U.S. patent application number 10/356720 was filed with the patent office on 2003-11-13 for remote health management system.
Invention is credited to Brown, Stephen J., Gunabushanam, Gowthaman.
Application Number | 20030212579 10/356720 |
Document ID | / |
Family ID | 29406615 |
Filed Date | 2003-11-13 |
United States Patent
Application |
20030212579 |
Kind Code |
A1 |
Brown, Stephen J. ; et
al. |
November 13, 2003 |
Remote health management system
Abstract
A system and method that remotely accesses and diagnoses the
medical condition of an individual patient and of each patient in a
group of patients and provides treatment based upon the diagnoses
of the individual patient and the risk stratification the
individual patient assumes in the group of patients.
Inventors: |
Brown, Stephen J.;
(Woodside, CA) ; Gunabushanam, Gowthaman;
(Hyderabad, IN) |
Correspondence
Address: |
BLACK LOWE & GRAHAM PLLC
816 Second Avenue
Seattle
WA
98104
US
|
Family ID: |
29406615 |
Appl. No.: |
10/356720 |
Filed: |
January 30, 2003 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
60379330 |
May 8, 2002 |
|
|
|
Current U.S.
Class: |
705/2 ;
600/300 |
Current CPC
Class: |
G16H 10/60 20180101;
G16H 10/20 20180101; A61B 5/7275 20130101; G06Q 10/10 20130101;
G16H 40/20 20180101; A61B 5/087 20130101; G16H 50/30 20180101; G01N
27/3271 20130101; G16H 40/67 20180101; A61B 5/0022 20130101; G16H
50/70 20180101; A61B 5/14532 20130101; A61B 5/411 20130101; A61B
5/0871 20130101; A61B 5/0205 20130101 |
Class at
Publication: |
705/2 ;
600/300 |
International
Class: |
A61B 005/00; G06F
017/60 |
Claims
The embodiments of the invention in which an exclusive property or
privilege is claimed are defined as follows:
1. A remote health management system, the system comprising: an
accessing system having a profile computer, the profile computer
further having a monitoring application, a task scheduler, a
content assignor, a dialog generator, a script generator, and a
report generator; a database in communication with the accessing
system, the database having a protocol section, a profile section,
a task list section, and a dialogs library section; and a
communication network connected to the accessing system, the
communication network further connected to a plurality of remote
terminals, each remote terminal connected to a monitoring device
operated by an individual patient, an external data source, a
healthcare provider, and a managed care organization, such that the
medical condition of each individual patient is diagnosed, a
risk-stratification is established of each individual patient in a
group of patients, and a treatment provided to each individual
patient based upon the diagnosis of the individual patient and the
risk stratification that the individual patient assumes in the
group of patients.
Description
PRIORITY CLAIM
[0001] This application claims priority to U.S. provisional
application serial No. 60/379,330, filed May 8, 2002.
FIELD OF THE INVENTION
[0002] The present invention relates to a modular interactive
system and method for remote health management, and in particular
to an automated content delivery program able to connect remote
users across independent platforms to a central database of
libraries whereby a patient's health can be scored dynamically, and
appropriate health management methods may be instituted that is
appropriate for the patient's health profile, disease state,
comprehension capacity, personal attitudes and medical needs,
including that due to co-morbid conditions.
BACKGROUND OF THE INVENTION
[0003] The invention relates to the field of health management,
particularly to an automated interactive system and method for
remotely interacting, across independent platforms with a group of
patients with one or more disease states, and with additionally
possible co-morbid conditions on a regular basis with a purpose to
educate and inform the individual about his/her health condition,
to motivate the individual to change health related behavior,
secure compliance with medical regimens, to monitor health related
parameters, and to intervene early with an ultimate view to improve
the health status of the individual monitored client.
[0004] In the United States alone, over 100 million people have
chronic health conditions, accounting for an estimated $700 billion
in annual medical costs. Because of the continuous nature of these
health conditions, and in an effort to control these medical costs,
many healthcare providers have initiated outpatient or home
healthcare programs for their patients. The success of these
programs is dependent upon the healthcare provider's ability to
effectively monitor patients remotely, and to detect and intervene
at an early stage in order to prevent the patient's medical state
from becoming more complicated, expensive and difficult to manage.
In addition, the program's success is also dependent on its
sustaining the patient's interest and continued participation in a
process that often extends to the remaining term of an individual's
life.
[0005] Managing a chronic disease or ongoing health condition often
requires the monitoring and controlling of a physical or mental
parameter relating to the health condition. Examples of these
parameters include blood glucose in diabetes, respiratory flow in
asthma, blood pressure in hypertension, cholesterol in
cardiovascular disease, weight in eating disorders, T-cell or viral
count in HIV, and frequency, severity or timing of episodes in
mental health disorders.
[0006] Since the patients themselves monitor their health
condition, the clinician is often limited to learning each
patient's status strictly through patient initiated events, such as
an emergency visit, an urgent care visit, a phone call, or other
patient initiated event that results in delivery of the patient's
latest medical data. Even with the current availability of remote
monitoring devices that store and transmit medical data from a
patient's home to a clinic, the clinician must still wait for
medical information whose arrival depends on the patient's
initiative.
[0007] As a result, the majority of the clinician's time is spent
with the patients who are the most motivated and eager for a
response, or patients whose conditions have become acute and
require immediate attention, while the greatest opportunity to
improve care and prevent conditions from exacerbating remain
unknown and hidden with the less motivated or "pre-acute" patients
who do not visit the clinician or transmit their medical data.
[0008] The less motivated patients often develop urgent medical
needs that could have been prevented with prior medical management.
Consequently, the cost of treating their chronic health conditions
is much higher than one might expect given the sophistication of
current medical monitoring devices.
[0009] In addition, the management of well motivated patients
differs considerably from the strategy employed in managing high
risk patients who aren't driven to initiate care because they do
not perceive a crisis or are less motivated to change their
behavior. Thus it is important to determine the level of motivation
in the individual patient when deciding the plan of management.
[0010] A patient health status reporting system that summarizes and
stratifies by risk-potential, the data received from the patients
as a group would help the healthcare provider identify those
patients who are in the greatest need of the provider's attention,
and would help increase the provider's efficiency and
productivity.
[0011] Unfortunately, most existing healthcare information systems
are only designed to display medical data on an individual patient
basis. Few systems have been developed that enable clinicians to
view medical data for an entire group of patients simultaneously.
Consequently, it is extremely difficult for a healthcare provider,
such as a clinician or a nurse to prioritize his or her time and
efforts in a manner that optimizes care and minimizes costs and
complications for a given group of patients.
[0012] The success of a health management program in chronic health
condition also depends on the program's ability to modify the
health related behaviors of the patient. Examples include changing
the dietary habits, and exercising habits in a patient with
diabetes; smoking cessation in patients, who have suffered heart
attacks, etc. A patient's compliance to medical advice varies
considerably with the patient's perception of his/her health
condition, healthcare provider; level of knowledge regarding
his/her health condition, personal beliefs, motivational drivers,
etc. In order that that the patient receives the best medical
advice, and with a view to improve the ultimate prognosis of an
individual with a given condition, it becomes necessary that the
healthcare management plan takes into account the above factors,
and that it is customized to the individual.
[0013] Notwithstanding the methods to improve the compliance in the
patient, with patients on prolonged follow up, there often develops
resistance to the health management plan. This resistance may
develop as a result of symptom-relief in the patient and his/her
consequent inability to appreciate that the underlying disease
process is unchanged or may be worsening (the patient feels that
he/she `doesn't need the medication anymore`). It may also develop
in response to the nature of content presented to the individual
i.e. the patient finds queries regarding a particular context
intrusive to his/her lifestyle. It is important to detect
resistance early and suitably modify content so as to prevent the
further development of resistance, and improve compliance in the
patient.
[0014] In some diseases such as asthma and allergy, and in the
mental health conditions, the precise diagnosis is not always known
to the healthcare provider. Further, in these conditions, even
after the diagnosis is made, the best treatment is not always clear
and may need to be evaluated over time. In these patients, dynamic
monitoring of the patient may help understand the condition better,
and formulate the ideal medical management plan in the given
patient.
[0015] It is also important to determine the reliability,
consistency and accuracy levels of the information that is inputted
into the system, given that the future medical management of the
patient is dependent on this data. This is especially more so in
those cases where the data provided by the patient is the only
source of information, and in the field of medical research.
[0016] Additionally, in the field of medical research, it is
necessary to analyze the patient data in order to better understand
patient diagnosis and needs. The system presents a method by which
routinely collected data from patients over multiple healthcare
facilities may be integrated and this information may be used to
understand subgroups of patients who may respond differently to
treatment or benefit from different treatment options. In addition,
the invention also presents a method by which patients may be
selected for enrollment in studies.
[0017] Further, it is also advantageous that any remote health
management system be compatible with a range of communication
protocols and devices, in order that the patient communicates using
the media and remote apparatuses that he/she is most comfortable in
using, and has ready access to. Differing remote apparatuses and
communication networks have varying requirements and limitations
and advantages with regard to data display and transfer. There are
advantages with specific media that may be utilized in ensuring a
more satisfying interaction of the patient with the healthcare
provider, a greater involvement in the disease management process,
and ultimately a better prognosis in the patient's disease state.
Current systems are incapable of automatically optimizing content
to the remote apparatus, type and speed of communication network,
and to individual preferences.
[0018] Current systems are incapable of automatically administering
a management plan that is relevant to patient's profile, updating
the profile in response to replies received from the interaction,
and highlighting to the provider those aspects of the patient's
condition that require his/her greatest attention. Current systems
are also incapable of risk-stratifying the individual patients
within the group.
[0019] Current systems lack the capability to analyze the
reliability, consistency and veracity of the replies, and validate
the information inputted into the system. Further, there is no
system in place that enables a researcher or a healthcare provider
to select research subjects either prospectively or retrospectively
for study on the basis of data contained within the profile.
Current systems do not easily allow the collected data to be
integrated over multiple healthcare facilities and utilized for the
purpose of medical research, impeding the conduct of large
multi-centric studies. Finally, current systems do not help the
healthcare provider in making a diagnosis and in determining the
medication and health management that is most suited to the
individual patient.
[0020] This and other advantages of the invention will become
apparent on consideration of the ensuing description below.
SUMMARY OF THE INVENTION
[0021] The invention is a system and method that automatically
initiates a remotely communicated surveillance and health
management process to interactively obtain a patient's
health-related profile, update and analyze the profiles, categorize
the patient's profile by risk-stratification methods using multiple
profiles from other patients, and administer a health-related
management plan that is relevant to the patient's updated profile
and risk-stratification. The invention further provides researchers
and healthcare providers to prospectively and retrospectively
select patients for study based on data contained within each
profile, and each updated profile of each patient.
[0022] The invention further provides a system and method to
highlight to a patient's provider the patient's condition that
require his/her greatest attention using the updated and
risk-stratification data obtained by remote iterative and
interactive querying of the patient based on updated data from the
patient in comparison to updated data from each patient in a study.
The invention also assists health care providers in diagnosing and
in determining the medication and health management that is most
suited to the treatment of each patient, individually, or, in light
of risk-stratification analysis, in determining the treatment based
upon the risk that a given patent assumes in a group of
patients.
[0023] In accordance with still further aspects of the invention,
the collected data is integrated over multiple healthcare
facilities and utilized for the purpose of medical research to
enhance the conduct of large multi-center studies.
BRIEF DESCRIPTION OF THE DRAWINGS
[0024] The preferred and alternative embodiments of the present
invention are described in detail below with reference to the
following drawings.
[0025] FIG. 1A is a schematic block diagram illustrating the
components of the system according to the preferred embodiment of
the invention.
[0026] FIG. 1B is a schematic block diagram illustrating the
architecture of the system and its connections to multiple
individuals at different types of remote terminals according to the
invention.
[0027] FIG. 2 is a block diagram depicting interdependent
characteristics (operators) of a dialog;
[0028] FIG. 3 is flow chart depicting the steps in creating and
storing of content data from a dialog;
[0029] FIG. 4 is a flow chart diagram depicting the creation of the
programming statements using a Dialog Editor Platform;
[0030] FIG. 5 is a schematic block diagram illustrating the
generation of an interview form according to the method of the
invention.
[0031] FIG. 6 is a schematic view of an interview form appearing on
the screen of the remote terminal of FIG. 1A.
[0032] FIG. 7 is a schematic block diagram illustrating the
functions of individual component programs at the profile computer
of the remote health management system.
[0033] FIG. 8 is a block diagram illustrating the three dimensional
aspects of the dynamically determined risk state output scale;
[0034] FIG. 9 is a schematic view of the report generator
interface, as it appears to the healthcare provider.
[0035] FIG. 10A is a flow chart illustrating steps included in the
method of the invention.
[0036] FIG. 10B is a continuation of the flow chart of FIG.
10A.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
[0037] Referring to FIG. 1A, the remote health management system 10
includes a profile computer 20 having a monitoring application 22,
task scheduler 24, content assignor 28, dialog generator 30, script
generator 32 and report generator 34. Profile computer 20 is
connected to a database system 40. Database system 40 includes
protocols database 42, profile database 44, tasklist database 46
and dialog library 48. Profile computer 20 and database system 40
are networked to a modem M1 for connecting profile computer 20 and
database system 40 to a communication network 50.
[0038] An individual 70 desiring access to system 10 is located at
remote terminal 72. Remote terminal 72 is connected to
communication network 50 through a modem M2 such that remote
terminal 72 accesses system 10 for interactive health management
sessions through modem M2 and communication network 50. It is
obvious that many more remote terminals can be connected to
communication network 50 for accessing system 10.
[0039] A monitoring device 74 for monitoring a health condition of
individual 70 is connected to remote terminal 72. Monitoring device
74 is capable of producing measurements of a physical
characteristic of the health condition and of uploading the
measurements to remote terminal 72 for transmission to system 10.
In one possible embodiment, individual 70 is a diabetic and
monitoring device 74 is a blood glucose meter for measuring blood
glucose levels of individual 70. In another embodiment, individual
70 is asthmatic and monitoring device 74 is a peak flow meter for
measuring the individual's peak flow levels. Specific techniques
for connecting a monitoring device to a remote terminal for remote
monitoring of an individual's health condition are well known in
the art.
