Limiting incoming telephone
calls may also serve as a comfort factor for the patient, since the
patient can feel secure that he or she will not be bothered by
annoying telemarketing callers or malicious/fraudulent callers.
Alarms for Low and Full Battery--The PWD lets the patient know, via
vibrate, alarm and/or voice prompts, when the energy storage
element (e.g. battery) is low, and that the patient should connect
to a charger, or when charging is complete
Live GPS Tracking & Directions--A geo-fence may be defined
(e.g., for early Alzheimer patients) such that if the patient
wanders away from his/her home, the PWD detects, using built-in
GPS, that the patient is beyond the defined geo-fencing region and
alerts their caregiver with accurate location information. In one
embodiment, the PWD evaluates its position periodically (e.g.,
every 15 minutes), and if the PWD moves beyond an established
geo-fence, the PWD begins continuous tracking and sends alert to
the caregiver along with location information. The tracking
continues until the PWD is back within Geo fence range. If the
patient plans to intentionally move beyond the geo-fence boundaries
(e.g., for a visit to a healthcare provider), the caregiver can
disable GPS tracking through the caregiver mobile app until the
patient returns to the within the established geo-fence region.
Fall Detection and Emergency Calling--The PWD may detect, using
positional sensor input, when a patient has fallen. In response to
a detected fall, the PWD may initiate an emergency call to a local
call center or to a caregiver. If the patient is able to
communicate during the emergency call, the patient may be able to
provide information as to subsequent steps. If the patient is not
able to communicate, the local call center or caregiver follows a
predetermined response protocol for sending help to the
patient.
Vitals Collection: In some embodiments, the PWD provides a hub for
collecting "vitals" information from devices such as blood glucose
monitor, blood pressure monitor and heart rate monitor, and sends
the collected data to the backend application of the webserver to
report to physicians, hospitals or other healthcare providers. In
some embodiments, the information may be collected periodically
based on a frequency set in the database.
Patient Profile Manager--This module of the PWD app initializes the
PWD database with patient profile information including medication
regimen, reminders and other details. Once operational, any updates
to the profile information is received and processed by this
module.
Battery Alarms--The PWD periodically monitors the state of charge
of its energy storage element (i.e., the PWD battery). When the
levels reach below a predetermined threshold (e.g., 20%), the PWD
provides an audible indication (e.g., beeps) along with a display
of the current battery state of charge and a voice instruction to
begin recharging the PWD. If the patient does not tap to confirm
the recharging notice, the reminders may again be presented
periodically, and an alert may also be sent to the caregiver. When
the PWD is fully charged, or the PWD recharging source is removed
from the PWD, a reminder is sent to the patient instructing him or
her to once again wear the PWD. If the PWD is not worn after a few
reminders, an alert is sent to the caregiver and the caregiver can
call the patient from mobile app or take other action with the
patient.
In other embodiments, the PWD may be a commercially available
smartwatch, modified to provide the functionality described herein.
The commercially available smartwatch may be modified either
physically or by having specific apps written for it.
Caregiver Mobile App
The smartphone or tablet application is available for the
caregivers and family members, among others, to be able to monitor
and communicate in real-time with the patient through the PWD. It
should be understood that the concepts described herein may be
implemented with alternative components, e.g., a pendant or armband
device instead of (or in addition to) the smartwatch. Similarly,
the smartphone or tablet application may be implemented on any
other platforms suitable for performing the functions and/or
capabilities of the described embodiments.
Initial Setup
When the caregiver buys or otherwise initially acquires a PWD for
use by a patient, the caregiver can download a caregiver app into
their smartphone or tablet from an app store or other source and
create an account. The caregiver may then register the patient
through the caregiver app, and either enter medication details
manually, or download prescription data from an associated
pharmacy, provided the caregiver can provide the proper
authorization and consent.
When the caregiver enters the medication details manually, they
will be prompted to take pictures of the medicine and upload to the
coordinated care platform. The caregiver app sequentially guides
the caregiver to complete the registration. A caregiver can
register and manage multiple patients (e.g., both parents). FIG. 7
illustrates an outline of an example setup procedure of a PWD
according to an embodiment. Example caregiver app functionalities,
which may be performed by the caregiver app and/or the CSS of the
described embodiments, are presented below.