[0040] Communication network 50 further connects a healthcare
provider 60 of individual 70 to profile computer 20. Provider 60
has a medical record database 62 for storing electronic medical
records of individual 70. Medical record database 62 is connected
to communication network 50 through a modem M3 such that profile
computer 20 receives through network 50 the stored electronic
medical records from database 62. Similarly, communication network
50 connects a managed care organization 56 of individual 70 to
profile computer 20. Organization 56 has a medical claims database
58 for storing medical claims data of individual 70. Medical claims
database 58 is connected to communication network 50 through a
modem M4 such that profile computer 20 receives through network 26
the stored medical claims data from database 58.
[0041] Communication network 50 also connects profile computer 20
to an external data source 52 having additional material database
54. Database 54 contains additional dialog libraries, additional
protocols and educational materials which may be used to induce a
modification in the health related behavior of individual 70.
Database 54 is connected to communication network 50 through a
modem M5 such that database system 40 may transfer additional
educational materials from database 54 to remote terminal 72
through network 50.
[0042] It is obvious that a plurality of remote terminals 72 and
monitoring devices 74 may be used by the same or different
individuals 70 to communicate with remote health management system
10 at any point of time. Further, a given individual 70 may prefer
one terminal for data entry, and another for viewing multimedia
components, depending on his/her individual taste. For example, an
executive with Diabetes may prefer to enter data by the use of a
Personal Digital Assistant, but may prefer to receive videos of
aerobics on his digital TV at home to help him with his weight
reduction program. When this executive is on a business trip, he
might wish to answer the queries but skip the digital TV bit. In
addition individual 70 may have preferences about the form and
manner of presentation of content, including such entities as
color, font, and background visualization preferences within remote
terminal 72.
[0043] Further, with the passage of time, newer communication
protocols, remote terminals and monitoring devices may become
available, and individual 70 may wish to use these newer devices.
It would be inefficient for the healthcare provider 60 to create a
separate dialog and script for each of the supported device types.
Even if this were done, there is the added possibility of the older
data being incompatible with the newer systems. To circumvent the
above problem, when dialogs 100 that are created by healthcare
provider 60 include multimedia and other components as datafiles
136. Script generator 32 reformats these so called `master-dialogs`
into scripts that are interpreted by remote terminal 72 and
monitoring device 74.
[0044] In the preferred embodiment, communication network 50 is a
public communication network such as the Internet, and system 10,
remote terminal 72, healthcare provider 60, managed care
organization 56, and external data source 52 connect to the public
communication network through the use of modems, as illustrated in
FIG. 1A. Alternatively, communication network may be a wireless
communication network, cellular network, telephone network, or any
other network which allows the fore-mentioned devices to exchange
data with profile computer 20. For clarity of illustration, only
one remote terminal 72 is shown in FIG. 1A. However it is to be
understood that remote health management system 10 may include any
number of remote terminals, and a single individual 70 may
communicate with profile computer 20 through the means of more than
one remote terminal 72. Specific techniques for networking computer
systems and the electronic devices mentioned above for on-line
interaction are well known in the art.
[0045] FIG. 1B is a schematic block diagram illustrating the
architecture of the system 10 and its connections to multiple
individuals at different types of remote terminals according to the
invention. Referring to FIG. 1B, multiple individuals 70 use
different remote apparatuses 72 (72A, 72B, 72C, 72D, 72E and 72F)
in order to receive dialogs and transfer replies and physiological
measurements (parameters) to healthcare provider 60 and research
worker 60A, both having access to system 10 via remote terminal 64
in signal communication with network 50G. Network 50G is in signal
communication with the monitoring application 22 via modem M1G.
Apparatus 72A is preferably a personal computer connected to remote
health management system 10 through a high speed internet
connection 50A. Here, monitoring application 20 connects to modem
M1A through a high speed internet data server (not shown). The
advantage of using communication network 50A in delivering content
to individual 70 in this instance is that the content presented can
be data intensive, including rich graphics and multimedia (video
and sounds). Further, the data processing, and storage capabilities
of the personal computer may be utilized so that the individual has
greater control over the information inputted into the system, as
will be described in the alternative embodiment.
[0046] Remote apparatus 72B is a handheld mobile communication
device such as a wireless enabled personal digital assistant (PDA)
or a wireless access protocol (WAP) enabled mobile phone. System 10
communicates with apparatus 72B by a wireless data server and Modem
M1B connected to wireless network 50B. An advantage of using
apparatus 72B is that individual 70 has greater connectivity when
compared with apparatus 72A.
[0047] Remote apparatus 72C is a digital television that uses
digital television network 50C for data communication. Modem M1C is
used in transferring data in the digital television format from
monitoring application 22 to the network 50C.
[0048] Remote apparatus 72D is a standard dial tone multi-frequency
(DTMF) telephone which communicates with system 10 via standard
telephone lines. Voice data is routed from monitoring application
22 using a voice data server in system 10 through modem M1D from
and to individual 70 via remote apparatus 72D in signal
communication with network 50D. Similarly, data received from the
individual is used to make changes in the profile. Specific
techniques to do this are well known in the art.
[0049] Remote apparatus 72E is designed to execute script programs
received from the system, and transfer replies and measurements to
the system via communication network 50E, that may be an internet
connection, standard telephone line, digital television network or
a wireless connection in signal communication with modem 1E of
System 10. In the preferred embodiment, apparatuses 72A, 72B, 72C,
72D and 72E are directly in connection with system 10. In addition,
these apparatuses are connected to atleast one remote monitoring
device for measuring physiological variables of individual 70.
Remote apparatus 72F is a communication enabled monitoring device
whereby, connectivity is enabled within the device itself. The
enabled monitoring device 72F uses network 50F that may use the
internet, standard DTMF telephone line, wireless network or a
digital television network to communicate with system 10. An
advantage of using the enabled device 72F is reduced cost of the
entire unit for the individual.
[0050] Healthcare provider 60 utilizes a remote terminal 50G in
order to access patient profile databases, requisition reports, and
to manually assign script programs to the individuals, either
singly, or to an entire group of patients. Further the system may
be also be used to schedule consultations and fix appointments with
the individual monitored patient. Remote terminal 50G is preferably
a workstation connected to the system using a secure internet
connection. Further, any person desirous of communicating with
individuals 70 may be additionally granted access to the system
using the remote terminal (workstation) 50G. This may include an
administrator who wishes to enroll patients, collect, analyze and
report data. Further, a researcher who wishes to enroll patients
for the purpose of research, or who wishes to access the data
within the profile database may additionally be granted access to
the system.
[0051] In the preferred embodiment, the components of system 10,
such as monitoring application 22, task scheduler 24, dialog
generator 30, script generator 32, profile database 44 and tasklist
database 46 are all physically located within the profile computer
20 and database system 40. In an alternative embodiment, the remote
apparatus additionally includes one or more components that are
contained within the profile computer in the preferred embodiment.
In one instance, the individual's profile database 44 may be
stored, atleast in part within the remote apparatus itself. The
profile database 44 on the remote apparatus is synchronized with
that on the database system 40. An individual would not need to
establish a connection with the profile computer every time he/she
wishes to enter data into the system. This is particularly
advantageous where the individual has to reply to the same or
similar queries over a period of time. For example, a patient
undergoing treatment for mood disorders may be asked to maintain an
electronic `diary` where he/she answers queries regarding his/her
mood. In the alternative embodiment, data entered each day would be
used to build the patient's profile database. At the end of a
period of time, the apparatus would transfer the newly added
profile variables and this information would be used to update the
individual's profile. Alternatively, the electronic `diary` may
also be presented to the healthcare provider at a personal
visit.
[0052] The profile database also includes information regarding the
individual's preferred media, personal preferences, and
visualization options. Henceforth, in instances where mass media
modalities such as digital television are used for communication,
the program logic may be broadcasted to all the individuals, and a
personalized experience may be created through local processing and
locally stored profile variables
[0053] The monitoring application and components of the protocols
database may additionally be incorporated within the remote
apparatus. The monitoring application within the remote apparatus
scans the profile database for `suspicious` correlates of data, as
defined in the protocols database. If at any time, a combination of
variables in the profile database is suggestive of an acute crisis
or a worsening disease state, then the monitoring application is
activated, and this establishes a connection with the profile
computer. At regular intervals of time, the profile database,
protocols database are updated for newer information and
protocols.
[0054] The alternative embodiment has the advantage of lowering the
communication costs with the system. A further advantage with the
alternative embodiment is that the individual exercises greater
control over the data that he/she wishes to provide to the system,
which is preferable to many individuals.
[0055] FIG. 2 is a block diagram depicting the interdependent
characteristics (operators) of a dialog 100 in the system matrix.
The interdependent characteristics include the dialog's general
attributes 110, library attributes 120 for the aspect of care
addressed and dialog content 130, the active content of dialog 100.
Dialog attributes 110 define such characteristics of the dialogs
such as the name of the dialog, the unique code that is used by the
system to refer to the dialog, version of the dialog, the name of
the dialog creator, the date of creation and last update of the
dialog. Dialog attributes 110 further includes information
regarding the previous versions of the dialog, and references to
those dialogs that are derivatives of the current dialog, besides
containing a reference to the previous versions of the dialog.
[0056] Library attributes 120 of a dialog include those attributes
that define the place of the dialog in the library. Besides
allowing for the searching, updating, and classification and
retrieval of dialogs by healthcare provider 60, the information
contained within library attributes 120 is also referenced by
content assignor 28 in order to assign a particular dialog to the
individual 70 from amongst the plurality of dialogs within the
library of dialogs. Library attributes 120 includes a health
condition referential 122, disease attribute referential 124,
client profile referential 126, and a priority referential 128.
Health condition referential 122 includes information on the
disease state or health related condition and attribute in the
management of which the dialog would typically need to be employed.
The states that are defined by health condition referential 122 are
not necessarily exclusive and can often overlap, and be subsets of
other health condition referentials particularly so when the
disease or health related condition is present in a diverse
population. For instance there may be one attribute that relate to
diabetes, and another that relates specifically to juvenile
diabetes (a form of diabetes). Obviously, dialogs containing health
condition referential 122 that relate to diabetes can be freely
applied to all patients who have juvenile diabetes but not vice
versa. An advantage of the above is that it decreases the number of
dialogs that have to be created for managing conditions that
present in a diverse population.
[0057] Disease attribute referential 124 declares a specific
attribute of the disease or health related condition that relates
to the said dialog 100. Disease attribute referential 124 in a
disease such as diabetes may include references to organs that are
damaged as a result of the disease in the longer term as a result
of poor glucose control, such as that for the kidneys, the blood
vessels, the eyes and the feet, to name a few. It can further
include follow up material that relates to patient's compliance
with medication, patient's compliance with dietary modification, a
weight reduction program, and a module that ensures that the
patient is monitoring his/her blood sugar regularly. Thus, diseases
and health related states are divided into a plurality of
attributes, and these attributes are referenced in the disease
attribute referential 124.
[0058] Patients (individual 70) may vary widely in their
educational background, comprehension capacity, motivational
drivers, attitude and perception of their condition, and preference
of communication medium. In order to communicate well with the
patient, it becomes necessary to optimize the communication to one
that is best suited to a particular patient. Client profile
referential 126 defines the characteristics of individual 70 in
whom the particular dialog 100 would likely be effective and
pertinent. Thus, at the time of assignment of a particular dialog
100 to individual 70, either by the automatic content assignor 70
or by healthcare provider 60, it is ensured that dialog 100 is
individualized to the patient not only in terms of the disease
condition or requirement in the patient's health condition, but
also in terms of the patient's individual preferences.
[0059] Priority referential 128 describes the relative importance
of the dialog in terms of urgency, value of the particular
information in the management of the health related condition and
health attribute status of individual 70. Priority referential 128
includes information that enables the system to prioritize from
among a set of possible dialogs that would need to be served on
individual 70. Further, priority referential 128 defines the
sequence in which dialog 100 appears in a particular
communication.
[0060] Dialog content 130 further includes queries 132, replies
134, datafiles 136 and actions 140. Query 130 may be a question
statement to a patient regarding a particular aspect of the disease
or health related condition. Text markers at the beginning of a
particular query are used to identify a given query 132 within
dialog 100. Each query 132 is followed by a list of possible
replies 134, and the patient is asked to select the best reply 134.
Of course, it is possible that, in the case of some queries 130,
more than one reply is possible. In such a case, the patient is
asked to choose all the replies 134 that apply. Actions 140 define
the next step to be followed for each of the possible replies. Of
course, when more than one reply 134 is allowed for a particular
query 132, actions 140 would define the procedure to be followed
for each of the possible allowed combinations. Alternatively,
default actions 140 may be assigned within a dialog 100 for that
reply 134 to which no action 140 has been defined. For instance,
one such default action would be to go to the next query. Further,
more than one action 140 may be assigned to a particular reply
134.
[0061] Possible actions 140 includes next function 142, jump
function 144, newdialog function 146, database function 148,
playfile function 150, alert function 152, measure function 154 and
terminate function 156. When Next function 142 is applied to a
particular reply 134, the next query is put to the user. Jump
function 144 instructs the program to go to specified query 134
within dialog 100. Newdialog function 146 instructs the program to
terminate the present dialog and download another dialog 100 from
communication network 50 or that stored at remote terminal 72.
Database function 148 instructs the program to make amendments and
additions in profile database 44 and tasklist database 46. Playfile
function 150 instructs the program to download and run a specified
datafile 136 from communication network 50 or that stored at remote
terminal 72. It is of particular value where multimedia components
such as digital audio and video streams containing health related
messages may be played to individual 70.
[0062] Alert function 152 instructs the program to send an alert
message to healthcare provider 60 informing him/her of the receipt
of specified reply 134 to query 132. Alert function is intended to
be used in case a given combination of replies received from
individual 70 is likely to lead to a poor health status in the near
future, and/or when immediate intervention is required and/or the
healthcare provider 60 would need to be informed about the reply
134 urgently. Measure function 154 contains instructions to remote
terminal 72 to collect measurements from monitoring device 74.
Health related parameters may be collected from individual 70 and
sent to remote health management system 10 by this function.
[0063] Terminate function 156 instructs the program to terminate
the execution of specified content 130. Terminate function 156
additionally includes an arithmetic or logical method for
computation of the results of a specified group of queries 132 and
replies 134. For example, dialog content 130 can include a number
of queries, with each of the replies being assigned a score 158.
Terminate function 156 may be used to instruct the program to
compute arithmetically the final score 159 for a given set of
queries 134. Alternatively, logical methods may be used to derive
score 159 for each of one or more combinations of replies 134.
Alternatively, any of actions 140 may be assigned to final score
159. An advantage of terminate function is that long standardized
questionnaires for risk determination in a variety of conditions
may be applied to individual 70 in their original format and scored
automatically, and the final score 159 be further used to execute
action 140. It will be obvious to anyone versed in the art that a
number of methods and programming languages are available to write
programs that can interpret dialog content 130 and perform any of
the above functions.
[0064] FIG. 3 is a flow chart depicting the steps in creating and
storing of dialogs 100. A healthcare provider's first task is to
assign a name, and provide the general attributes of the
dialog-to-be-created as depicted in block 350. Next, the user
defines the health condition to which the dialog will primarily
refer (health condition referential 122) at block 352. The provider
then identifies an aspect of care at block 354 to which the dialog
will primarily refer (disease attribute referential 124). Following
this, the provider identifies the profile of the individuals to
whom the dialog is customized (client profile referential 126) in
block 356, and the priority given to the dialog and the health
context that it attempt to modify (priority referential 128) in
block 358. Once the naming conventions and the attributes of the
dialog 100 are assigned, the provider creates dialog programming
statements at block 359, in a graphical programming environment as
embodied in FIG. 4. New dialog content is then stored in the dialog
library 48 at block 360.
[0065] The provider who has access to create new content does so
using a simple dialog composer as embodied in FIG. 4. FIG. 4 is a
diagram depicting the creation components of a dialog Editor
Platform 362. First, a provider is presented with a program icon
palette 365 of programming statements that are represented as
graphic symbols (icons) that can be dragged from the palette of
available statements into a dialog construction platform 370. In a
typical embodiment of the present invention, the provider drags a
start query icon 372 and a multi-pronged reply icon 375 from an
icon palette down to the construction platform 370. The provider
then activates a dialog box for each icon by clicking on it with a
mouse and specifying a query associated with that particular icon,
for example, a Start Query Dialog 374. Next, in a Reply Dialog 380,
the provider assigns one or more actions 140 of reply action group
382 to each reply 378. The reply actions group 382 includes one or
more actions 140 i.e. the execution of an arithmetic function on
risk value R, and/or appending/modification of data within database
system 40, execution of follow up queries, playing multimedia files
on remote terminal 72 and collecting measurements from monitoring
device 74. In those actions 140 which have pointers to follow up
queries, next query icon 385 is dragged onto the construction
platform along with an associated reply icon 375, in the same
manner as described for reply icon 375.
[0066] The same steps 388 are repeated for each of the follow up
queries, till all the queries and follow up queries within the
dialog 100 have associated actions 140 assigned to them. By
clicking on customization icon 390, the provider activates
customization dialog box 392. Customization dialog box 392 allows
the user to customize the dialog to different kinds of remote
terminals 72 and monitoring devices 74, and make additions to the
output over and above the baseline dialog. For instance, the
provider can include specific multimedia components to be served on
individual 70 in addition to/lieu of any of the actions 140
wherever the remote terminal 72, monitoring device 74 and
communication network 50 support it. The advantage of customization
dialog box is that it allows the provider to make additions to
existing dialogs and add multimedia and other components without
changing the basic structure of the queries that are applied to
individual 70. In this manner, a `master-dialog` can be initially
created that includes data components of all supported remote
terminals 72. As and when newer and enhanced versions of remote
terminals 72 become available, the corresponding components may be
selectively appended and modified within dialog 100.
[0067] By clicking on the output icon 394, the provider activates
the output dialog box 396. At any time during or after the dialog
creation process, the provider can review the dialog created, using
a simulation interface to an appropriate appliance, or
alternatively the provider can review the actual dialog content in
a text only overview window.
[0068] Once all the follow up queries, replies and output dialogs
are formulated and put onto the construction platform 370, the
newly created dialogs are stored in a dialogs library 48 from where
it may be incorporated by provider 60 and content assignor 28 in
care management programs and for future updating and editing.
[0069] Referring to FIG. 5, the profile computer 20 includes the
dialog generator 30 and is capable of generating an interactive
dialog 100 and of transferring it through communication network 50
to the display unit of remote terminal 72. At the time of initial
registration of individual 70 in the system, dialog generator 30
generates a registration dialog 100R and transfers it through
communication network 50 to remote terminal 72. Registration dialog
100R includes data fields for a name 80, a language 81, current
medical condition 83, and co-morbid health condition 84 of
individual 70. Profile database 44 has storage capability for
storing profile record 45 that includes, name 80A, language 81A,
current health condition 83A and co-morbid health condition 84A of
individual 70.
[0070] In the preferred embodiment, dialog library 48 stores
hundreds of dialogs 100 in various languages relating to possible
health conditions of individual 70, such as asthma, diabetes,
nicotine addiction, etc. To narrow the focus of registration dialog
10OR, dialog generator 30 communicates with content assignor 28,
which tailors an interview dialog 1001 in dependence upon language
81 and current and co-morbid health conditions of individual 70.
For example, in FIG. 5, individual 70 has indicated his language 81
as "ENGLISH" current medical condition as "DIABETES" and co-morbid
health condition 84 as "SMOKER" so content assignor 28 assigns
interview dialogs 1001 that contain English language queries
pertaining to diabetes and smoking related behaviors of individual
70. As illustrated in FIG. 5, language 81 further includes French
and Spanish. Other languages may be included. Current medical
condition 83 also includes asthma and depression. Other medical
conditions may be included. Co-morbid condition 84 also includes
alcoholism, obesity, hypertension, and hyperlipidemia. In an
alternative embodiment, content assignor 28 uses data from medical
records database 62, medical claims database 58, and/or additional
material 54 to assign registration dialog 100R individual 70.
[0071] Registration dialog 100R is illustrated in greater detail in
FIG. 6. Registration dialog 100R contains a first category of
queries 160 relating to the current health condition of individual
70. Category 162 is divided into a first subset of queries 162 for
determining current diseases or symptoms of individual 70 and a
second subset of queries 164 for determining the pattern and
history data of the individual's health condition. For example, in
the example in which individual 70 is a smoker, subset 162 asks
about any current diseases or symptoms smoking has caused in
individual 70. Similarly, subset 162 asks for the pattern and
history data of the individual's smoking habit.
[0072] Registration dialog 100R further includes a second category
of queries 166 for determining the motivational drivers of
individual 70. Category 166 is divided into six subsets of queries
including longevity 168, quality of life 170, family life 172,
social responsibility 174, social acceptability 175 and economy 176
for determining a value placed by individual 70 on various
motivating factors for modifying his or her behavior. Subset 168 is
for determining the value placed by individual 70 on longevity. For
example, where individual 70 is a smoker, subset 168 includes
queries to determine if the prospect of living a long life would
provide sufficient motivation to quit smoking.
[0073] Subset 170 is for determining the value placed by individual
70 on the quality of his or her life. In the smoking example,
subset 170 includes queries to determine if an improvement in
smoking symptoms, such as no longer suffering coughing fits, would
provide sufficient motivation to quit smoking. Subset 172 is for
determining the value placed by individual 70 on family life. In
the smoking example, subset 172 includes queries to determine if an
improvement in family life, such as no longer harming relatives
with second hand smoke, would provide sufficient motivation to stop
smoking. Subset 175 is for determining the value placed by
individual 70 on social acceptability. In the smoking example,
subset 175 includes queries to determine if an improvement in
social acceptability, such as no longer offending people with bad
breath, would provide sufficient motivation to quit smoking.
[0074] Subset 174 is for determining the value placed by individual
70 on social responsibility. In the smoking example, subset 174
includes queries to determine if an improvement in social
responsibility, such as not burdening society with the cost of
Emphysema treatment, would positively motivate individual 174 to
quit smoking. Subset 176 is for determining the value placed by
individual 70 on economy. In the smoking example, subset 176
includes queries to determine if the cost savings associated with
no longer purchasing cigarettes would provide sufficient motivation
to quit smoking.
[0075] Registration dialog 100R also includes a third category of
queries 178 for determining the comprehension capacity of
individual 70. Category 178 is divided into four subsets of queries
including age 180, language skills 182, reading habits 184, and
educational background 186 for determining various comprehension
capacity factors. Subset 180 is for determining an age of
individual 70 and subset 182 is for determining language skills of
individual 70. Similarly, subset 184 is for determining reading
habits of individual 70 and subset 186 is for determining the
educational background of individual 70.
[0076] A fourth category of queries 188 is for determining a media
preference of individual 70. Category 188 is divided into three
subsets of queries including picture 190, text 192 and video games
194 for determining if the individual 70 prefers pictures, text, or
video games, respectively. Of course, these types of media are just
examples of possible media choices and other media, including mixed
media selections, are possible in alternative embodiments.
Registration dialog 100R and educational fulfillment bank 22 may
offer other media choices in alternative embodiments, such as
computer videos, musical lyrics, or hyper-text links.
[0077] Referring to FIG. 7, monitoring application 22 is designed
to generate a profile database 44 that includes a motivational
driver profile 202, a comprehension capacity profile 204, and a
media selection profile 206 from the questionnaire replies.
Monitoring application 22 further includes a confirmation program
for sending a confirmation form 198 to remote terminal 38. The
confirmation form includes a summary of each generated profile 202,
204, and 206 so that individual 70 may confirm each of the
generated profiles. Profile database 44 is also designed to store a
profile record of individual 70 including his or her name, current
health condition, and confirmed profiles. Additionally, profile
database 44 includes medical claims data 58D received from claims
database 58, electronic medical records 62D received from medical
record database 62, and device measurements 74D received from
monitoring device 74. Thus, profile database 44 includes
representing the medical condition 83 of individual 70, co-morbid
health conditions 84, health related parameters, and progress
regarding the conditions.
[0078] Content assignor 28 is designed to parse through data
contained within profile database 44, and selects those dialogs 100
from profile database 44 that are most appropriate to the health
condition, profile, and needs of individual 70. The selected
dialog(s) 100 are exported to script generator 32 that
individualizes the dialog to individual 70, stores it in the
dialogs library 48, and converts it to a format that is recognized
by remote terminal 72 and sends it over communication network 50
via modem M1. The communication network 50 exchanges data between
the monitoring device 74 and the individual 70 through the remote
terminal 72 and the modem M2.
[0079] FIG. 8 is a block diagram illustrating the three dimensional
aspects of the dynamically determined risk state output scale which
is subsequently used by content assignor 28 to determine the most
appropriate dialogs 100 for assignation to individual 70. The
X-axis 200 defines the relative risk of individual 70 on an
arbitrary relative scale. Replies 134 to query dialogs 132 sets the
risk score 205 at a certain level on the scale, and further replies
134 may be used to modify the assigned score 205 in either
direction. Actions 140 of dialog content 130 may so be programmed
that the new risk score 205 is a value derived from arithmetic
modification of the risk score 205 that was assigned previously to
a specified parameter in the risk profile of individual 70. Over a
period of time, replies 134 can lead to the creation of an
accumulated risk profile. Additionally, replies 134 to dialogs 100
that are incorporated as a value in a mathematically calculated
risk state may incorporate other answers as well, creating a
composite, weighted risk state. The Y-axis 210 refers to the actual
aspect of care in which the risk will be incorporated. The Z-axis
220 incorporates the expression of risk, i.e., whether the risk is
assigned to a sign or symptom 232, a behavior 234, or a knowledge
expression 236. An individual 70 on follow up for long periods may
additionally be assigned follow-up risk 230 that compares the
health related parameters, compliance to medication etc over a long
period of time, and charts the course of the health related
parameters of individual 70. Similarly, the individual 70 on follow
up for long periods may additionally be assigned follow-up
resistance risk 229 that compares the health related parameters,
resistance to medication etc over a long period of time, and charts
the course of the health related parameters of individual 70. This
dynamic model allows for very sophisticated risk profiling
including risk trend alerts, composite risk profiling by aspects of
care and profiling by risk expression, as will be described in the
charts below. Further, the dynamic risk `foot prints`, or the
pattern of risk scores 205 at any time can serve as triggers for
automated content selection.
[0080] While a considerable amount of health related data may be
available regarding an individual 70, in the interest of the health
provider's productivity, it is, pertinent that only relevant data
is presented. It is also preferable that the healthcare provider 60
views patients with similar health status as a group, and decides
the future plan of management accordingly. Report generator 34, and
report generator interface 300 are used to generate reports which
are used by healthcare provider 60 in following the health status
of individuals 70 or groups of individual 70.
[0081] FIG. 9 shows the report generator interface 300 as viewed by
healthcare provider 60. It includes fields Name 302, health
condition 304, co-morbid condition 306, profile parameters 308,
context type 309, follow up list 310, custom list 312, report type
314, export list button 316, store list button 318 and create
report button 320. Name 302 is a drop-down menu that includes the
names of all individual 70 that are managed by healthcare provider
60. Healthcare provider 60 has the option of selecting one or more
individuals 70 from the drop down menu, for whom he/she wishes to
view a generated report 320. Health condition 304 is also a
drop-down menu with a list of all the medical conditions 83 which
provider 60 is managing. Similarly, co-morbid condition 306
includes all the co-morbid conditions that are seen in individuals
70.
[0082] Profile parameters 308 allows healthcare provider 60 to
select individual 70 by any aspect of data available in their
profile database 44. Examples of this include can include (1)
individuals 70 who use digital TV as their remote terminal 72, (2)
individuals with diabetes with poor sugar control scores, (3)
individuals who are showing resistance to smoking cessation
messages, (4) individuals who have not entered data into the system
for more than 3 continuous days, (5) individuals with mood
disorders who have refused to answer queries regarding illicit drug
abuse, (6) individuals with risk scores below a predetermined level
in any context, etc. It will be appreciated that possible profile
parameters 308 that healthcare provider 60 can create or select are
limitless, and that the examples illustrated above are only broadly
indicative of the many parameters available to healthcare provider
60. In addition, it is also possible for healthcare provider 60 to
select a combination of two or more parameters, for the purpose of
analysis, and report generation.
[0083] Context type 309 is another drop down menu that allows
healthcare provider 60 to choose from one or more contexts or
attributes within health conditions 304 and co-morbid conditions
306.