Status feed: The caregivers may receive regular feeds regarding the
patient's status, including medication log and other activities
such as food, water, exercise, deviation from normal behavior, and
GPS tracking.
Alerts and notifications: The caregivers may receive one or more of
the following alerts and notifications, which are intended as
illustrative examples and are not meant to be limiting: (i) Missed
Medication alert, (ii) GPS alert, (iii) Low battery alert, (iv)
Watch idle/not worn, (v) Early Warning alerts, (vi) R U OK alert,
and (vii) ADL alerts.
When the alerts are received, the caregiver can log in to the
coordinated care platform through the caregiver mobile app and
review details of the alert. If the review indicates an issue, the
caregiver can either send text messages to PWD, or call the PWD,
from the caregiver app.
Dashboard: the coordinated caregivers get reports on a weekly basis
with details including medication adherence, activities, and R U OK
responses. Following reports are provided via dashboard:
Medication adherence logs: number of missed medications in the past
week, number of alerts generated for missed medications, number of
as needed medications (such as ibuprofen or over the counter
medicines), medication adherence trending.
ADL logs: number of missed reminders for food, exercise and
water.
R U OK logs: Number of missed R U OK response alerts.
Deviation logs: Deviation or change from normal
behavior/activities.
GPS logs: Number of GPS alerts.
Vitals logs: Number of times vitals (e.g., weight, body
temperature, blood pressure, heart rate, blood glucose)
collected.
Telephone logs: Phone call History and details.
Emergency logs: Number and characteristics of emergency alerts
issued.
Webserver
The webserver 106, shown in FIG. 1, implements the coordinated care
platform described herein. The webserver 106 includes a backend
application and a web-based front end, each of which runs on a
processor-based computer system suitable for hosting server
software.
The backend application portion of the webserver processes the data
sent from the PWD, compares the currently sent PWD data to
historical PWD data, and computes one or more trending occurrences
based on the comparison. If the backend application detects changes
in the patient behavior, with respect to the patient's historical
data, the backend generates alerts to caregiver mobile app or
hospital dashboard. The coordinated care platform uses JAVA REST
API to communicate with web frontend, the pharmacy, the PWD and the
caregiver app. The coordinated care platform also uses Google Cloud
Messaging (GCM) Service to send messages and alerts. Some
embodiments of the coordinated care platform may provide one or
more of the following functions:
PWD Management--Includes (i) coordination of the assignment and
initialization of PWDs for patients, (ii) registration of PWDs and
patients, (iii) the setup by caregivers, (iv) downloading
medication information, (v) receiving alerts, (vi) assembling data
for the real-time behavioral points that were collected by and sent
from the PWD, among others. As described herein, in an example
embodiment the number of real-time behavioral points collected by
the PWD may exceed 250.
Interface to Caregiver App--Management of caregiver accounts,
communication with caregiver app to notify of real-time status
information, missed medication alerts, and emergency alerts, among
other alerts.
Early Warning Predictive Analytics (EWPA)--Periodical Collection,
real-time processing of behavioral points from seniors based on the
engagement of the PWD, generation of alerts if anomalies detected
from the normal behavioral pattern
Storage--some embodiments use a combination of relational (i.e.,
Structured Query Language--SQL) and Non Structured Query Language
(NoSQL) database implementations to store patient profiles,
medication regimen and file system to store images of all
medications and behavioral data points collected from the
patient.
Security--The implementation and maintenance of authentication,
authorization and consent forms with respect to all components of
the CCS, implementation of a HIPAA compliance platform, and
executing verification procedures for caregiver login. Digital
signatures such as XML DigSIg or JWT or images will be used for
verifying authorization. Other security techniques known in the art
may also be used.
Pharmacy Interface--Pharmacy and Electronic Health Records (EHR)
interface for one-touch download of prescription information and
setup of the PWD. The pharmacies have application programming
interfaces (API) to access prescription information. The Fast
Healthcare Interoperability Resources (FHIR) describes REST API to
get prescription information from EHR. The coordinated care
platform may be integrated with pharmacy APIs, so that a pharmacist
can download the prescription information, from their
pharmacy-based software to the PWD, by clicking "PWD setup" from an
application available to the pharmacist (i.e., the pharmacist's
access point into the CCS). The pharmacist's app may be a web-based
application or an application that is downloaded to the
pharmacist's computer platform. Alternatively, as described herein,
a caregiver using a caregiver mobile app may coordinate, through
the pharmacy, downloading of prescription information to the
PWD.
Web Front End--Provides a web-based interface for caregiver login,
registration, setup, weekly and on-demand dashboards, to name a few
examples.
Real-Time Instrumentation Data and Predictive Analytics
The described embodiments merge the ambulatory assessment and
intervention, Ecological Momentary Assessment (EMA) and Ecological
Momentary Interventions (EMI), into a single device (PWD). The PWD,
coupled with advanced sensors, enables meticulous real-time
monitoring to measure a patient's engagement with their treatment
and adherence to their medications, by capturing the following
instrumentation variables: Battery level and time when watch was
connected to, and taken out of, a charger. Watch movement--sensor
output, such as accelerometer, gyroscope, and GPS values, captured
periodically, for example every 15 minutes. Other sensor output may
also be collected Time when a reminder is displayed and time when
the reminder is "auto-snoozed" and/or when patient acknowledges
taking the medication associated with the reminder. Time when
medication is displayed and time when it is confirmed as taken or
missed. Time when a motivational message or survey question is
displayed and time when responded. Time when patient requests for a
HELP callback, when the call is received and its duration. Time
when patient initiated or received a call to/from the medical staff
and its duration
Each medication reminder may produce a significant number of
behavioral data points to be collected. Medication regimens with
several medications therefore multiply those data point quantities.
Example medication reminder data points may include (i) time when
reminder was displayed, (ii) time when the screen was tapped to
listen to the medicine list, (iii) time when the message/display
was read, (iv) duration the display was presented, (v) time when
the confirmation was tapped, (vi) time when reminder was snoozed,
(vii) time when the second reminder was displayed, (viii) time when
second time the reminder was auto-snoozed, (ix) time when the third
reminder was displayed, (x) time when the alert was sent to the
backend
Movement of the patient wearing the PWD may also result in a
substantial number of behavioral data points to be collected. For
example, movement may produce the following data to be collected:
(i) time when the patient wears the watch in the morning, (ii) time
or times when there is a change in movement, (iii) start time for
inactive period (rest, sleep), (iv) end time for inactive period,
(v) start time for charging the PWD, (vi) end time for charging the
PWD, (vii) start time for idle period (i.e., no movement), (viii)
end time for idle period, (ix) start time for rapid movement (e.g.,
exercise), (x) end time for rapid movement (exercise), (xi) start
time when the PWD is not worn by the patient (but is not on the
charger), and (xii) time when the PWD is worn again.
Similar data points may be collected with respect to ADL reminders,
RU OK responses, PWD cellular telephone usage, among others.
The coordinated care platform can analyze the collected data points
to predict potential outcomes. For example, certain combinations of
points may be associated with potentially dangerous situations.
Examples of such combination of points that can trigger potential
danger include: (i) long idle time during an active period, (ii)
missing a medication reminder and/or PWD not being worn or being
idle during medicine or meal time, (iii) long idle period after
rest or nap time, and (iv) unusually long rapid movement
period.
Connected Healthcare Platform
The CCS described herein may be used to provide a connected
healthcare platform. This connected healthcare platform may connect
seniors (i.e., elderly patients), in real-time, with caregivers
such as family, physicians, providers and pharmacy. The CCS may
alert the caregivers to any anomaly in seniors' health or behavior.
The PWD features the following aspects related to a connected
healthcare platform:
Medication Reminders & Alerts--to take meds on time, shows
medication name, image and dosage with voice instructions, with
alerts to clinic on any missed medications
Real-time Customizable Interventions--Occasional questions multiple
times a day (such as YES/NO, side effects, pain levels) to assess
patient's physical and emotional health
Patient Safety--fall detection, one-touch phone call to clinic, and
water resistant (can be used in the shower or the kitchen)
Early Warning System--collects multiple real-time behavioral
engagement and compliance data points every day to assess and alert
on any unusual behavior and adverse events.
Real-time Vitals Monitoring--Bluetooth hub to connect to Bluetooth
based vital sign monitors such as heart-rate, blood pressure, blood
glucose, weight-scale
Activity Monitoring--Passively track behavior and environmental
exposures in real-time with minimal burden on participants
Connected Ecosystem--to clinicians, emergency numbers and trial
management system.