[0084] Store list button 318 allows healthcare provider 60 to store
the actual list of names for future recall and use for later
reference at follow up list 310. Similarly custom list 312 allows
healthcare provider 60 to store the parameters that are used to
create the list, for future reference and follow up. Export list
button 316 allows healthcare provider 60 to send the list to other
programs within system 10, including monitoring application 22,
task scheduler 24 and content assignor 28. Export list 316 may be
used to assign specific dialogs 100 to individuals 70 by directly
using the report generator interface. Additionally, specific tasks
may be queued within the task scheduler 24 in order to execute them
(such as delivering a dialog) at a later date. Custom list 312 may
be exported to monitoring application 22 to create a new protocol,
in order that all future individuals 70 fulfilling the said
criteria are automatically assigned the specified dialog 100, or
action 140.
[0085] Report type 314 menu allows healthcare provider 60 to chose
from among a set of available types of reports. Different types of
reports are suited to depicting different aspects of information
about the patients and group of patients. One possible report type
is a risk map, whereby the risk factors of selected individuals 70
are depicted as a status map on the basis of their computed health
risk. Another possible report type is a temporal graph, whereby
lines are drawn showing the temporal progression of the computed
risk in one or more contexts. Color coding may be used to depict
different patients and contexts separately. A third possible report
type is one where individuals who are most in need of medical
attention are selectively listed by priority, urgency, name or any
parameter within database 40, using arithmetically computed risk
factor variables. A fourth type of report is a detailed patient
report where all the relevant parameters within the database is
listed, alongside the analysis reports of the content assignor, and
the rationale behind the automated script assignment.
[0086] FIG. 10A is a flowchart providing an overview of the various
steps involved in the system-individual interaction, described from
the level of remote health management system 10. FIG. 10B is a
continuation of the flow chart of FIG. 1 OA.
[0087] In step 502, monitoring application 22 is activated after
the establishment of a successful connection with remote terminal
72. In step 504, replies 134 are received to previously sent
dialogs. In step 506, actions 140 are performed at the level of the
system on the basis of received replies 134. These actions include
amending the profile database 44, adding new tasks to tasklist
database 46, and serving new dialogs 100 to individual 70. In step
508, content assignor 28 scans profile database 44, and assigns
dialogs 100 to individual 70 in the order of decreasing importance
in health management. In step 510, an additional list of dialogs
100 is received from tasklist database 46. In step 512, healthcare
provider 60 may additionally use report generator 34 to add dialogs
100 to that from steps 508 and 510. Alternatively, healthcare
provider 60 may add dialogs 100 to the tasklist database 46, which
are sent to individual 70 at a future time.
[0088] In step 514, monitoring application 22 groups the dialogs
into a logical sequence and exports an ordered and often truncated
list to script generator 32. In step 516, script generator 32
references profile database 44 to elicit information on the type of
remote terminal 72, monitoring device 74, and communication network
50 that will be used to communicate with individual 70.
Additionally the visualization and content presentation preferences
of individual 70 are referenced in step 518. In step 520, script
generator 32 references dialog library 48 and elicits relevant
dialog content 130 (including datafiles 136) that is compatible
with communication network 50, remote terminal 72 and monitoring
device 74.
[0089] In step 522, script generator 32 may additionally
communicate directly with remote terminal 72 and monitoring device
74 to elicit the characteristics of the devices and stored
individual preferences, and parameters within, in order to
customize the script. In step 524, script generator 32 creates a
customized program. Finally, in step 526, script generator 32
exports the customized script program to monitoring application
22.
[0090] In step 528, monitoring application 22 sends the script
program to individual 70 via communication network 50. Monitoring
application 22 waits until a specified time `T`, typically between
24 to 48 hours, for individual 70 to respond to the script program
(step 530). In case responses are received, the application
continues along step 504. Alternatively, if replies are not
received within time `T`, monitoring application adds information
regarding the non-receipt of replies to profile database 44 (step
532). This information may be accessed and used separately by
report generator 34, task scheduler 24 and content assignor 28.
[0091] Content Assignment and Data Analysis Methods:
[0092] The system works on the following premise: since each of the
patients managed by the system has an individualized treatment
protocol which is based on the unique health related parameters of
the patient, ideally each patient must also be followed on an
individual basis by the healthcare provider. However, this approach
is inefficient in that it would require the health provider to
spend considerable time following those patients who otherwise
present little risk as far as compliance with medical advice is
concerned. In addition, these patients are usually well motivated
and likely to follow medical advice, and the management would
primarily consist of confirming and maintaining their low-risk
status.
[0093] On the other end of the spectrum is the other group of
patients who are less motivated to look after themselves, and less
likely to modify their disease related behaviors on their own
accord, and without continued encouragement and support from the
healthcare system. These patients need regular monitoring of their
health related parameters, sensitization to the need to modify
disease related behaviors, continued psychological support, health
education and encouragement to comply with prescribed treatment
regimens in order to prevent their disease condition from
worsening. This group of patients is responsible for a greater
proportion of total morbidity and mortality from chronic illnesses
than the rest of the population.
[0094] The method of managing patients varies significantly between
the two groups of patients. While the first group may be managed by
assigning content that is dynamic and individualized to the patient
profile automatically, the second group requires more active
participation and involvement of the provider in the management
process. In this second group, the provider would need to probe
deeper and develop novel strategies in order to treat the
individual's condition.
[0095] The two groups are, however not mutually exclusive. Some
patients who are initially poor risks may be gradually motivated
and encouraged so that they acquire the low risk status of the
first group. Likewise, it is also necessary to ensure that patients
belonging to the first group continue participating actively, and
remain involved in the upkeep of their own health, so that they do
not acquire the risk profile of the second group.
[0096] Stated simply, the system consists of a consists of a
patient profile, which contains details of the patient's condition,
and details of co-existing conditions (co-morbid conditions) that
may have a bearing on the way the patient is managed by system 10.
These details are expressed in the form of Risk factors (R) that
numerically depict the level or state of risk a patient is in, for
each of the numerous aspects of disease (context).
[0097] The profile may be formed on the basis of an initial dialog
(registration dialog 100R). In addition, profile data may be
derived from information contained within the patient's medical
records database (clinician's notes), claims database (from the
patient's insurer, employer, managed care organization, care
provider, etc) and additional data sources (laboratory data, etc).
This is preferably done by computer programs that use Natural
Language Processing algorithms to extract data from the diverse
sources to determine an initial profile. However, the patient's
initial profile may also be created, atleast in part, by the
healthcare provider manually.
[0098] Information contained within the profile is used to assign
content in the form of `dialogs 100`. In assigning content, the
system takes into account the relative importance of treating that
particular context of disease (P), the level of risk or the risk
state (R) the particular patient is in, and a correction factor (F)
that allows comparison of the R and P of different contexts.
[0099] Replies to the content are used to make changes to the
profile, and further content is assigned on the basis of the
updated profile. In this manner, a dynamic feedback loop is created
whereby one day's results change the patient profile, and the
updated patient profile is used to select the most appropriate
content to fit the need.
[0100] This simplified view depicts the underlying basis of content
assignment. However, in order to improve the effectiveness of the
system, it would be necessary to take into account
[0101] 1. Reliability of the data provided by the patient. If the
data provided by the patient is inherently unreliable, then health
education, analyses, and medical advice provided on this basis
would be fallacious.
[0102] 2. Resistance of the patient to modify health related
behavior. Resistance is inversely proportional to the level of
motivation in the patient. A well motivated patient is more open to
changing health related behaviors, with a view to improve his/her
health status. However, motivation per se is not static, and it is
possible to increase the patient's level of motivation (openness to
change) by providing information that is in keeping with the
patient's motivational drivers (emotive or driving force that is
behind all human actions)
[0103] 3. The patient's attention span. Patients have only a
limited attention span, which further varies on a day-to-day basis,
depending on the patient's other commitments: and a health
management system that doesn't take this into account at content
delivery is bound to suffer from poor compliance and continuation
rates. Educational dialogs would need to be small enough to deliver
the message within the patient's attention span, yet be capable of
delivering more information on a specific aspect wherever requested
by the patient.
[0104] 4. The patient's personal preferences and type of remote
terminal 72. In order that the system be most effective, it is
preferable that the content presentation be customized to the
patient's preferences and utilize the data display capabilities of
the remote terminal to the fullest extent possible.
[0105] Method to Prioritize in Automated Content Assignment:
[0106] Let
[0107] C be the context or attribute of the patient's health
condition which is computed and followed by the system, and is the
basis of which further automated content assignment. C in a patient
with diabetes can include the blood glucose control attribute,
medication compliance attribute, feet care attribute and long-term
complications attribute.
[0108] N be the identification number of a patient.
[0109] R be the risk score that is computed in a particular context
for a particular patient at time T
[0110] T is the time designator of a particular point in time.
[0111] M, the modifying factor is applied to R on the basis of
replies received to queries
[0112] L be the reliability index of any given R value.
[0113] Further, let P be the priority sequence value of a context
within any health condition.
[0114] F be the care provider defined correction factor for R
values of any specified context. F values vary with the R scores of
the specified context.
[0115] Z, the cumulative sequence is obtained by multiplying R, P
and F. for a context.
[0116] W, the resistance factor indicates the degree to which a
patient exhibits resistance to prescribed therapy.
[0117] Risk Scores (R Values)
[0118] R values signify the risk state or level of the patient. The
R variable depicts the risk state or level of risk the individual
within a particular context. It is also possible to create
cumulative R values that are derived either arithmetically, or
through the use of logical expressions, from different variables in
the profile. Variables within the profile database, including the R
value may be discrete or continuous variables.
[0119] Discrete variables include the individual's actual responses
to queries, and are stored as such within the profile database,
whereas continuous variables are arithmetically derived from the
replies of the patient. In those instances where continuous
variables are used in a context, the profile database additionally
records the discrete variables, and the methods used to compute the
continuous variable (numerical figure). Variables can also be a
logical expression or a binary state.
[0120] When a patient is first registered into the system, he/she
is assigned Risk values on the basis of data from medical records
database 62, medical claims database 58 and additional material 54,
by the use of Natural Language Processing Algorithms. Further, R R
values are subject to numerical manipulation on the basis of
replies 134, measurements 74D, besides that from data contained
within medical records database 62, medical claims database 58 and
additional material 54.
[0121] R can be any positive decimal number greater than Zero. In
case where R values have not been assigned from any of the
abovementioned sources, R takes the default value of 1. A patient
with greater R values for a given context has a higher risk of
developing health related complications pertaining to that context.
Conversely, a patient with a lower R value is better placed than a
patient with a higher R value with regard to that particular
context.
[0122] In the instance where R is a discrete variable, it instead
takes on the value of a state, the R state. R states include text
strings (`Feeling great`, `Not feeling so good`, `Terrible`), or a
logical value (`Yes`, `No`, `Not sure`).
[0123] R values are also serially followed in time (Time designator
`T`). Newer R values are usually mathematical functions of recent R
values, and are directly linked to replies through the means of
actions 140. For instance, healthcare provider 60 may create dialog
100 such that when individual 70 chooses one particular reply, the
new R value equals the recent most R value multiplied by factor
`M`, where M is any positive number.
[0124] ACTION 1: IF REPLY=`YES` THEN NEW R=OLD R multiplied by
0.5
[0125] ACTION 2: IF REPLY=`NO` THEN NEW R=OLD R multiplied by
2.0
[0126] In the instance where discrete variables are used, the
actions may be so modified that,
[0127] ACTION 1: IF REPLY=`YES` THEN NEW R=`Yes`
[0128] ACTION 2: IF REPLY=`NO` THEN NEW R=`No`
[0129] If assigned value `M` for a particular reply is greater than
one, then, the risk R increases, since any number multiplied by a
positive number greater than one increases in value. On the
contrary, if M lies between 0 and 1, the R value would decrease.
So, provider 60 would assign M values to replies 134, in the form
of actions 140 based on what a particular reply signifies in any
individual's health condition and context. However, it would also
be possible for healthcare provider 60 to `reset` R to default, or
any other value. The mathematical function applied here is:
[0130] ACTION 3: IF REPLY=`NOT SURE` THEN NEW R=1.0
[0131] ACTION 3: IF REPLY=`NOT SURE` THEN NEW R=`Unknown`
[0132] Even as new R values are generated by the system as a result
of replies 134 to queries and measurements 74D from monitoring
device 74, older R values are archived within the profile database.
Serially following R values of a patient is used in determining the
progress made in the management of the patient's condition, in
comparing the success of management protocols in the given patient,
and in monitoring the progression of the chronic disease that is
irrespective of management.
[0133] R values of different contexts for a given patient vary
independently of each other. However, it would be possible for the
provider to create summary R values that are a mathematical
function of two or more R values.
R.sub.CUMULATIVE=R.sub.1*R.sub.2*R.sub.3*R.sub.4
[0134] In the instance where discrete variables are used,
[0135] IF REPLY1=`YES`; REPLY2=`NO`; REPLY3.gtoreq.300,
[0136] THEN R.sub.CUMULATIVE=`High risk`
[0137] ELSE IF REPLY1=`NO`; REPLY2=`YES`; REPLY3.gtoreq.300,
[0138] THEN R.sub.CUMULATIVE=`Moderate risk`
[0139] ELSE IF REPLY1=`NO`; REPLY2=`YES`; REPLY3.ltoreq.300,
[0140] THEN R.sub.CUMULATIVE=`Low risk`
[0141] ELSE R.sub.CUMULATIVE=`Error: Incompatible data`
[0142] Reliability Index (L Values)
[0143] L, the reliability index of a given R value is indicative of
level of confidence with which the R value depicts the actual risk
faced by the patient within a context. L is proportional to the
total quantum of data pertaining to a context that has been
inputted, and subsequently used by the system in the computation of
R values. L values for an R value assigned by default would equal
zero, since no data was used in deriving the default value. It
would also be possible for individuals 70 to be assigned negative L
values indicating poor reliability of the veracity of the
individual's responses.
[0144] L values are to ensure standardization and uniformity of R
values, comparability of R values of different patients, and to
correct a possible source of error within the system. Consider the
following example; two diabetics, patient A and patient B with
similar R values in the context of blood glucose control have been
on follow up for 2 years, and 1 month respectively. Surely the
confidence that may be placed by provider 60 on the R values of
patient A is more than that of patient B, since considerably more
data has been used in computing the R value of patient A. Patient A
may have been a high risk patient who has subsequently modified his
disease related behaviors, while patient B may have been assigned
her R value by default. If one were to compare the results of a
particular intervention, such as a new drug in the two patients
without considering their L values, there would likely be
fallacious results.
[0145] Further, R values are also used in decision making when
replies are received. For instance, say patient A is assigned a low
risk in the context of blood glucose control. If an isolated blood
sugar measurement 74D is returned as high, then the test result may
even be ignored along with the additional assignment of the task of
repeating the test after a few days. On the contrary, if patient B
were to return the same high blood glucose test result, considering
that the validity (L value) of his/her R value is still uncertain
(a default R value gives no information at all), the further follow
up would likely be different. In this case, the assignment of an M
value less than one would be appropriate.
[0146] It is obvious that as and when more data is entered into the
system on a particular context, the L value for the particular
derived R value would increase, since more computations would have
been performed on the original R value to derive the current R
value. It would be additionally possible for a provider to exclude
R values from the automated decision making process when it's
corresponding L value is below a defined value.
[0147] It is additionally possible to program L values to decrease
when responses from individual 70 are inconsistent over time and
incompatible with existing available data in the system. This is
particularly relevant in the diagnosis and treatment of psychiatric
conditions. In certain psychiatric conditions, such as Munchausen's
syndrome, patients tend to live in a false world of their own
creation. It becomes very difficult to diagnose the existence of
this condition in the first place since the patients seem to be
perfectly at ease with their own perception of events, which is
however, grossly inconsistent with reality. Strategies employed by
mental health professionals include repeating the same question to
the patient after a span of time, repeating the query in another
context and rephrasing the query and putting it to the patient.
When gross inconsistencies are observed in any of the above, the
suspicion of a disease process increases.
[0148] The same strategy is implemented in the system in the
following manner: consider dialog D that queries the patient about
a particular aspect of his/her life. Provider 60 simultaneously
creates three more `clone` dialogs, D.sub.1, D.sub.2 and D.sub.3
which elicit the same information, albeit by putting the queries in
a different manner, or as variants. Then provider 60 schedules
dialogs D.sub.1, D.sub.2 and D.sub.3 to appear on different days,
and in different locations within a daily session (the beginning,
the middle or the end) and interspersed between other queries. The
interval between the patients' undergoing the two `clone` dialog
sessions may be days or even weeks.
[0149] Replies to D.sub.1 are stored in the profile database.
Additionally, actions may also be programmed to modify the value of
R within a given context. Alternatively, the replies to D.sub.1 may
be directly compared with prior data within profile database for
coherence, consistency and compatibility in view of R. When replies
to D.sub.2, and D.sub.3 are received, they are directly compared
with that of D.sub.1. In case there is a significant variation in
the replies to the three dialog `clones` the L values will
decrease, and, as a consequence it is determined that data received
from the patient is not entirely reliable.
[0150] This would cause the healthcare provider to discount earlier
data that the patient may have supplied, including that in a
different context of disease. Further, it may be considered
necessary at this stage to corroborate patient supplied information
with that from the patient's friends, relatives, and significant
others who know the patient since the unreliability of patient
supplied information is established.
[0151] Content Assignment Mechanism
[0152] Different contexts represent different aspects of the
patient's condition. Likewise, these contexts occupy different
priorities in the management of the disease process. Some contexts
would need to be managed expediently, while others, though
pertinent to the disease management process, could be managed at a
later date.
[0153] Consider patient A in the diabetes example above. Patient A
is also a smoker, and smoking cessation encouragement modules are
particularly relevant in this patient, since diabetics who smoke
have a very high risk of developing coronary arterial disease. Thus
smoking cessation remains a high priority. However, if patient A
were to additionally have trouble controlling his/her blood sugar
levels (measurements repeatedly return in the higher range) then
controlling the high levels of blood sugar takes precedence over
smoking cessation. On the contrary, if measurements 74 of patient
A's blood glucose were to widely fluctuate with some values in the
lower value range (this occurs in a variant condition called
brittle diabetes) then preventing low blood glucose (hypoglycemia)
is of more priority than preventing occasional high blood glucose
values. This is because, in the presence of hypoglycemia, the brain
starves of energy, and this can permanently damage the cells of the
brain. Thus it is obvious that even within a disease condition,
there are some contexts occupy a higher priority in immediate
management (and in the context of the system, in immediate content
assignment, since the attention span of the patient is limited)
[0154] Within the system, the provider assigns higher priority
sequence values (P values) to those contexts, which if test true
for high risk behavior, signs or symptoms, require immediate
management and intervention, while that which may be managed at a
later date are assigned lower values. P value can be any positive
number, and is assigned by provider after considering the P values
that have been assigned to other contexts within the same and
related health conditions. Similarly, educational modules that deal
with contexts which are assigned higher P values are automatically
assigned higher P values than that of other educational
modules.
[0155] The consideration of the R and P values of all the contexts
of a patient's condition is otherwise sufficient to manually assign
content by priority. However, in order to automate the content
selection process and make the risk values of different contexts
comparable, it is necessary to include a correction factor, F that
varies with the computed R for each of the contexts. F values are
assigned by the provider for each range of R values of a context C.
This information is stored in a lookup chart, such as Chart 1
below:
1 CHART 1 R-F lookup chart CONTEXT, C P value FOR R= F= Explanation
Hypoglycemia 100 <1 1 >1.0 risk is assigned prevention 1-1.99
30 when there have been >2 100 one or more such episodes (as
determined from dialogs) Glucose 90 <1 1 control 1-2.99 20
3-6.99 35 7-10.99 80 >11 200 Smoking 50 <0.1 0 <0.1
signifies that cessation 0.1-0.99 10 patient is not a smoker.
1-4.99 25 F = 0 is to make sure that 5-9.99 40 the patient never
receives >10 90 smoking cessation modules, as it is irrelevant.
Weight 45 <1 0.1 <1 means the patient reduction, 1-2.99 10 is
maintaining optimum 3-9.99 30 weight, and any advice >10 60
would only be towards encouraging weight maintenance Exercising 30
<1 0.1 1-5 4.0 >5 8.0 Foot care 30 <1 0.1 1-5 4.0 >5
8.0 . . . more . . . . . . . . . . . . contexts The contexts
depicted above are for the purpose of illustration only. By no
means must it be considered comprehensive. In the actual diabetes
treatment program, monitoring would be done on a far greater number
of contexts than that depicted here, and in charts 2 and 3.
[0156] The hypoglycemia prevention context is assigned a very high
P value=100, and when R>2, the F value equals 100. Similarly in
glucose control, when R=8, F equals 80.
[0157] In prioritizing at the time of content assignment, content
assignor 28 uses Z values which are numerically derived by
multiplying R, P and F for each of the contexts. Dialogs
corresponding to the context with the highest derived Z value are
assigned to the patient at the start of each new session. However,
in prioritizing content assignment for those contexts where R
exists as a state (logical expression, text string or value), the
above expression will not hold true. In this instance, the provider
has the following options
[0158] Firstly, the healthcare provider may arbitrarily assign R
values for each of the possible R states. For example, the R state
`Feeling great`, to the query `How are u feeling today` to a
patient with depression may be assigned a low R value, since it
represents a better prognosis in a patient with depression; and
`Feeling terrible` may be assigned a high R value.
[0159] Second, the healthcare provider may write the follow up into
the actions itself. In this case, if the patient replies `Feeling
terrible`, he/she is asked further queries that attempt to elicit
the cause of his/her low mood. Alternatively the healthcare
provider may create an action whereby the provider is immediately
alerted to the fact of such a reply, so that it may be followed up
by a telephone call, a personal visit, or the scheduling of an
appointment.
[0160] Third, the healthcare provider may so program the protocols
database such that when the R state for a particular context in a
patient returns certain value, text string or logical expression,
the protocols database automatically adds a specified dialog to the
tasklist database for that individual.
[0161] Fourth, the healthcare provider may use the reporting
interface and manually assign dialogs to those individuals whose
profile database returns true for a specified R state, or a
combination of R states.
[0162] Chart 2 (Z-value chart) shows the derived risk factors, and
computed Z values for the same patient A at point of time=`1
November 2001`
2 CHART 2 R value = Corresponding (From F = Z value = Context, C P
value= profile data) (from Chart 1) P * R * F Hypoglycemia 100 2.8
100 28000 prevention Glucose 90 2.5 20 4500 control Smoking 50 2.2
25 2750 cessation context Weight 45 2.1 10 945 reduction Exercising
30 4 4 480 Foot care 30 2.5 4 300
[0163] In patient A, the top priority (and the assigned content
dialogs) would be towards preventing the possibility of a
hypoglycemic episode (Z=28000). Next in priority would be to
control the patient's blood glucose levels (Z=4500), and so on. At
this point of time, automated content would primarily focus on the
above two contexts, though `maintenance` content would be assigned
for the remaining contexts, namely smoking cessation, weight
reduction and foot care. Further, regardless of the priority
assigned to any context, when significant resistance is encountered
(high W values), the automated content assignor provides only
maintenance content for that context.
[0164] Over the next few weeks, the patient interacts with the
system, and as a result, learns more about his/her condition, and
is motivated to change some of his/her health related behaviors. It
is also possible, especially in the case of diseases that run a
more rapid course, that there has been progression of the condition
in the meantime, which is independent of the medical management. As
a result of the above the R values of the patient change in either
direction for different contexts. At this point of time, the
Z-value chart would read in the manner of chart 3 given below
3 CHART 3 R value = Corresponding (From F = Z value = Context, C P
value= profile data) (from Chart 1) P * R * F Hypoglycemia 100 0.9
1 90 prevention Glucose 90 1.6 20 2880 control Smoking 50 1.5 25
1875 cessation context Weight 45 1.5 10 675 reduction Exercising 30
3 4 360 Foot care 30 2.1 4 252
[0165] Here, the highest priority is given for (and consequently,
maximal dialogs assigned would relate to) glucose control Z=2880.
Following this, priority would be assigned to smoking cessation and
weight reduction respectively. The reduction in the R value of
hypoglycemia prevention context is likely the result of patient
education about the symptoms of hypoglycemia and the techniques
that are used to prevent hypoglycemia when self administering
insulin. Z values that are at this level (z=90) may be safely
managed by `maintenance dialogs`, which reinforce in the patient,
at regular intervals, the need to avoid hypoglycemia. Similarly, if
significant resistance were encountered to smoking cessation
context meanwhile, then the system would revert to `maintenance
dialogs` in that context. In the above example, the Z values for
glucose control have also reduced from 4500 to 2880, but this still
represents an area of high priority, and at this juncture, a
significant portion of the content would still relate to this
context of health care. The R value for foot care has decreased
from 2.5 to 2.1, but with a Z score of 252, this remains one aspect
of the disease that would require the provider's attention.
[0166] For reasons of clarity, the above example illustrates the
change in assigned content after the passage of a span of a few
weeks. However, it must be understood that in the actual system,
dynamic updating of the profile and content assignment would take
place in real time, such that the system responds in real time to
the changing profile of the individual patient. Maintenance dialogs
are communicated with patients who are otherwise at low risk states
for a particular context, in order to ensure the patient's
sustained interest in the disease management process, as also to
reinforce in the patient knowledge and correct attitudes regarding
the disease context.
[0167] Measurements 74D from monitoring device 74 may be used, in
addition to modify the computed R values by assigning `M` values to
ranges of values of 74D. In the case of diabetes, monitoring device
74 may be a glucometer, and dialog 100 may be so programmed that
when blood glucose levels are above a certain value, the old R
value (R.sub.GLC-OLD) gets multiplied by a certain factor M1 to
yield a new R value (R.sub.GLC-NEW). In the logical expression
given below, M1 is the returned post dinner blood sugar value from
monitoring device 74 of patient A, in milligrams per deciliter
(measurement 74D).
4 ACTIONS: IF M.sub.1 >500, THEN R.sub.GLC-NEW=R.sub.GLC-OLD *
3.5; THEN FOLLOW PREACUTE DIALOG 1001. ELSE IF
300.ltoreq.M.sub.1.ltoreq.500, THEN R.sub.GLC-NEW=R.sub.GLC-OLD *
1.8; THEN FOLLOW EDUCATIONAL DIALOG 16. ELSE IF
200.ltoreq.M.sub.1.ltoreq.300, THEN R.sub.GLC-NEW=R.sub.GLC-OLD *
0.8; THEN FOLLOW CONGRATULATORY DIALOG 22. ELSE IF M.sub.1<200,
THEN FOLLOW LOW PD-GLUCOSE DIALOG 11. IF LOW PD-GLUCOSE DIALOG
11=POSITIVE, THEN R.sub.HYPOGLY-NEW=R.sub.HYPOGLY-OLD * 1.6
[0168] If the measured blood glucose values exceed 500 mg/dl, then
the risk factor increases by a factor of 3.5 and PREACUTE DIALOG
1001 is followed. If the measured blood glucose were to be between
300 and 500 mg/dl, then the risk factor would increase by 1.8 and
an educational dialog would follow. However, if the returned values
were in the normal range (200-300 mg/dl) a congratulatory statement
would follow. If the values were on the lower range, a hypoglycemia
prevention statement would follow. In this manner, it is possible
to detect a developing complication and intervene early, in order
to prevent the condition from worsening and presenting at a stage
where it is more difficult and expensive to manage.
[0169] In addition, R values can be modified on the basis of the
results of laboratory tests, data contained in medical records
database 62, medical claims database 58 and additional material
database 54. In the diabetes example, a lipid profile test done at
the laboratory of the healthcare facility or elsewhere (additional
material database). Similarly the failure of the patient to keep up
an appointment at healthcare provider 60 can also be used to
increase the R value for a context.
[0170] Resistance Factor (W)
[0171] W, the resistance factor is particularly relevant in the
management of chronic illnesses, where there is a tendency for
patients to become non-compliant with medical regimens with the
progression of time. While individualization and customization of
the dialogs and management plans helps to prevent this to a great
degree, it is also necessary to identify and quantify developing
resistance at an early stage, and suitably modify the mode of
management. W, the resistance factor quantifies the degree to which
the patient is non-compliant with the treatment plan, whatever be
the reason, in order that alternative approaches are explored.
[0172] Resistance in a particular context is inversely proportional
to the motivation of the patient with regard to the particular
context. Further, this motivation is amenable to change, as a
result of information provided to the patient on the importance of
that context in the entire health management process. So, when a
patient is exhibiting high resistance in a context, the dialogs
assigned to the patient take on the `maintenance mode`.