Big Data Insights in Real-time--the described embodiments include
machine-learning predictive analytics that provide customizable
real-time trends, insights and outliers into patient's daily and
historical engagement/compliance, health/activity status and
response to therapy.
Privacy & Security--HIPAA compliant, encrypted, de-identified
database.
Predictive Analytics Early Warning System
As described herein, the PWD may collect a substantial number of
real-time behavioral data points every day, to be analyzed by an
Early Warning Predictive Analytics (EWPA) machine learning system
within the coordinated care platform to alert the caregiver (via
our mobile app) in case of any behavioral anomaly. The real-time
measurements from the PWD are stored in the backend database. An
EWPA predictive algorithm (and its associated model), based on
weighted distribution, evaluates historical data trends, predicts
the behavior of the patient based on a predetermined number of
prior responses and assesses the trending of the patient's
compliance. If the adherence decreases (or is tending towards
decreased adherence), then a trigger is generated to send alerts to
the caregiver to nudge the patient get back on track. The real-time
data can be used to predict the patient's adherence behavior and
alert nurses or physicians proactively for any timely intervention.
Predictive analytics described herein use the real-time data points
to calculate a "compliance index" and "engagement index" for each
patient.
In some embodiments, the compliance index may be a function of the
following values:
Idle time--(During day-time or active time) Use accelerometer
readings from the PWD (when not charging) every 15 mins to check if
the PWD is moving, or idle (e.g., left on a desk)--assumption is
that if the PWD is moving, it is being used and not idle.
Number of Auto-snooze requested--Determine the number of times the
patient requested auto-snooze (i.e., activated an appropriate
response input to a particular reminder auto-snoozed (means no
response) in a day, then aggregate over all days within a
predetermined time interval.
Number of Alerts issued--As described in more detail elsewhere
herein, if the patient does not respond to a medication reminder
within a predetermined time interval, an alert is raised for missed
medication. Determine the number of alerts issued within a
predetermined time interval.
Charging time--Determine the amount of time the PWD was left on
charger--The PWD notifies the patient (i.e., produces beeps) to
indicate low charge, full charge, or other charge states, with
voice announcement to take it out and wear
Battery dead time--Determine the amount of time the PWD remained
without charge until it was fully charged and restarted.
In some embodiments, the engagement index may be a function of the
following values:
Medication Reminder--Determine the amount of time elapsed between
when the reminder was issued (e.g., displayed or sounded) to when
the patient acknowledged the reminder. This determination may
indicate if the patient is just dismissing the reminder without
compliance, or actually taking the medication and then dismissing
it.
Daily Motivational Messages and other Reminders--This relates to
ADL messages or R U OK messages. Determine the amount of time
elapsed between when the message was displayed to when the patient
acknowledged the message. This determination may be used to gauge
the patient's level of interest in reading the message.
Proof of reading--evaluate PWD position, based on sensor output
(e.g., accelerometer, magnetometer, gyroscope, etc.) to see if the
PWD has moved to a position or level that evidences the patient
reading the PWD display.
FIG. 8 illustrates an example graph that displays potential actions
with respect to combinations of compliance index values and
engagement index values. For example, a combination of the
compliance and engagement index falling in the darkly shaded region
802 indicates that it is time to generate an alert.
In one embodiment, the EWPA prediction uses Exponential Weighted
Moving Average (EMA) formula for determining indices. The indices
are determined every day. Then decreasing weighting is assigned to
each datum exponentially, but never becoming zero. Continued
decreasing of the EMA value may be indicative of a tendency to
deviate from normal behavior.
In one embodiment, the EMA for a series Y may be calculated
recursively: S.sub.1=Y.sub.1, and for t>1,
S.sub.t=[.alpha.Y.sub.t][(1-.alpha.)S.sub.(t-1)]
The coefficient .alpha., a constant smoothing factor between 0 and
1, represents the degree of weighting decrease. A higher .alpha.
discounts older observations faster. Y.sub.t is the value at a time
period t, and S.sub.t is the value of the EMA at any time t.
The EWPA prediction uses these indices (and its historical trends)
to learn the patient's behavior and help predict and personalize
the intervention during follow-up phone calls or in-person
meetings.
In an illustrative example, assume that the compliance index
maximum value is 10 and the engagement index maximum value is 10.