`Maintenance dialogs` differ from the `conventional dialogs` in
that these dialogs are
[0173] 1. presented to the patient at a frequency that is less than
other contexts with similar Z values,
[0174] 2. intended to stimulate the patient's interest and
curiosity, rather than stimulate the patient to change his/her
health related behavior,
[0175] 3. intended to convey information in a non-opinionated
manner, and allow the patient to draw his/her own conclusions,
and
[0176] 4. atleast one reply to each query or teaching statement
allows the patient to terminate the dialog relating to that
context.
[0177] Chart 4 is similar to Chart 3, with the addition of a column
depicting the resistance (W) values for each of the contexts.
5 CHART 4 R Corre- value = sponding From F values profile (from Z
value = Context, C P value= data Chart 1) P * R * F W value
Hypoglycemia 100 0.9 1 90 2.6 prevention Glucose 90 1.6 20 2880 2.5
control Smoking 50 1.5 25 1875 25.1 cessation context Weight 45 1.5
10 675 0.5 reduction Exercising 30 3 4 360 0.5 Foot 30 2.1 4 252
1.0 care
[0178] The W values for smoking cessation are quite high (25.1) in
this patient. So the patient is put on maintenance dialogs for this
context till the issue is resolved by a patient-provider personal
interaction, or the W values show a decreasing trend as a result of
the maintenance dialogs. A pro-active stance of attempting to
achieve total smoking cessation is deferred until then, for the
abovementioned reasons. If it is determined, as a result of the
patient-provider interaction(s), that the patient will actively
work towards achieving complete cessation, the W values and/or R
and L values may be `reset` to default levels by the provider, by
accessing profile database directly, or by the use of a custom
action at dialogs 100.
[0179] As is evident from Chart 4, resistance values are
independent of R, P, and F values. This is because the method for
the derivation of W values is different from the method of
derivation of the remaining values. R depicts the risk status of
the individual within a context, while W has more to do with the
level to which the patient is motivated (or resistant) to the
context under question.
[0180] Consider another example: Patient H is on follow-up for
coronary arterial disease (clogging of the blood vessels, the cause
of heart attacks). Patient H is also a chronic smoker, and it is
well known fact that quitting smoking remains the single best
intervention in this patient that will reduce the patient's risk of
developing heart attacks and strokes in the future. So the content
assignor 28 would commence intensive efforts to achieve smoking
cessation in this individual, by the use of education modules.
Hopefully, this will cause the individual to quit smoking. However,
a subset of patients may be put-off by these messages, and consider
them intrusive to their lifestyle, which may lead to an increased
dropout rates and non-compliance with the remainder of medical
management provided by the system. In these patients, smoking
cessation still remains a long-term goal but it is recognized that
conventional dialogs may prove to be counter-productive. Instead,
maintenance dialogs such as the one depicted below, are
communicated to the patient.
[MAINTENANCE DIALOG IN A RESISTANT SMOKER]
[0181] 1. A RECENTLY CONCLUDED STUDY CONFIRMS THE FACT THAT SMOKERS
ARE TWICE AS MORE LIKELY TO DEVELOP IMPOTENCE DUE TO VASCULAR
CAUSES THAN NON-SMOKERS.
[0182] A. OK [ACTION: GO TO NEXT DIALOG]
[0183] B. MORE [ACTION: W.sub.SMOKER-NEW=W.sub.SMOKER-OLD*
0.8;]
[0184] [ACTION: GO TO DIALOG `SMOKER-IMPOTENCE-101`]
[0185] C. SKIP [ACTION: W.sub.SMOKER-NEW=W.sub.SMOKER-OLD* 1.5]
[0186] [ACTION: GO TO NEXT DIALOG]
[0187] The patient is informed about the harmful effects of
smoking, more as a teaching statement or a news headline, than as
an attempt to induce him/her to quit smoking. The idea here is to
provide information, and stimulate the patient's interest, and to
arouse the patient's curiosity. Responses to statement include
`Ok`, `More` and `Skip`. If the patient chooses `Ok`, the next
dialog within the program is served to the patient. If the patient
chooses `More`, it means that the patient is curious about this
aspect. So, in addition to providing the patient more details of
the study, the value of resistance in the context of smoking
cessation is reduced by a multiplying it with 0.8. On the contrary,
if the patient decides to `Skip` the statement, the value of
resistance is increased.
[0188] When resistance increases above a certain
operator-determined value, it is obvious that the problem is
unlikely to be solved by remote dialogs alone, and alternative
approaches are explored, including bringing up the topic at a
personal interaction, such as an interview, or a telephone
conversation.
[0189] After any dialog is presented to the patient, a marker is
placed in the profile of the patient regarding the same, in order
that the content assignor does not automatically present the same
dialog to the patient more than once. However, the same dialog may
be presented to the patient at the discretion of the provider.
[0190] Methods to Avoid the Development of Resistance--
[0191] Dynamic dialog paradigm--In this paradigm, possible replies
to query include `Skip Query`. When this option is chosen by the
patient, the program at remote terminal 72 proceeds to the next
query, and a marker indicating the particular reply is added to
profile database 44. Additionally, a task may also be added to the
tasklist database 46. This task can include serving the dialogs to
the patient at a later date and in an alternative format.
[0192] When implemented, dynamic dialog paradigm allows the patient
to dodge a query. This is because; some queries can be very
personal or embarrassing to patients. Further, not all patients are
forthcoming when initially queried regarding such topics as
domestic abuse, usage of psychotropic substances, etc. In diseases
such as mood disorders, it is not uncommon for patients to abuse
psychotropic substances in a misguided attempt to `treat` their
condition. Though the topic may be better discussed with the
patient at a personal interview, it is advantageous to screen
patients using the remote monitoring system, to identify those
patients in whom this aspect needs to be further elucidated.
[0193] By allowing the patient to `dodge` a query, the healthcare
provider identifies those individuals who have reasons not to reply
to the query. In the instance that a patient `Skips` a query, this
aspect is either explored at a future personal visit, or by a
dedicated dialog for eliciting responses from this subset of
patients.
[0194] If this option were not available to the individual, it is
likely that a considerable proportion of patients would simply deny
the condition, thus making the data unreliable. Thus, dynamic
dialog paradigm has the net effect of increasing the reliability
and veracity of the data entered into the system and available to
the healthcare provider.
[0195] Online dialog dynamism--In order to sustain the patient's
interest and continued attention in health management and
educational materials, it is necessary that the content adapts
itself to the attention span and interest of the patient. It is
further necessary that patients progress through the program at
their own pace, and that the content matches evolving patient
status or the evolving understanding of the patient condition.
Since the patient's perception of the health benefits to be derived
from the system tend to vary with time, it is important that the
focus of interaction be dictated by the user wherever possible and
that the dialog is responsive to the individual in real time
(Online dialog dynamism) For example, on a particular day,
individual 70 might not be keen on completing his/her educational
module, while on another day, a topic in the module might stimulate
the individual's interest, and he/she may wish to learn more about
it.
[0196] Online dialog dynamism allows patient's replies to change
the presented content in real time without the invoking of
communication system 50. Replies to a teaching statement can
include `Ok, `More` and `Skip`. If the user selects `Ok`, the
teaching statement proceeds along the originally intended course.
However, if `More` is chosen by the patient, additional latent
teaching statements from content 130 are invoked. On the contrary,
if `Skip` is chosen, the teaching statements are terminated, and
are invoked at a later date. `Skip` invoked within online dialog
dynamism will not normally modify the resistance factors, though it
is possible to program it to do so.
[0197] Customization--At the time of registration of the patient
into the system, the patient is queried regarding the preferred
remote terminal 72 and monitoring device 74 with which the patient
is to communicate with the system. Other preferences include
whether the patient prefers audio, videos and other multimedia
components (including video games) as a part of the content
(wherever supported by the remote terminal and communication
network 50). Another point where customization may be implemented
is in the frequency, timing and total amount of dialogs served to
the patient, as a part of the interaction.
[0198] Additionally, customization may be in relation to the scheme
and mode of presentation of content on the display screen of remote
terminal 72. In the preferred embodiment, the customization
components are stored in profile database 44 in the profile
computer 20. In an alternative embodiment, these components are
stored in remote terminal 72 and/or monitoring device 74, and
script programs received from profile computer 20 are automatically
customized at the level of the terminal 72, and device 74, and then
presented to the patient.
[0199] Customization is aimed at making the system more
user-friendly and capable of optimally utilizing the data transfer
capabilities of communication network 50, and the display and
processing capabilities of remote terminal 72 and monitoring device
74. In one embodiment, the patient provides the details of his/her
devices and communication network. For example patient may be asked
to choose among a list of supported devices and type of content. In
an alternative embodiment, the profile computer 20 detects the type
of remote terminal 72, monitoring device 74 and communication
network 50, and optimizes the content for delivery. For example
patients who are connected by a low-speed 36 kbps modem would have
lesser amount of multimedia content served to them than those
patients in whom communication is established by high speed
internet. Specific techniques of implementing this are well known
in the art.
[0200] A few examples of specific instances where customization of
the dialogs to the remote terminal 72, monitoring device 74 and
communication network 50 is particularly advantageous are given
below.
6 PDA (Personal Digital Assistants)- Integration of Patients may be
reminded to take medications on time, script programs with the in
the form of alerts. daily task planner Appointments with the
healthcare provider may be program within the automatically
scheduled on the basis of entries in the planner. PDA. Serial
recording Graphical representations of the parameters, e.g. blood
of parameters measured glucose levels may be created at the level
of the terminal 72 by device 74. itself, with regular updating of
the profile computer. The patient may communicate with the system
less often, with the one-time transmission of stored data in the
intervening period. Symptom diary A patient with depression,
epilepsy, asthma, etc may maintenance create entries, and even
store measurements 74D in a symptom diary program that stores the
data within the terminal 72. At regular intervals, or when the
local processing algorithm determines that the patient is at
increased risk, he/she is advised to contact the healthcare
provider immediately. In the case of asthma, the additional
advantage is that the daily activities of the patient may be
correlated with the PEFR (Peak Expiratory Flow Rate) values of the
patient, which is an objective measure of the severity of disease,
in order to find what factors actually are responsible for the
exacerbation of the patient's condition. Additionally,
subjective-objective studies may be performed whereby, for example,
the PEFR is compared with the patient's subjectively reported
symptoms. Digital TV- Enhanced rich There are no limits as to the
bandwidth of transmitted multimedia content content within Digital
TV, and rich multimedia and audio suited to the patient's
containing data streams may be transmitted. disease state. Patients
who need to exercise are shown provided daily videos of weight
reducing exercises to help them in their workouts. This may be
scheduled at the patient's convenience. Patients may be shown
videos of former smokers recounting their personal experiences with
smoking cessation, as a psychological support mechanism to
encourage them to try and quit smoking. Support may extend into the
critical quitting phase, especially where the patient is
experiencing withdrawal symptoms. PC (and also PDAs) connected to
High-speed internet- Speech Instead of receiving written queries, a
computer generation and voice generated face reads out the queries.
Replies are received by recognition programs deciphering the voice
of the patient, so that no text input is required by the patient.
The same applies also to the maintaining an electronic symptom
diary. Monitoring device 74 directly connected to communication
network 50- Automatic Bathroom scales (device 74) of patients with
measurements of congestive heart failure are connected through
terminal 72 or physiological parameters directly to profile
computer 20. An increasing weight over a with direct few days or
weeks is used to activate a script program in the communication to
terminal 72, either directly or after going through profile profile
computer 20. computer 20. A similar method is adopted in the case
of diabetes, whereby an unusually high blood sugar level is used to
activate a script program in terminal 72. In the case of
asthmatics, a low PEFR is used to activate the script program in
terminal 72.
[0201] Though specific instances are displayed for specific remote
terminals 72, with the increasing convergence of technology, it may
be possible in the near future for some remote terminals to perform
the functions that have been described with other remote terminal
72. Further, the specific instances listed above are for
illustrative purposes only, and are not meant to limit the scope of
the invention. Many other variations of the above may be generated
and transmitted to patients in alternative embodiments.
[0202] Serial Follow Up of R, L and W Values, and Its Implications
in Patient Management:
[0203] Since the R values are measured in arbitrary units, serial
follow up of the values in a patient provides far more information
than that provided by a single value. Different combinations of the
trend in R, L and W values with respect to time identifier `T`
suggest a variety of possibilities, and this may be used in patient
management. This is illustrated by factor combinations 1 through 5
below:
7 FACTOR COMBINATIONS RISK FACTORS, RELIABILITY, RESISTANCE,
MANAGEMENT R L W IMPLICATION STRATEGY 1 Isolated More More
Patient's Follow-up dialog to high risk result or or less constant
or less constant health status is determine the cause of this
rapidly worsening. Patient sudden deterioration in health
deteriorating risk may be `pre-acute`. status. To rule out
fallacious factors in one or results. Alert provider to the more
contexts. case, who may establish contact with the patient by phone
or personally. 2 Slow More or More or Probably Confirm
deterioration in R less constant less constant disease progression
whether disease is really factors in a that is irrespective
progressing, by the use of previously well of the management.
diagnostic tests and measuring patient disease related parameters,
and at personal visits to healthcare facility. Manage accordingly.
3 Patient More Increasing Patient Reduce the content on is
gradually or less constant probably feels that those contexts where
the patient deteriorating with the system is not is showing
resistance. Likewise, regard to R helping him/her increase the
content on those factors (R value much since he/she is contexts
where the patient has is increasing). doing quite well on requested
more information at the health related any time. Provide
`maintenance parameters anyhow. content`, that teaches the patient
There is a high the importance of continued likelihood of losing
follow up and compliance, even this patient to if otherwise
asymptomatic, or follow-up. feeling well. 4 Any Decreasing High
Information This aspect would value or low for that context. about
this context is need to be explored at a personal reliability in a
not reliable when interview, and after the patient's single context
obtained remotely confidence has been secured. by the system. This
context may, in fact hold the key to the diagnosis, and the
patient's condition may never be cured completely without resolving
this. 5 Any Decreasing Any Information Obtain value or low
reliability value. obtained from the information from other sources
spread over patient is inherently such as the patient's friends,
multiple unreliable. relatives, and employers/co- contexts.
workers. Consider the possibility of malingering. In psychiatric
patients, rule out underlying psychosis.
[0204] Factors Combination 1:
[0205] Progressively deteriorating (increasing) R values over
multiple contexts in a patient over a short period of time, with
more or less constant reliability and resistance suggests that the
patient's health status is deteriorating. The patient may be in the
process of developing an urgent `pre-acute` condition, which, if
left untreated may rapidly progress into full blown disease. The
first step would be to confirm the fact of the existence of the
pre-acute condition, and this is followed by the actual management
process. In some cases, however, the confirmation step may be
skipped and one may directly go to the step of intervening
urgently. Approaches for urgent intervention in the patient may
include a personal visit to the patient's home, fixing up an
appointment at the healthcare facility, a telephone call to the
patient and/or his/her loved ones, and the usage of dialogs that
enable the more intensive monitoring of the patient, atleast until
the patient is considered out of risk. The advantage of earlier
detection and intervention in the patient's condition is that
interventions done at this time are more efficacious, less
expensive, and demand fewer resources on the healthcare system--in
terms of time and personnel.
[0206] Earlier detection of `pre-acute` conditions for the purpose
of intervention is implemented by the use of standardized
questionnaires and dialogs that look for the presence of `sentinel
signs and symptoms` that represent the earliest warning signs of
deteriorating patient's health. In the case of diabetes, include an
isolated abnormally high blood glucose measurement (diabetic
hyperosmolar coma) or an episode of sweating accompanied by
palpitations and drowsiness (hypoglycemic attack).
[0207] Logical expression below depicts the PREACUTE DIALOG 1001
that is assigned when there is an isolated blood sugar test result
of greater than 500 mg/dl (precursor of diabetic hyperosmolar
coma)
[0208] PREACUTE DIALOG 1001 [M.sub.1>500]
[0209] QUERIES:
[0210] 1. DID YOU FORGET TO TAKE ONE OR MORE DOSES OF YOUR
MEDICATION?
[0211] A. YES [ACTION: SERVE DIALOG `COMPLIANCE 21`]
[0212] B. NO [ACTION: GO TO NEXT QUERY]
[0213] 2. DID YOU HAVE AN UNUSUALLY LARGE MEAL?
[0214] A. YES [ACTION: SERVE DIALOG `COMPLIANCE 23`]
[0215] B. NO [ACTION: GO TO NEXT QUERY]
[0216] 3. DO YOU HAVE ANY OF THE BELOW: FEVER/ FEELING UNWELL/
NAUSEA?
[0217] A. FEVER [ACTION: SERVE DIALOG `ACTIVE INFECTION 91`]
[0218] B. FEELING UNWELL [ACTION: SERVE DIALOG `MALAISE 92`]
[0219] C. NAUSEA [ACTION: SERVE DIALOG `ACTIVE INFECTION 92`]
[0220] D. NONE OF THE ABOVE [ACTION: GO TO NEXT QUERY]
[0221] 4. PLEASE REPEAT A BLOOD GLUCOSE SAMPLE {REPEAT M1}
[0222] A. REPEAT M.sub.1>500 [ACTION: GO TO NEXT QUERY]
[0223] B. REPEAT M.sub.1<500 [ACTION: ADD INFO TO
`ERROR.LOG`]
[0224] 5. YOUR BLOOD SUGAR LEVELS ARE DANGEROUSLY HIGH!!! IF LEFT
UNCHECKED YOU MAY DEVELOP SERIOUS COMPLICATIONS. PLEASE SELF
ADMINISTER 8 UNITS OF INSULIN AND CHECK YOUR BLOOD SUGAR LEVELS 8
HOURLY.
[0225] A. OK [ACTION: FOLLOW DIALOGS `INTENSE MONITOR 67`;]
[0226] [ACTION: ADD TASK `CHECK BLOOD SUGAR 11`8 HRLY]
[0227] [ACTION: ALERT PROVIDER IMMEDIATELY]
[0228] Since this high a blood glucose value places the patient at
a particularly high risk for developing complications and has
important implications for further management, it is necessary to
know the exact circumstances under which the patient's blood
glucose has reached these levels. This is done by asking the
patient to reply to PREACUTE DIALOG 1001. Pre-acute dialog 1001
contains queries that attempt to rule out less worrisome causes of
the high glucose levels. For instance, if the patient has forgotten
to take his/her regular dose, then he/she is advised to take it
immediately, and obtain a repeat blood glucose value after 2-3
hours. Likewise, if the patient has had an unusually large meal
that particular evening, he/she is advised to add a little extra
insulin to the usual nightly dose. Alternatively, if the patient
reports that he/she has been running a fever or has been feeling
unwell for a few days, an urgent visit to the healthcare facility
is necessitated, since this level of blood glucose may be the
earliest sign of developing diabetic hyperosmolar coma, a
potentially serious condition where, in the presence of active
infection, the blood glucose control mechanisms are seriously
compromised, and which may rapidly progress to uncontrollable
systemic infection, coma and death.
[0229] Further, if a repeat blood glucose test returns as normal,
and it is determined that the high initial levels were a result of
a technical fault in measuring device 74 (ERROR), then the
information is stored in the ERROR.LOG file, and no further action
needs to be taken, other than recalibrating or replacing monitoring
device 74.
[0230] In the management of depression, a sentinel sign can include
the patient's daily rating of his/her mood on a scale from one
(`feeling low`, depressed) to ten (`feeling great`, happy). Other
sentinel symptoms include the patient's sleep patterns, appetite,
feeling of tiredness and activity, etc. The progressively worsening
of the patient's risk factors on these contexts over a short span
of time, such as a few days is picked up by the monitoring
application, which in turn, may send dialogs to explore the reasons
for the patient's worsening mood, or fix an appointment with the
patient, or alert the healthcare provider depending on the preset
protocol as determined by the provider. The provider may
additionally initiate a phone call to the patient, or make a
personal visit to the patient, or assign additional exploratory
dialogs with the patient (Was there some recent setback at the
workplace/at home, to dear ones/in relationships? Is the patient
regularly taking his/her medications? Have suicidal thoughts ever
intruded (suicidal ideation) into the patient's consciousness?
etc)
[0231] For obvious reasons the patient may be advised to report to
the facility immediately if he/she reports suicidal ideation. In
addition to receiving alerts regarding the worsening mood status of
the patient, the patients' non-response to dialog programs would
constitute sufficient reason to alert the provider. This is
especially more so when accompanied by a recent decreasing mood
level in the patient (increasing R values).
[0232] Similarly, in the case of patients with congestive heart
failure, a history of increasing swelling of the feet (pedal
edema), decreasing urine output and episodes of breathlessness on
lying down (orthopnea) would constitute `sentinel signs` and be
monitored for the purpose of diagnosing `pre-acute` conditions. In
this case an electronic weight scale serves as monitoring device
74.
[0233] Factors Combination 2:
[0234] It is a well known fact in medicine that, while the
treatment of chronic diseases improves the quality of life and
increases longevity, the underlying pathological process remains
unaltered, and often worsens with time. This necessitates an
increase in the dosage of administered drugs, or the consideration
of alternative modes of therapy. Further, the therapeutic effects
of drugs and management tend to decrease on long term usage. For
instance, diabetes patients develop the phenomenon of insulin
resistance, whereby the body requires an increased dosage of
insulin to control the blood sugar level to the same extent as
earlier. Patients on oral hypoglycemic agents may require a
switchover to injected insulin. Similarly, patients with HIV
infection on anti-retroviral therapy would exhibit lowering CD4
counts, as a result of developing resistance of the virus to
drugs.
[0235] Common to the above processes is the fact that the
increasing risk factors takes place over a span of months, or even
years. In these patients, there is a gradual deterioration of risk
status (increasing value of R), especially the values that are
derived as a result of objectively measured physiological
parameters, and scored questionnaires. So, when this is seen in a
previously well patient with more or less unchanged or improving L
and W status, this indicates primary progression of disease. To
confirm the above, the healthcare provider may additionally
prescribe diagnostic laboratory tests, such as tests for
anti-insulin antibodies in the case of diabetes, and Viral Load
tests in the instance of HIV infection.
[0236] Factors Combination 3:
[0237] Notwithstanding the mechanisms to increase the patient's
compliance with medical regimens such as dynamic dialog paradigm,
online dialog dynamism and customization, in a small subset of
patients, resistance (W) to the system may show a gradual increase
with time. This may be accompanied by a simultaneous increase in
risk factors (deteriorating patient's health status). In patients
with low pre-existing risk factors (well-motivated patients), this
may be because the patient feels that he/she has not much to gain
from the system, since his/her health status and management is low
risk anyhow. On the contrary, some patients may be intrinsically
less motivated, or less inclined to look after their health and may
show high risk factors with high resistance.
[0238] Regardless of their present health status, both
patient-groups are still at risk of deteriorating. Patients who
exhibit this phenomenon will show gradually increasing resistance
(W) values over multiple contexts, with or without accompanying
deteriorating R values. Reliability values may also be adversely
impacted. Continuing to remotely manage these patients without
taking into account the developing resistance will have a negative
impact on the effectiveness of treatment regimens and continuation
rates for these patients.
[0239] Managing these patients would require that the patient is
kept on `maintenance content`, content that is primarily
educational in nature, and emphasizes upon the importance of
continued follow up and compliance with medical advice, even if the
patient is otherwise asymptomatic. Further, the patient is taught
that many diseases remain under the surface and progressively hurt
the body, and emerge after a `latent period` even if they do not
cause symptoms meanwhile. Further, there are as few as necessary
dialogs initiated which pertain to the contexts where resistance is
encountered.
[0240] However, some of these issues are likely to be better
resolved at personal interviews and the monitoring application is
crucial in that it serves to detect early and alert the provider to
this happening, in order that a personal interaction is
initiated.
[0241] Factors Combination 4:
[0242] Every effort is made to query patients on sensitive topics
in a non-opinionated fashion and in an open ended manner (by
allowing the patient to elaborate). However some patients may still
feel uncomfortable about answering queries that relate to sensitive
topics. Some of these queries are relevant to the health condition,
and can have a significant impact on the management, e.g. a history
of psychotropic drug abuse in a psychiatric patient, a history of
alcoholism in a depressed patient. Often these queries hold the key
to the diagnosis and the patient's condition may never be treated
wholly without resolving the particular issue. In addition, some
queries, while not directly related to the patient's condition, if
returned positive, can have a significant impact on management,
e.g. a 70 year old lady on follow up for congestive heart failure
is queried about possible elder abuse. The query however is needed
to rule out one of the commonest causes of poor health in these
patients, and is in the best interests of the patient.
[0243] As has been described earlier, patients are allowed the
option of skipping the query, so that the provider is alerted to
the fact that the patient has reasons whatsoever to `dodge` the
query. The reasons can range from the queries' being embarrassing
to the patient (e.g. the screening for elder abuse query) to the
patient's being uncomfortable with that particular aspect of
his/her own life (e.g. alcoholism and drug abuse in a depressed
patient) may simply wish to skip it.
[0244] High resistance implies that the patient has reasons not to
answer the query, whatever the reason. When there is high
resistance in a single or a few isolated contexts, with or without
low reliability for these contexts, the information regarding this
context within profile database would need to be elaborated by the
provider. This aspect could be explored at a personal interview and
after the patient's confidence has been secured. Alternatively this
is done through the usage of dialogs that are more sensitive in
their language, and of dialogs which look for indirect pointers
towards the condition of interest. Alternatively, `pointer queries`
are used which query the patient on topics that are commonly
associated with the health condition of interest.
[0245] For example, consider patient D on follow up for depression
who gives a history of occasional alcohol intake. However, he says
that he has no more than a couple of drinks a week, and denies that
he has ever consumed alcohol as a result of his feeling `low`.
Ruling out alcoholism (different from `social drinking` and
moderate alcohol consumption) is very important in this patient,
since his depression may in fact be a result of chronic intake of
alcohol. It is a well known fact that alcoholics tend to downplay
the actual level of their intake, so the patient's denying a
history of alcoholism would not be of much value in the decision
making process.
[0246] Instead the system attempts to assign risk scores for
alcoholism by looking for indirect markers within medical claims
database 58, medical records database 62 and additional material
54. These include a history of frequently presenting in the
emergency room with cuts and wounds (acquired as a result of the
drunken state), frequent lung infections (from aspiration of
stomach contents), a history of being in a de-addiction program,
the physician notes of prior consultations, results of old/recent
laboratory tests that are associated with a high alcohol intake (a
fatty liver, and cirrhotic changes seen on a CT scan done for some
other reason, a raised blood Gamma-glutamyl-transferase enzyme
levels), etc.
[0247] Further, the patient may be administered adaptations of
standardized screening tests for alcoholism, such as the CAGE Test,
either at a same sitting, or spread out over many sittings. If the
screening tests and/or any of the indirect markers test as
suspicious for alcoholism, this aspect is explored further through
the use of dedicated dialogs or at personal communications with the
patient.
[0248] In addition, patients who are former alcoholics, and who are
on follow up for some other condition may be monitored for a
recurrence of uncontrolled drinking in the same manner as detailed
above. In this case, these patients may be monitored and provided
supportive management in order to prevent the alcohol recurrence
status.
[0249] Factors Combination 5:
[0250] When the information obtained from the patient over multiple
contexts has a low reliability, or the reliability factors are
decreasing, this indicates that the information obtained from the
patient is inherently unreliable. Healthcare management in these
patients cannot be done on the basis of patient provided inputs
alone, since the patient provided data is seen to be unreliable
(inconsistent with known data) and possibly fallacious.
[0251] Malingering is the condition where a healthy person
simulates symptoms of disease for personal gain. Malingerers often
present to medical providers as difficult to diagnose, a typical
cases. While this gives the provider an initial clue to the real
cause of the patient's symptoms, it is seldom possible to make a
sure diagnosis of malingering on this basis, in an individual
patient. Further, some malingerers do not simulate new symptoms but
simply exaggerate the severity of the disease state for the sake of
gain, which may be financial compensation from a former employer,
or an insurer; not having to go to work and sympathy from dear
ones. Malingering is responsible for a significant portion of
wasteful healthcare expenditure, and detecting this is a priority
in any healthcare system.
[0252] An important characteristic of malingering is a frequent
change in the patient's `story` and inconsistency between the
history and the severity of symptoms. It is often difficult for the
provider to keep track of every single detail of the
patient-provider interaction in order to seek inconsistencies in
the history. Even if this were to be done, an attempt to visualize
the data in the context of malingering would not prove to be
worthwhile in view of the time, effort and cost expended.
[0253] This however may be done through the use of the system in
the following manner: when the reliability is low for one or more
contexts and there is an inconsistency in the risk factor scores
obtained from different sources, the suspicion of malingering is
raised. This is because malingerers do not report groups of
symptoms in the manner that they usually present to a provider.