If the collected behavioral data indicates that the patient takes
their medicine each time and no alert is raised then the value of
10 is assigned for compliance. If the data indicates that the
patient looks at the PWD, reads or hears the instructions, then the
engagement index is assigned a value of 10. If the patient scores
10 for both it is great. If, however the patient starts missing
medicines and alerts need to be raised, then the compliance value
will be some value less than 10, which triggers an alert to be sent
to the caregiver. The caregiver may respond to the alert, for
example by calling or visiting the patient to determine possible
remedial actions.
In another illustrative example, the patient takes their medicine
at the first reminder prompt. After a while, the patient's
responsiveness slips to require a second reminder (engagement index
will be lower) each time. As more time goes by, the patient's
responsiveness slips to third reminder. Now looking at historic
behavior, the trend is to slowly miss the medicine. The reason
could be that the patient is not interested in taking the medicine,
or even not hearing or seeing the reminder. This could mean that
the patient's health is slowly deteriorating. On the other hand, it
could simply be that the patient is busy and knows the third
reminder will come. In this case an alert is sent to the caregiver
to respond appropriately to the potential non-compliance
trending.
Another illustrative example is based on engagement index. In this
example, the accelerometer readings are sampled every 15 minutes
and the samples will show if there is movement of the patient's
hand. During their typical active period, one should see some
constant movement. If the patient's resting period gradually lasts
longer than usual, or there is no activity for more than an hour
during active period, then the patient may be experiencing a health
problem. An alert is sent to the caregiver, and the caregiver may
respond to the alert, for example by calling or visiting the
patient to determine possible remedial actions.
These historical data point trends, when analyzed over the course
of the patient's near-term and long-term treatment, help categorize
high risk patients (non-engaged or trending towards) versus other
patients (reasonably well engaged), prompting a targeted and timely
intervention by the medical staff to the high risk patients,
thereby contributing to higher adherence rates.
As an example of an alert to family caregiver, suppose the patient
wakes up every day between 7:00 AM and 7:15 AM. Then one day,
behavioral data from the PWD does not indicate movement even as of
8:00 AM. This data presents an anomaly in behavior and results in
an alert to the caregiver to call and check-in with the patient to
if a health issue need to be resolved, or if some other acceptable
reason for the anomaly exists.
FIG. 9 illustrates an example prediction data pipeline of the CCS
predictive analytics platform according to the described
embodiments. The CCS predictive analytics platform collects data in
real-time (reported by the PWD periodically, for example every 15
minutes) from all the PWDs deployed. All these data streams are fed
into the EWPA prediction system. The backend consists of a Big Data
pipeline processor, predictive analytics engine and a user
dashboard with reporting engine to provide results in real time to
caregivers and various third party healthcare providers. The
backend applications are hosted remotely in the webserver (i.e., in
the Cloud).
The Coordinated Care Platform enhances the traditional phase 3
clinical trial process by providing a wearable dedicated for
real-time monitoring of the patient and their vitals during the
trial. The PWD may also serve as a safety alert device with fall
detection and emergency calling features. The PWD also passively
collects large quantities (but still high quality) of patient's
positional data, vital sign data, and physiological data, in
real-time. This technology enables increased volume and speed of
data collection, as compared to manual methods in traditional
clinical trials, leading to further reduced data acquisition costs.
This also helps monitor patients' physical and emotional health in
real-time outside of the clinic, and could help reduce expensive
site visits by the patient.
The backend Big Data data pipelines can be customized to ingest
various types of data streams. The (machine learning) AI-based
behavioral analytics algorithms with Early Warning System alerts
(on potential non-severe Adverse Events) feed on these (raw and
derived) data points, making real-time pharmacovigilance possible
and improving associated risk mitigation procedures. The metrics,
its associated variability levels and acceptable limits can be
customized (or personalized) to every trial. The algorithms and
insights from Big Data system can effectively help design better
follow-up trials. The PWD and the connected ecosystem make outcomes
better, faster, and more meaningful, and reduces data collection
costs. As compared to using a diary/notebook, the accurate
real-time (and on-time) reporting of status and feedback helps
increase the efficacy of clinical trial outcomes and success.
While this invention has been particularly shown and described with
references to example embodiments thereof, it will be understood by
those skilled in the art that various changes in form and details
may be made therein without departing from the scope of the
invention encompassed by the appended claims.
* * * * *
References