This is different from Factors Combination 4 in that the
unreliability is spread over multiple contexts, and that these
contexts do not necessarily deal with `sensitive topics`, or those
topics where an ordinary person would be uncomfortable in relating
to a provider.
[0254] Further, these patients are likely to exhibit high
resistance, and this is spread over multiple contexts. This is
because these patients have a constant underlying fear of being
`caught` and so try to avoid supplying additional data as much as
possible. So, they are far more likely to choose `Skip`, when
compared with a patient who is genuinely suffering from the
condition.
[0255] The mechanism may be further refined such that within a
dialog, there are no more than one or two queries of interest,
which are interspersed within general queries pertaining to
`neutral` topics, i.e. those topics which are likely to put the
individual off-guard with regard to the true significance of the
query.
[0256] Factors Combination (Miscellaneous):
[0257] If a new intervention is planned for a specific condition,
and is in relation to a particular context, such as a novel drug
for the control of blood glucose, it would be possible to choose
those patients who are having greatest difficulty in controlling
their blood sugar levels based on their computed R values. This
group of patients would likely represent the diverse background of
patients with diabetes, with regard to other parameters. If a
randomized study comparing the efficacy of the two treatments is
planned, then the control group (patients who are not treated at
all, or those who are placed on standard treatment) can consist of
patients with comparable R values. The R values of patients
following the new treatment can be followed to see the response of
the glycemic control to the new drug. In this manner, R values may
be applied in the context of conducting research studies, and
clinical trials in the evaluation of new drugs, etc.
[0258] Improving R values that relate to the symptoms and quality
of life factors, after the institution of a new drug would suggest
that the new drug is more effective in controlling this given set
of symptoms in the patient, and is more likely to be accepted in
this patient.
[0259] A final advantage of the system is that there is huge
research benefit in identifying parameters that can correlate to
subgroups of patients who respond differently to treatment. All
data relevant to the patient's health condition, including the
patient's replies to specific queries are stored in the profile
database, each in a separate column. Additional columns describe
the response of the patient to specific medical interventions (in
the form of the change in the R values as a function of time, i.e.
pre-treatment and post-treatment values of R)
[0260] Chart 5 depicts an example profile database for a group of
10 patients on treatment for depression, who additionally consume
alcohol. Some of the patients being followed up agree to cease
consuming alcohol consumption completely, and are given support and
motivation to do so, while the remainder patients are placed on the
regular follow up schedule for depression. For the sake of clarity,
only three contexts (1, 2 and 3) are shown in the chart. Cumulative
L and W values for the patient are shown. However, in the actual
study, the R, L and W values of more contexts, and more patients
would be taken into account.
8 CHART 5 R, P, F VALUES OF SPECIFIED CONTEXT R, CEASE APRIL 2002
JANUARY 2003 PATIENT ALCOHOL R.sub.1 R.sub.2 R.sub.3 L.sub.ALL
W.sub.ALL R.sub.1 R.sub.2 R.sub.3 L.sub.ALL W.sub.ALL #1 YES 3.5
1.6 2.5 6.1 0.7 1.6 1.2 2.4 10.1 10.4 #2 NO 1.6 5.1 1.9 2.8 0.9 1.5
5.5 1.7 2.5 0.9 #3 YES 11.5 5.3 4.1 9.1 1.4 6.1 2.1 2.8 10.1 1.2 #4
NO 1.2 1.4 1.7 12.1 0.4 1.1 1.2 1.5 12.2 0.4 #5 NO 10.6 6.2 3.8
14.1 1.7 8.1 4.2 3.4 11.2 1.8 #6 NO 5.5 4.0 2.2 8.1 5.4 3.9 6.1 2.6
8.2 4.3 #7 NO 3.4 1.3 2.8 7.5 0.9 2.8 1.2 2.6 7.4 0.7 #8 YES 1.0
1.2 1.1 9.8 0.6 1.0 1.3 1.2 9.6 0.5 #9 YES 5.4 4.1 2.1 0.4 1.7 3.7
6.4 2.3 0.6 1.6 #10 NO 10.6 6.2 3.8 14.1 1.7 8.1 4.2 3.4 11.2
1.8
[0261] A researcher wishes to know whether depressed patients who
successfully cease alcohol consumption have a better prognosis than
those who don't. In order to find out, he/she searches the entire
profile database of patients on follow for depression to select
patients for inclusion into one of two groups--test and control
group. The test group consists of individuals who have completed
the alcohol cessation program, while the control group consists of
those who have not opted to cease alcohol consumption. The
researcher now matches patients in the groups such that there is a
patient with similar R, L and W factors in the test group for every
patient with similar factors in the control group. Matching for
factors is done for those contexts that are known to have a bearing
on the long term prognosis of the depression, including the
age-group, sex of the patient, the dosage of the drugs administered
to the patient and most important, the initial risk state of the
patient's depression. Of course, matching is not done for the
context of alcohol cessation (Context `CEASE ALCOHOL` in the
chart), since this is what the researcher is interested in
studying.
[0262] Matching is preferably done by the use of automated
algorithms that match patients who exhibit similar values for the
risk factors that are to be matched. The matching algorithm also
excludes the patients whose data has
[0263] 1. low reliability, since a study done on unreliable data is
bound to decrease the quality of the study
[0264] 2. high resistance, since such patients are unlikely to
adopt or continue the new management in any case. Management in
these patients is directed towards decreasing the resistance first,
and then attempt behavior modification.
[0265] 3. very low risk factors of interest--the change in the
values of R in the context of interest is likely to be so low, that
the likelihood of errors is greatly increased in this case.
[0266] Chart 6 shows an algorithm for the automated assignment of
patients to the two groups. Individuals whose risk factors and
values meet allocation criteria as decided by the researcher are
automatically assigned into the groups as shown below.
CHART 6
[0267] TEST GROUP
[0268] IF CEASE ALCOHOL=`YES`
[0269] AND R.sub.1>2.5
[0270] AND R.sub.2>2.0
[0271] AND R.sub.3>1.5
[0272] AND L.sub.ALL>4.0
[0273] AND W.sub.ALL<3.0
[0274] THEN PATIENT =`TO TEST GROUP`
[0275] CONTROL GROUP
[0276] ELSE IF CEASE ALCOHOL=`NO`
[0277] AND R.sub.1>2.5
[0278] AND R.sub.2>2.0
[0279] AND R.sub.3>1.5
[0280] AND L.sub.ALL>4.0
[0281] AND W.sub.ALL<3.0
[0282] THEN PATIENT=`TO CONTROL GROUP`
[0283] EXCLUDED GROUP-REASON LOW L VALUES
[0284] ELSE IF L.sub.ALL<4.0
[0285] THEN PATIENT=`EXCLUDED GROUP REASON: LOW L VALUES`
[0286] EXCLUDED GROUP-REASON HIGH W VALUES
[0287] ELSE IF W.sub.ALL>3.0
[0288] THEN PATIENT=`EXCLUDED GROUP REASON: HIGH W VALUES`
[0289] EXCLUDED GROUP-REASON LOW R VALUES
[0290] ELSE IF R.sub.1<2.5
[0291] OR R.sub.2<2.0
[0292] OR R.sub.3<1.5
[0293] THEN PATIENT=`EXCLUDED GROUP REASON: LOW R VALUES`
[0294] It is obvious that there may be more than one reason to
exclude a patient from the study. On the basis of the exclusion
criteria, patients are assigned to the groups as shown in Chart 7
below.
9CHART 7 INCLUDED PATIENTS TEST CONTROL EXCLUDED PATIENTS GROUP
GROUP REASON #1 #7 LOW L VALUES #9 #8 #2 HIGH W VALUES #6 #3 #5 LOW
R VALUES #4, #8
[0295] Once this is done, the researcher has two groups of patients
who are comparable in all respects, except that one group has
ceased alcohol consumption at that point of time, while the other
group continues to consume alcohol. Now, the researcher analyzes
the difference between the present R factors of the two groups. In
addition, it is also necessary to verify that none of the patients
enrolled in the study have developed an unacceptably high
resistance or low reliability during the course of the study.
[0296] Standard statistical methods are used to test the level of
significance of the R factors as a result of the alcohol cessation.
R factors of more than one context, which are scored on the basis
of replies, may be used to measure the improvement in the risk
status of the patient. Alternatively, the R values are derived from
standard questionnaires that score the risk status of the patient
objectively e.g. SF12, SF36, Minnesota Living with Heart Failure
Questionnaire, Geriatric Depression Scale. Alternatively,
laboratory tests may be used to evaluate the effectiveness of the
management e.g. in the case of blood glucose control, Serum
Hb.sub.AIC values, that give an indication on the long term (80-100
days) level of blood sugar control may be used. If there is a
statistically significant benefit (which may not be attributed to
chance alone) from these analyses, it is concluded that alcohol
cessation in depressed patients is helpful in the treatment of
their depression. In this manner a researcher is able to prove or
refute earlier hypotheses.
[0297] The advantage in using the system for the purpose of
research is that the method of storage of data in the form of R, L
and W values is ideally suited to risk assessment. In case of
doubt, since all inputted replies to queries would additionally be
archived by the system, it is possible to validate the assigned R,
L and W values. Further, since the data for patients over multiple
healthcare facilities is stored in the same database format, it is
possible to integrate the results over multiple healthcare
facilities. This makes it convenient to conduct large multi-centre
studies, and further validates the results of the study.
[0298] In addition, the cost of conducting the study is greatly
reduced, since the data is already preformatted for the purpose of
comparison. It is of importance to note that any context can be
studied in relation to any other context over a large group of
patients. It is also possible to conduct studies comparing the
effectiveness of different healthcare facilities management
protocols by a similar method. Further, it is possible to randomly
assign patients to one of two groups (prospective randomized
controlled trials), each receiving different treatments and
comparing the effects of the different treatments. This information
may be further added into the new protocols for better automated
content assignment. So this method may also be used in the
evaluation of different protocols of management in the system, and
be used to refine the protocols assignment process in any group of
patients.
[0299] Process Governing Dynamics:
[0300] In view of the above, it is apparent that the provider
utilizes a combination of automatically selected and manually
assigned content to elicit further information from the monitored
individuals. Automated content utilizes a combination of the level
of risk of the patient, the immediacy/urgency of delivering the
content to the individual, the individual's reliability and
resistance, and the individual's profile variables such as
comprehension capacity, motivational drivers, etc. Automatically
assigned content is queued in the tasklist database, for
communication with the individual at the next connection. However,
the provider may also assign content manually from the report
generator interface, and has the additional option of modifying,
appending and deleting content from the tasklist database.
[0301] The monitoring application regularly scans the individual's
entire profile and seeks variables, and combinations of variables
within the profile, for which there are defined protocols within
the protocols database. Protocols are program statements that
instruct the monitoring application to perform actions 140 when a
given set of conditions are fulfilled. Actions 140 include alerting
the healthcare provider to the presence of a combination of
variables that may signify an impending `pre-acute` state in the
patient. The protocol shown below is for patients with mood
disorders. A gradually worsening mood over many days is often the
first sign of an incoming depression in these patients. Since these
patients are at risk for committing suicide in their depressed
state, it is important that their mood fluctuations be continuously
monitored, and that these individuals are prevented from
deliberately harming themselves.
[0302] PROTOCOL
[0303] (1) IF PATIENT REPLY RECEIVED=`FALSE`
[0304] AND MOOD.sub.RECENT=`FEELING TERRIBLE`
[0305] THEN ACTION: ALERT PROVIDER, `PATIENT NOT RESPONDING`
[0306] (2) ELSEIF MOOD.sub.CURRENT=`FEELING TERRIBLE`
[0307] AND MOOD.sub.CURRENT-1=`FEELING FINE`
[0308] AND/OR MOOD.sub.CURRENT-2=`FEELING FINE`
[0309] THEN ACTION: FOLLOW WORSENING_MOOD DIALOG
[0310] (3) ELSEIF MOOD.sub.CURRENT=`FEELING TERRIBLE`
[0311] AND MOOD.sub.CURRENT-1=`FEELING TERRIBLE`
[0312] AND MOOD.sub.CURRENT-2=`FEELING TERRIBLE`
[0313] THEN ACTION: FOLLOW PERSISTENT DYSPHORIA DIALOG
[0314] (4) ELSEIF MOOD.sub.CURRENT=`FEELING FINE`
[0315] AND MOOD.sub.CURRENT-1=`FEELING GREAT`
[0316] AND MOOD.sub.RECENT=`FEELING GREAT`
[0317] THEN ACTION: FOLLOW MOOD_SURVEILLANCE PROTOCOL
[0318] In the first condition in the protocol, the monitoring
application is activated when replies to queries are not received
in time by the system. Given that the patient has recently stated
that he/she has been feeling terrible, the lack of reply is likely
to suggest a worsening mood level in the patient, since patients
who are very depressed are unable to sum up enough energy to even
perform the activities of daily existence. The action here is to
alert the provider, so that he/she may initiate personal contact
with the patient, since dialogs are unlikely to be of any benefit
(the patient has ceased responding to the remote apparatus)
[0319] The second condition of the protocol, if returned true,
implies that the patient's mood has been worsening. This may be
because of external factors, such as problems with family, and at
the workplace, or may be without any apparent reason. The reason
for the worsening mood is explored using the WORSENING_MOOD dialog.
If it is determined that the patient's mood is worsening without
any apparent reason, then such a patient is kept on close follow
up. Alternatively, the provider may schedule a priority appointment
with the patient.
[0320] The third condition of the protocol, if returned true,
implies that the patient has been steadily feeling terrible. This
may be because of the patient's failure to take medication or due
to a recent event, or even due to the failure of the medication to
act in the particular patient. The reason for the patient's
persistently low mood is elucidated by the PERSISTENT_DYSPHORIA
dialog.
[0321] In the fourth condition, the patient's mood has been
deteriorating, though it is still at the level of `feeling fine`.
This patient would require more intensive monitoring for worsening
mood, and the monitoring application switches the patient to the
MOOD-SURVEILLLANCE_PROTOCOL, whereby there is more frequent
communication between the patient and the system, and the threshold
for alerting the provider to a worsening mood is lowered.
[0322] While the preferred embodiment of the invention has been
illustrated and described, as noted above, many changes can be made
without departing from the spirit and scope of the invention.
Accordingly, the scope of the invention is not limited by the
disclosure of the preferred embodiment.
* * * * *